Chemi Peres

1. Snapshot and testable edge thesis

Investor and scope. Peres founded and managed the listed Mofet vehicle in 1992, then joined Rami Kalish in 1996 to raise and manage Polaris II, the successor platform later named Pitango. Mofet, Polaris I, Polaris II, later Pitango funds, specialist/continuation entities, board roles, and nonprofit initiatives are separate objects. Wharton, formation passages · Globes 2000, institutional history

Testable edge thesis — [researcher inference · moderate]. In Israeli cross-border technology venture from 1992 through the later Pitango platform era, Peres's repeatable advantage is best described as institutional translation under changing constraints: identify whether capital duration, market feedback, leadership complementarity, customer access, or vehicle mandate blocks a technically strong company; then assemble the relationship, vehicle, and cross-border bridge that can remove that bottleneck as the evidence changes. profile thesis

Mechanism and scope. The mechanism should operate most clearly where strong technical formation meets a small home market, scarce commercialization capacity, or a customer/regulatory interface. It should be weaker where the bottleneck is basic product quality, founder integrity, price, or demand that no network or capital structure can repair. [researcher inference · strong]

Predictions. Unseen deal records should show dated deferrals with explicit evidence contracts; re-entry only after the named uncertainty changes; vehicles matched to duration and specialist authority; founders identifying a specific market/team/governance contribution; and the same behaviors in ordinary companies and failures. [researcher inference · strong]

Falsifiers. The thesis fails if pass/IC records show no such contracts, founders and partners assign the work elsewhere, vehicle choice was incidental, ordinary cases show no intervention advantage, or complete vintage-adjusted cash flows show no improved outcomes after team, access, sector, era, and outliers are controlled. The public record currently covers 12 hand-researched possible Peres-linked cases, only six sufficiently reconstructed, against a live 140-card Pitango page and a firm claim above 300 companies; career coverage and audited personal/fund returns are not computable. denominator reconciliation [researcher inference · strong]

2. Core philosophy

Peres's own words first. Across the direct record, he favors founders who are adaptable, persistent, honest, and expert; capital whose information cadence matches long-duration company building; portfolios balanced across sector and stage; global-market bridges from a small home economy; and ambition beyond an acquisition or valuation milestone. Wharton founder filter · Globes 2007 diversification · Human Founder, 47:42–51:27 [investor-stated · strong]

His recent essays add three claims: reference sets must change with technology and capital regimes; a geography can provide innovation density or fast feedback without being the ultimate market; and dual-use/B2G companies need public institutions as customers and constraints. Calibrate Bigger, “How we got here” · PMF essay, X/Y sections · B2G essay, five drivers [investor-stated · strong]

Synthesis — [researcher inference · moderate]. The operating sequence is: identify a structural change; choose a vehicle that can wait; test the founder's learning and truthfulness; specify what the home ecosystem can and cannot validate; use networks as evidence channels rather than proof; scale only after the reference set and customer authorization justify it; then debrief and resize when facts break. writings cross-source synthesis · talks recurring frameworks

The public doctrine is stronger on aspiration and institution building than on valuation, ownership, reserve allocation, rejected deals, and stop rules. Any usable version must add those controls rather than treating optimistic language as underwriting. [researcher inference · strong]

3. Sourcing

• Longitudinal contact. The Taboola retrospective describes three meetings across an evolving model and a decisive referral by Zvi Limon. That supports keeping a relationship open while evidence changes, but IPO selection and unresolved dates prevent a universal hit-rate claim. Taboola essay, opening and lessons [documented behavior · moderate]
• Cross-institution networks. Mofet/Dash Polaris co-investment, Israel–U.S. institution building, strategic LPs, co-investors, and company boards created multiple routes to founders, capital, customers, and references. One Via/Japan story is evidence of a selected bridge, not a channel-level outcome cohort. Shany note · Mizuho note, 30:07–30:51 [researcher inference · moderate]
• Small-market search. The consistent target is Israeli technical density with global commercialization potential; a U.S. subsidiary was visible at Shany, and later doctrine separates innovation from customer-feedback scale. Shany lab · PMF essay [researcher inference · moderate]
• Access boundary. Public roles and long tenure are documented access-relevant facts; family-name and military/technical-network access are hypotheses, not established mechanisms. No source reviewed shows that either hypothesis broadened deal introductions, and no funnel by channel or matched counterfactual shows whether any access route converted better. [researcher inference · insufficient public record] Taboola proxy, director biography
• Reproducible sourcing rule. Log originator, referrer, first meeting, why the company was not yet underwritable, next evidence, recontact trigger, and competing access path. Attribution at exit must be compared with this contemporaneous log. [researcher inference · strong]

4. Picking

Founder. Peres's stated filter favors adaptable, persistent, honest domain experts rather than founders whose plan never changes. Taboola adds complementary leadership and a culture of transparent coordination; Human Founder adds recruiting and organizational DNA at scale. Wharton founder filter · Taboola note · Human Founder note [investor-stated · strong]

Market and product. In a small home market, the IC must state whether a local test resolves technical efficacy, user behavior, procurement, unit economics, or only urgency. The X/Y framework becomes investable only when foreign regulation, distribution, price, and customer authority are separately tested. PMF essay, “Changing Global Footprint” [researcher inference · strong]

Evidence threshold. Taboola's 2013 packet included self-reported 7.5x revenue growth, named publishers, mobile launch, operating hires, and international use of funds; Via's 2015 packet included cumulative/weekly rides, members, vehicles, and a constrained initial corridor. Those facts make the decisions replayable, but valuation, security, ownership, audited cohorts, IC dissent, and partner division of labor remain missing. Taboola lab · Via lab [contemporaneous record · strong]

Red flags and kill criteria — [researcher inference · strong]. Kill or defer when customer evidence cannot travel to the target market; founder adaptation cannot be distinguished from narrative drift; one government buyer controls authorization and cash; defense and civilian requirements conflict; valuation requires unprecedented outcomes without ownership math; or partner capacity cannot cover the next proof. B2G note · dual-use note

The record does not disclose a numerical scorecard, false-positive/negative founder cohort, pass sample, or Peres-specific IC vote. Picking skill therefore remains a qualitative doctrine with a few observable decisions, not a measured classifier. [researcher inference · insufficient public record]

5. Deal and portfolio mechanics

Peres publicly argued for sector diversification, stage balance, and a large partnership able to support companies through crises. In 2007 the platform described roughly 1,000 proposals and 10–12 investments annually. The same record contained eight closures among 56 Fund III companies and an adverse interim mark, so balance is a risk-allocation claim, not proof of superior return. Globes 2007 note [investor-stated · strong]

Pitango III's reported plan was 30–35 companies, $6–12 million of lifetime capital each, initial-stage entry, meaningful ownership, and follow-on support; Yoram Mankovski, not Peres, stated the detailed policy. By 2007, reporting cited 56 companies. Globes 2000 · Pitango III lab [contemporaneous record · strong]

Public evidence is insufficient on entry ownership targets, price discipline, securities, reserves, pro rata thresholds, concentration bands, decision rights, and exit contracts. Taboola's $15 million and Via's $27 million are total rounds, not Pitango checks. Via's retained shares and Taboola's partial affiliated-fund realization are issuer facts, not audited fund or personal returns. portfolio forensics [contemporaneous record · strong]

Operational mechanic — [researcher inference · strong]. Each IC should link the milestone to check, security, ownership, reserve requirement, responsible partner, and next decision date; each follow-on should be re-underwritten against prospective fund contribution rather than prior narrative. The record supports the need for this control but does not prove it was Peres's published procedure.

6. Vehicle and incentive fit

The vehicle is part of the strategy, not proof of the strategy. Private structure can create patience and opacity simultaneously; specialist and continuation vehicles can create expertise and allocation conflicts simultaneously. [researcher inference · strong]

7. Team attribution and work with founders

Peres's founder doctrine preserves adaptability and agency: the investor can debate an exit, but the founder remains at the wheel. That principle is morally attractive and operationally incomplete where LP liquidity, employee concentration, fiduciary duties, or conflict diverge. Globes 2014 note [investor-stated · strong]

8. Views on the craft

• Power law. Peres acknowledges that only a small fraction of startups become large, but the public record does not quantify contribution concentration. Pareto is therefore a research-ordering lens, not a literal claim. Globes 2014 note · Pareto ranks [researcher inference · strong]
• Fund size. A larger platform can supply sector/stage breadth and partner support, yet Pitango III demonstrates deployment, attention, and vintage risk. Size is justified only by ownership opportunity, specialist capacity, and reserves, not ambition alone. Pitango III lab [researcher inference · strong]
• Access versus selection. Networks can find companies and customers, but no channel-level conversion or rejected-deal sample separates access from judgment. [researcher inference · insufficient public record]
• Governance. Peres favors founder agency, candid debriefing, and patient support; GoRefer supplies one public stop, Webbing one ambiguous price/exit case, and no board-minute cohort. GoRefer lab · Webbing lab [documented behavior · moderate]
• Exits. An IPO or sale is not the end of company building and is not a return until ownership and cash flows are known. Via's retained stake and Taboola's partial realization demonstrate the distinction. Globes 2014 note · economic ranking [researcher inference · strong]
• Great investor. The evidence-backed ideal is a truth-seeking team builder who matches people and capital to the bottleneck, gives founders room to adapt, and remains accountable to milestones and attribution. Whether Peres repeatedly met that standard is only moderately supported. [researcher inference · moderate]

9. Signature bets and decision process

Do not read the outcome before making the decision. Each case freezes the available evidence, unknowns, alternatives, risks, provisional underwriting, invest/pass choice, kill criteria, terms/ownership logic, and next-check milestones before the reveal.

1. Shany, 1993 — a bootstrapped exporter and two separate emerging vehicles.
2. AudioCodes, late 1990s — technical/market validation with disputed partner credit.
3. Taboola, February 2013 — invest after business and operating evidence resolves, not at the first dream.
4. Via, April 2015 — dense local evidence against city-expansion and regulatory risk.
5. GoRefer, 2001–2002 — the required failure lab: stage proof and preserve cash when customers disappear.
6. Pitango III, 2000 — underwrite a fund's pacing, reserves, partner load, and closed-market scenario.
7. Webbing, before 2021 — weigh a public listing against private growth and strategic exit.

These labs teach more about evidence boundaries than personal returns: Shany is retrospective, AudioCodes is team-disputed, Taboola and Via are platform cases, GoRefer is a shared stop, Pitango III is a fund-level adverse test, and Webbing lacks the board record. [researcher inference · strong]

10. Mental models and maxims

The field guide reconciles eight distinct supportable rules rather than treating every slogan as a model:

1. Convert “not yet” into an evidence contract.
2. Back adaptable truth-seekers and build the complementary team.
3. Separate innovation density, test-market evidence, and scale-market proof.
4. Match vehicle duration, disclosure, and authority to the work.
5. Diversify risks and capital sources without diversifying away attention.
6. Recalibrate reference sets in both directions before scaling ambition.
7. Underwrite public and dual-use markets through independent authorization and economics gates.
8. Debrief losses, preserve residual value, and earn the next check.

Each is a reconstruction unless the model's evidence-status bullet says otherwise. The guide separates documented applications from researcher-applied illustrations and keeps impact-without-return-concession, “giraffes,” LP strategic value, and founder-controlled exits in the candidate ledger where implementation evidence is thinner.

11. Evolution and contradictions

The stable belief is adaptability—of founder, company, vehicle, and reference set. The unresolved tension is whether continuous adaptation is falsifiable: after success it can look like learning; after failure it can look like drift. Dated decision records and explicit kill criteria are the necessary control. [researcher inference · strong]

12. Failures and limits

• GoRefer. Customers disappeared; founders and co-investors shut down while cash remained. The record supports a specific stop, not a portfolio-wide loss-control system or documented later procedure. Globes 2002 [contemporaneous record · moderate]
• Shopserve. It closed after the Palo Alto content-company experiment, but no company-specific thesis, warning, board response, or cash outcome survives in the researched record. Globes 2001 [contemporaneous record · insufficient public record]
• Pitango III. Press-reported interim marks moved from favorable to sharply adverse and later less adverse under different dates and definitions. The series cannot establish final performance but directly challenges a seamless diversification story. mark trajectory [contemporaneous record · strong]
• Webbing IPO path. Abandoning a listing over valuation and later selling strategically may show discipline; without the rejected valuation, board vote, ownership, and proceeds, the counterfactual is unresolved. CTech 2023 [researcher inference · moderate]
• Misses. A 2007 funnel implies many annual passes, but no named pass sample or outcome-matched anti-portfolio was found. Founder, market, and network heuristics therefore cannot be scored for false negatives. Globes 2007 [contemporaneous record · insufficient public record]
• Selection and survivorship. Current portfolio pages, issuer filings, press profiles, and retrospectives select different visible subsets. Six of 12 researched possible Peres-linked cases remain insufficiently reconstructed; career coverage is not computable. denominator [researcher inference · strong]
• Era and role dependence. Israeli policy, communications/Nasdaq demand, dot-com capital, founders, co-investors, and multiple partners can produce outcomes without a distinctive Peres-only edge. sensitivity controls [researcher inference · strong]

13. Historical operating environment

Early 1990s. Israel had deep technical capability but a young venture market and small local demand. Inbal public vehicles and Yozma-backed private partnerships were different responses to the financing gap. Shany's pre-financing U.S. subsidiary supplied a rare ex-ante export signal; Intel's acquisition was not knowable. Wharton, Yozma/Inbal passages · Shany note [contemporaneous record · strong]

Late 1990s. Private fee/carry structures could finance a support team; Nasdaq and the communications wave supplied exit opportunity. AudioCodes must be read inside syndication, sector, and public-market conditions, not as timeless personal selection proof. historical context [researcher inference · strong]

Dot-com and aftermath. Large capital pools met collapsing customer and financing demand. GoRefer and Shopserve are company-level failures; Pitango III's shifting marks are a fund-level measurement warning. No contemporary industry base rate or final cash-flow series supports a relative performance rank. Pitango III lab [contemporaneous record · strong]

2010s–2020s. Taboola and Via arrived with visible operating evidence and long paths to public markets. Later stage/sector vehicles and continuation entities could match specialization and duration, while increasing allocation, valuation, transfer, fee, and conflict complexity. Via 424B4, n.12 [researcher inference · moderate]

Current doctrine. AI, quantum, dual-use, resilience, and B2G reflect real technological and institutional questions, but the processed record offers essays and mandates rather than a complete current cohort. Crisis adoption, government budget, and military validation must not be mistaken for voluntary civilian product-market fit or unconstrained export demand. dual-use note · B2G note [researcher inference · strong]

14. Modern VC translation (as of 2026-08-22)

15. Comparison with completed peers

These three completed peers are dimension controls, not a performance ranking. Yigal Erlich is the closest Israeli vehicle/ecosystem comparator; Jim Breyer is the closest institutional/global-network and long-board-tenure comparator; Trae Stephens is useful only for present-day defense/dual-use authorization and role-conflict questions. Their eras, geographies, vehicles, and evidence quality differ. [researcher inference · strong]

Bounded conclusion — [researcher inference · strong]. Erlich shows the public architecture that made the 1990s private-manager market possible; Peres shows one operator's move from a public venture vehicle into that private architecture. Breyer tests whether network and board visibility can be separated from platform/team/outlier economics. Stephens tests whether Peres's current dual-use/B2G doctrine has enough authorization, productization, and conflict controls for a later regime. None supplies a valid synthetic return comparison.

16. The strongest case against greatness

Skeptical thesis — [researcher inference · strong]. Peres's visible reputation may be explained by access to the Yozma-era ecosystem, Pitango's team and vehicles, founder execution, favorable technology/exit windows, selected retrospective narratives, and a few public outcomes—without evidence of repeatable Peres-specific selection alpha or audited personal/fund economics. A prominent family/public profile is an additional visibility confounder to test, but the reviewed record does not establish that family standing generated deal access or returns.

1. Access and timing. Israel's technical base, Inbal/Yozma policy, Dovrat–Shrem, Kalish, U.S. markets, and Nasdaq demand were preconditions no individual created alone. historical environment [researcher inference · strong]
2. Family and public visibility. A prominent surname and institutional roles may widen doors and later narration, but the causal channel and matched comparison are absent; it must remain an alternative, not an accusation or fact of performance. [researcher inference · insufficient public record]
3. Team misattribution. Shany was Mofet plus Dash Polaris; AudioCodes includes a syndicate and Eyal Kishon; Taboola names Limon and multiple participants; Via and filings show multi-entity/shared-principal power. people map [contemporaneous record · strong]
4. Missing denominator and survivorship. The current 140-card directory is far below the firm's 300+ claim; the 12-case research set is deliberately finite; six are unresolved; passes and complete historical failures are absent. denominator [researcher inference · strong]
5. Economics are not personal returns. Company acquisition prices, round sizes, position marks, retained shares, and partial vehicle cash are not cost-basis-adjusted DPI/TVPI/IRR or Peres proceeds. economic ranking [contemporaneous record · strong]
6. Outlier sensitivity. Removing Taboola and Via removes the richest cap-table cases and only disclosed partial realization, leaving no rankable economic record. ex-outlier test [researcher inference · strong]
7. Adverse ordinary record. Shopserve and GoRefer closed; Pitango III's marks were adverse and unstable; AudioCodes credit is ambiguous; no public before/after process change proves learning efficacy. failures [contemporaneous record · strong]
8. Doctrine after outcomes. Taboola and Shany are late retrospectives; “giraffes,” dual use, B2G, and current specialization lack comparable outcome cohorts. writings limits [researcher inference · strong]
9. Era dependence. Communications/Nasdaq, dot-com capital, digital distribution, and current strategic-government demand may produce non-repeatable advantages. historical context [researcher inference · strong]

The favorable thesis wins if original pass/IC records, role ledgers, board minutes, founder/co-investor testimony, allocation/conflict controls, and complete vintage cash flows repeatedly connect Peres-specific bottleneck removal to better ordinary and ex-outlier outcomes. The skeptic wins if deal access was mostly warm/institutional, decisive work belonged to partners/founders, vehicles did not improve decisions, public doctrine was not implemented, or full economics were ordinary after fees, failures, and vintage controls. [researcher inference · strong] terminal uncertainty ledger

17. Analyst study guide, glossary, and go deeper

Recommended reading path

1. Read the biography through the 1992 Mofet / 1993 Polaris I / 1996 Polaris II control.
2. Process the authored writings, especially Taboola, Calibrate Bigger, PMF, and dual use.
3. Process the spoken talks, beginning with Wharton part 2, Globes 2007, IsraelTech, Calcalist 2026, and Human Founder.
4. Complete every decision lab before reading its reveal. Do not skip GoRefer or Pitango III.
5. Use the mental models as diagnostic checklists, then return to the skeptical thesis.

Investor-specific glossary

Reusable sourcing checklist

• Record source, referrer, first contact, warm-access dependency, competing channels, and conflicts.
• Preserve legal vehicle and date; never merge Mofet, Polaris I, Polaris II, later funds, boards, and nonprofit roles.
• Log why the company is not yet underwritable and the evidence/recontact trigger.
• Sample passes and ordinary outcomes, not only IPO retrospectives.
• Reconcile later credit with the contemporaneous sourcing/role ledger.

Reusable diligence and IC checklist

• Which bottleneck is technical, founder/team, customer, market transfer, capital duration, authorization, or governance?
• What is known, unknown, disputed, and unavailable at the freeze date?
• Which independent reference can falsify the founder and market narrative?
• What check, security, ownership, reserve, and milestone fit the risk; what result kills the deal?
• Does local evidence travel across regulation, procurement, language, price, channel, and voluntary demand?
• For dual-use/B2G, who authorizes, budgets, tenders, certifies, audits, renews, restricts export/end-use, and bears liability?
• Who sources, sponsors, diligences, approves, negotiates, governs, recruits, reserves, and decides the exit?

Reusable portfolio and governance checklist

• Pre-register sector/stage correlations, partner capacity, reserve coverage, vintage pacing, and closed-exit-market scenario.
• Re-underwrite follow-ons prospectively; never equate patience with no stop rule.
• Separate company outcome, security value, vehicle cash flow, net fund return, carry, and personal proceeds.
• Test the record ex-Taboola/ex-Via and against ordinary/failure cases.
• At every board decision, log founder agency, fiduciary duties, employee/LP liquidity, conflicts, alternatives, dissent, and counterfactual.
• At debrief, compare forecast with result, name root cause and changed control, preserve residual value, and require new evidence for the next check.

Ranked sources and reasons

1. Wharton, Building a Vehicle — best direct formation, vehicle-fit, founder-filter, and 1996 chronology source.
2. Globes 2007 — doctrine confronted by adverse performance and closure questions.
3. Taboola retrospective — richest named relationship/team narrative, with hindsight and date limits.
4. SEC/issuer controls in investments — strongest ownership, board, shared-power, retained-share, and partial-realization boundaries.
5. IsraelTech and Calcalist 2026 — strongest late direct chronology/debriefing and specialist-mandate accounts, with caption/ASR controls.
6. GoRefer and Pitango III adverse record — necessary antidote to winner-only doctrine.

Pareto 80/20 — what drove the record

Public evidence does not permit a literal 80/20 return calculation. No complete Peres-attributed investment denominator, audited fund cash-flow schedule, cost basis, dilution history, carry allocation, or career DPI/TVPI/IRR series was found. The tables therefore use Pareto as a prioritization lens and keep three questions separate: which cases have the strongest economic proxy, which changed the institution, and which teach the most. [researcher inference · strong] source-map coverage boundary

Directional economic-contribution ranking

Case paths: Taboola → Lab 3 · Via → Lab 4 · AudioCodes → Lab 2 · Webbing → Lab 7 · Shany → Lab 1

Dominant-outlier test. Removing Taboola—the only case with a disclosed partial realization—does not reveal a rankable remainder. Via supplies a residual share count, Webbing a company transaction price, AudioCodes a press description, and Shany an acquisition; none has the cash-flow series needed to compare contribution. The economically honest result is therefore not computable, not “diversified success.” [researcher inference · strong]

Historical-importance ranking

Case paths: Shany → Lab 1 · VocalTec → scoped ledger · AudioCodes → Lab 2 · Taboola → Lab 3 · Via → Lab 4

Decision-learning-value ranking

Case paths: Pitango III → Lab 6 · GoRefer → Lab 5 · Shopserve → failure ledger · Taboola → Lab 3 · Shany → Lab 1 · Via → Lab 4 · AudioCodes → Lab 2

Evidence boundary and searched denominator

Guiding questions and investor classification

This phase asked: (1) what can be ranked economically without invented returns; (2) which companies are genuinely linked to Peres rather than merely to Pitango; (3) how Mofet, Polaris II, later Pitango funds, and board roles changed incentives; (4) who else deserves decision and execution credit; (5) what was knowable at each material decision; (6) which failures survive in the public record; (7) what happens when Taboola and Via are removed; and (8) what missing evidence could overturn the assessment.

Peres is a modern, low-transparency, multi-partner VC operating through a public venture company, conventional limited partnerships, specialist vehicles, boards, and impact/regional platforms. He is not a 13F-style public-markets investor. Controversy searches found firm-governance and conflict questions but no sourced allegation of investment misconduct by Peres; those findings are handled separately below.

Search frame and deduplication. The review spans 1992–2026 and searched the live Pitango directory, firm/team pages, SEC EDGAR ownership/director filings, contemporaneous Globes and Israeli press, issuer/company releases, archived direct interviews, and the existing note/source map. Entity and name variants included Chemi/Hemi/Nehemia Peres, Mofet, Polaris, Dash/DS Polaris, Pitango, and fund-specific legal entities surfaced in filings. Companies—not financing rounds—are the unit: follow-ons, continuation transfers, syndicate entities, and duplicate cards remain within one company row. A case enters Research set B only with a public deal/vehicle link stronger than a platform card or shared-principal footnote; board biography alone is a retrieval lead, not investment proof.

Two denominators, kept separate

Universe A — current Pitango platform snapshot. On 2026-08-22, the live portfolio HTML exposed 140 unique company cards after deduplicating by data-company-id/slug: 69 with no displayed outcome status, 51 marked acquired, and 20 marked IPO. The extraction is reproducible from the page source by enumerating company-card attributes and displayed status. It is a current marketing directory, not a historical fund ledger; “acquired” and “IPO” are company states, not realized returns. Pitango portfolio directory

Pitango's May 2026 boilerplate separately claims 13 funds, more than $3 billion under management, more than 300 companies, and 91 exits/IPOs. It is a firm self-report with no named-company or cash-flow reconciliation. Even the minimum reading of “300+” exceeds the live directory by at least 160 companies; the difference may reflect removed historical companies, follow-ons, vehicle definitions, or marketing conventions and cannot be allocated from public evidence. Pitango/Tomorrow.io release, “About Pitango”

Research set B — finite hand-reconstructed core cases, not a universe. The source map and direct/contemporaneous accounts seeded 12 possible deal-linked cases for deeper reconstruction: Shany Computers, VocalTec, Orckit, NiceCom, AudioCodes, BackWeb, MystiCom, Shopserve, GoRefer, Taboola, Via, and Webbing. Only VocalTec, AudioCodes, BackWeb, Taboola, and Via appear in Universe A. Seven researched names being absent from the live page demonstrates a history/survivorship gap, but the 12 are not an exhaustive career list: an issuer biography separately names other current and former Peres boards—including Venn, Duda, Totango, Radwin, and Anywhere—that were not promoted into the deal set without transaction or sponsorship evidence. Globes 2001, early cases and failures · Israel Amplify Program Corp. S-1, Peres biography

Filings for Riskified, Radware, Retalix, Varonis, TScan, Rocket, Cerus, TopSpin, Voltaire, and Virtual NetCom were used as vehicle-control evidence where they show Pitango entities or shared principal power. They were not promoted into Universe B because those facts alone do not prove a Peres-specific sourcing, sponsorship, or board role. This conservative exclusion prevents a GP signature or shared voting footnote from becoming a personal deal claim. source-map filing routes

Mechanical reconciliation

Coverage fractions. The live-page extraction is 140/140 cards and the finite research set is 12/12 ledgered: six sufficiently reconstructed and six unresolved. This is research-set completeness, not deal completeness. Peres career coverage fraction: not computable. The denominator of unobserved Mofet holdings, exited/removed portfolio companies, other board/deal roles, passed deals, write-offs, syndicate investments, and later specialist vehicles is unknown. Pitango's self-reported totals—120 companies in 2007, close to 300 in 2022, and 300+ in 2026—use different dates and scopes and cannot supply the missing Peres-specific term. Globes 2007, “120 companies” answer · Mizuho note, 01:53–03:37 · 2026 firm release

Selection bias runs in both directions: current portfolio pages preserve visible exits and active companies while removing some failures; public profiles overselect named winners; SEC filings overselect companies that reached public markets; and direct retrospectives overselect memorable decisions. [researcher inference · strong]

Historical market, regulatory, and base-rate context

No sourced contemporary portfolio base rate permits comparing Peres with all Israeli venture investors across vintages. The only defensible local base rates are Pitango's own 2007 funnel and the reporter's interim Fund III company counts; later outcome knowledge is not inserted into earlier decision packets.

Pitango III mark trajectory — a measurement warning, not a return series

The surviving LP-related reports change date, definition, and measurement basis. They are useful as adverse signals, but they cannot be chained into a final performance series:

Conclusion. These observations establish an adverse and unstable mark trajectory, not final economics. Final net DPI, TVPI, IRR, fees, distributions, valuation policy, and cash-flow dates remain unknown; neither the favorable 2003 point nor the later negative points should be selected as “the return.” [researcher inference · strong]

Vehicle economics and incentives

Relative to a current LP-backed fund, the supported similarities are finite committed capital, annual management fees, profit participation, multi-partner governance, reserves/follow-ons, and exit-dependent liquidity. The differences are historical: Mofet's listed-company form, Yozma/Inbal market creation, Israeli export dependence, and later multiple specialist/continuation entities. Exact modern fee/carry comparisons are impossible without partnership agreements.

Fund duration, reserve policy, succession/key-person triggers, removal rights, extension terms, and end-of-life liquidation mechanics were not found for these vehicles; none should be inferred from current-market convention.

People, capital, and influence map

The 2018 allegations against former partner Rami Beracha concern Beracha, not Peres. Pitango said it had received no complaints and stated a zero-tolerance posture while Beracha took leave. This is a firm-governance stress test, not evidence of a documented portfolio-process change and not an allegation against Peres. Times of Israel 2018, report and firm response

Comprehensive scoped research ledger

Unknown is substantive here: it prevents board biography and platform marketing from becoming invented deal history. The 12 rows are complete for the stated research set, not for Peres's career.

No-hindsight decision labs

Each exercise stops at the decision date. Unknown terms, dissent, reserves, and governance are not silently filled with later outcomes.

Lab 1 — Shany, 1993: a lean exporter before the institutional ecosystem

Freeze date: 1993. Formation packet available then. Shany had bootstrapped, opened a US subsidiary in 1991, and sought its first institutional financing. Mofet and Dash Polaris were separate vehicles led by Peres and Kalish. Shany note

• Known: operating frugality, founder commitment, early US-market presence, two emerging local venture institutions.
• Unknown/disputed: check split, valuation, security, ownership, customer concentration, revenue, product diligence, board rights, dissent, milestones, reserves, and exit alternatives.
• Alternatives: keep bootstrapping, take one vehicle rather than two, seek strategic capital, or defer expansion.
• Provisional underwriting — researcher exercise: invest only if US reference customers and product economics validate the export wedge; stage capital against repeatable revenue and preserve founder control subject to basic governance. [researcher inference · provisional]

STOP BEFORE REVEAL — analyst worksheet

1. Would you invest then? Only conditionally; the US subsidiary is evidence of action, not product-market fit.
2. What would kill the deal? Unverifiable customer demand, no defendable product, or a round sized beyond evidenced milestones.
3. What ownership/terms fit the risk? Unknown historically; propose milestone-based preferred equity with pro-rata and information rights, not a fabricated target percentage.
4. What evidence unlocks the next check? Paid US references, retention, gross margin, implementation time, and founder-team capacity.

Reveal — decision and outcome. Mofet and Dash Polaris invested; Intel acquired Shany in October 1994. The transaction price and investor proceeds are not public. Peres/Touboul retrospective

Attribution and lesson. The supported lesson is cross-vehicle cooperation around a lean exporter, not “Peres picked an Intel acquisition.” Founders, both vehicles, and later Intel deserve separate credit. [researcher inference · strong]

Lab 2 — AudioCodes, late 1990s: communications-wave winner with disputed credit

Freeze date: February 1999, before the IPO. AudioCodes operated in voice-over-IP/network communications during an Israeli communications-company and Nasdaq expansion. Contemporaneous press reported that the Polaris/Chase/Genesis group held 14%; the issuer later recorded a 1997 DS Polaris/Chase/Genesis investment agreement and Eyal Kishon's board service from 1997. The original memo, group allocation, exact check, and a Peres-authored thesis were not found. Globes February 1999 · AudioCodes 2003 Form 20-F

• Known: communications technology fit the vehicle's historical sector; a syndicate rather than Peres alone held the reported stake; Eyal Kishon was the vehicle-linked director.
• Unknown/disputed: who sourced and sponsored, founder/reference diligence, check, valuation, security, ownership, reserves, board interventions, dissent, and IPO sell/hold policy.
• Alternatives: other communications vendors, wait for standards adoption, syndicate to reduce technical/market risk, or pass on an uncertain VoIP adoption curve.
• Provisional underwriting — researcher exercise: require interoperability proof, carrier/OEM references, unit economics, a standards map, and burn sufficient to survive platform shifts. [researcher inference · provisional]

STOP BEFORE REVEAL — analyst worksheet

1. Would you invest then? Only after independent technical and distribution validation; sector excitement is not diligence.
2. What would kill the deal? Proprietary lock-in without channel support, fragile interoperability, or financing needs beyond the vehicle's reserves.
3. What ownership/terms fit the risk? Not publicly reconstructable; demand pro-rata and board-information rights rather than inventing the historical stake.
4. What evidence unlocks the next check? Product certification, repeat customers, partner commitments, gross-margin durability, and burn-to-milestone.

Reveal — decision and outcome. AudioCodes became public. In May 1999, Globes valued the Polaris fund stake at $16.5 million; this was a post-IPO position mark, not realized cash or fund return. A 2001 profile later called AudioCodes a prominent Pitango II exit and associated Peres with it, but the contemporaneous issuer/syndicate record prevents sole credit. Globes May 1999 · Globes 2001

Attribution and lesson. It is directionally important economically but unusable as a personal return without the cap table and cash flows. The case teaches credit discipline more strongly than it teaches the original picking rule. [researcher inference · strong]

Lab 3 — Taboola, February 2013: invest after the model starts to resolve

Freeze date: 2013 Series D. Taboola reported 7.5x revenue growth in 2012, a mobile launch, new executives, and publishers including Bloomberg, The Weather Channel, and BusinessWeek. A $10 million Marker-led round had occurred six months earlier; Pitango led the new $15 million round. Taboola release

• Known: rapid self-reported growth, named customers, product expansion from video toward broader content recommendation, new operating leadership, existing investors, and an international-expansion use of proceeds.
• Unknown/disputed: audited cohort economics, customer concentration, algorithmic edge, brand-safety risk, valuation, security, Pitango ownership, reserves, IC dissent, and whether the decisive third meeting occurred in 2011, 2012, or another date.
• Alternatives: wait for audited retention, invest in another ad-tech platform, syndicate a smaller check, or pass because platform dependence and content quality could cap durability.
• Provisional underwriting — researcher exercise: convert “7.5x” into cohort/customer math; verify publisher renewal, revenue quality, recommendation lift, mobile economics, data rights, and leadership references. [researcher inference · moderate]

STOP BEFORE REVEAL — analyst worksheet

1. Would you invest then? Yes only if growth is retained and customer economics survive independent verification.
2. What would kill the deal? Concentrated publishers, low-quality traffic, weak incremental advertiser ROI, or model risk hidden by headline growth.
3. What ownership/terms fit the risk? Historical valuation is unknown; seek pro-rata, information, governance, and downside terms proportionate to ad-tech/platform risk.
4. What evidence unlocks the next check? Audited revenue bridge, cohort retention, publisher concentration, advertiser repeat spend, mobile margin, and reference calls.

Reveal — decision and outcome. Pitango led the Series D; Peres joined the board in 2013. Taboola became public in 2021. In November 2025, Taboola paid about $23.4 million for 6,387,648 shares held by certain Pitango-affiliated funds. 2025 proxy · 2025 8-K

Attribution and lesson. Peres's retrospective names Zvi Limon, other investors, and operating leaders. The supported mechanism is repeated contact plus independent validation and team formation; the later listing does not prove culture caused the return. [retrospective witness · moderate] Peres essay

Lab 4 — Via, April 2015: dense operating evidence, expansion risk

Freeze date: Series B. Via had operated one Manhattan corridor since September 2013; it reported more than 300,000 cumulative shared rides, tens of thousands weekly, 40,000 members, and over 500 licensed vehicles. Pitango led a $27 million round with Hearst, Ervington, and returning investor 83North; proceeds were for market expansion, product, and hiring. 2015 round report

• Known: real usage, a constrained operating geography, two years of operating history, licensed supply, a defined pooling product, named co-investors, and explicit use of proceeds.
• Unknown/disputed: retention, contribution margin, utilization by hour, incentives, regulatory exposure, valuation, security, Pitango ownership, IC dissent, and whether corridor density travels to new cities.
• Alternatives: deepen Manhattan before expansion, license technology to transit operators, stage the round by city milestones, or pass because local density may not generalize.
• Provisional underwriting — researcher exercise: fund one or two city replications with density, wait time, pooled occupancy, contribution margin, regulatory approval, and repeat-use gates. [researcher inference · moderate]

STOP BEFORE REVEAL — analyst worksheet

1. Would you invest then? Yes conditionally; the usage packet is unusually tangible, but one corridor may be a local optimum.
2. What would kill the deal? Negative mature-city contribution margin, subsidy dependence, regulatory prohibition, or declining service quality as geography expands.
3. What ownership/terms fit the risk? Historical terms are unknown; use milestone-linked reserves, pro-rata, information, and governance rights suited to capital-intensive rollout.
4. What evidence unlocks the next check? Replicated density, rider retention, driver utilization, city-level economics, regulator support, and expansion payback.

Reveal — decision and outcome. Pitango led the round and Peres joined the board as a Pitango-designated director. Via later became public. Its final 2025 prospectus attributed 5,520,762 pre-offering shares to Pitango-affiliated entities; 388,645 were sold in the offering and 5,132,117 / 6.8% remained. Via 424B4, directors and principal-stockholder table

Attribution and lesson. Founders Ramot and Shoval, prior investor 83North, co-investors, multiple Pitango vehicles, and Peres's board role are distinct inputs. A later Japanese market-introduction anecdote may show LP value-add, but one selected case cannot prove systematic causality. [researcher inference · moderate] Mizuho note, 30:07–30:51

Lab 5 — GoRefer, 2001–2002: failure and stop decision

GoRefer — preserve cash after customers disappear

Freeze date: financing before shutdown. GoRefer aimed to link buyers and sellers. A 2001 Hebrew Globes postmortem reported Pitango III and Apax Israel II in a $10 million round; its timing references are not fully consistent, and check split, security, ownership, customer data, and reserve commitment are unresolved. The accessible formation record is press, not an IC memo. Globes Hebrew 2001

• Known: an internet intermediation model, a substantial syndicate round, and later evidence that customers disappeared.
• Unknown/disputed: unit economics, customer concentration, product differentiation, burn, milestones, board votes, and how much was initially drawn.
• Alternatives: stage a smaller seed, pivot to a surviving customer segment, return capital immediately, or continue only with signed demand.
• Provisional underwriting — researcher exercise: fund a narrow milestone tranche contingent on repeat transactions, cost of acquisition, gross margin, and customer survival. [researcher inference · provisional]

STOP BEFORE REVEAL — analyst worksheet

1. Would you invest then? Not at full-round scale without transaction and retention evidence; a staged experiment might qualify.
2. What would kill the deal? Disappearing customers, no repeat usage, negative contribution after acquisition cost, or a market that collapses with financing.
3. What ownership/terms fit the risk? Unknown historically; tranche capital and liquidation protection matter more than a fabricated ownership target.
4. What evidence unlocks the next check? Paid repeat transactions, surviving reference customers, burn control, and a credible path to non-venture-funded demand.

Reveal — decision and outcome. GoRefer shut down. In a 2002 interview, Peres said customers disappeared and that founders and co-investors chose closure while cash remained; he presented small initial/follow-on funding as limiting cash loss. The Hebrew account said roughly one-third of the $10 million round was returned. No LP distribution or final fund loss was found. Globes 2002 · Globes Hebrew 2001

Lesson. The evidence supports a bounded governance behavior—stop with cash remaining—not a portfolio-wide claim that Peres consistently cut losses early. No later public process change is documented. [contemporaneous record · moderate]

Lab 6 — Pitango III, 2000: portfolio construction at the top of a cycle

Freeze date: 2000 fund construction. Pitango III was reported at $500 million after the Polaris and Eucalyptus teams combined. Its contemporaneous plan targeted 30–35 companies, $6–12 million of lifetime capital per company, initial-stage entry, meaningful ownership, and business models able to survive the bubble. General partner Yoram Mankovski—not Peres—gave the 20% return target and follow-on logic. Globes 2000

• Known: large committed capital, a combined multi-partner team, a finite company-count plan, lifetime-capital bands, early-stage entry, and a stated significant-ownership intent.
• Unknown/disputed: actual check pacing, reserve ratio, the numerical ownership target, concentration limits, loss budget, attribution of the full strategy beyond Mankovski's quoted elements, and dissent.
• Alternatives: smaller fund, slower deployment, narrower sector focus, more reserves, or a barbell of early and later stages.
• Provisional underwriting — researcher exercise: pre-register vintage pacing, exposure limits, reserve coverage, ownership thresholds, write-down rules, and a scenario where public markets close. [researcher inference · strong]

STOP BEFORE REVEAL — analyst worksheet

1. Would you invest then? Only with explicit pacing and downside scenarios; brand and prior exits do not substitute for fund construction.
2. What would kill the deal? For the fund commitment: no reserve model, deployment pressure from size, valuation indiscipline, or weak partner capacity per company.
3. What ownership/terms fit the risk? At fund level: adequate GP commitment, key-person/LP governance, transparent valuation policy, and alignment; exact historical terms are unknown.
4. What evidence unlocks the next check? Deployment by quarter, initial/follow-on split, sector/stage exposures, loss ratios, partner load, and cash distributions.

Reveal — later evidence. The strategy expanded beyond the 30–35-company plan: a 2007 article cited 56 companies, eight closures, and 18 exits from IVC; Kalish said the figures were not far from reality. The press-reported marks moved from positive 20.6% in a 2003 analysis to negative 18.25% at end-2005, negative 11% / 0.6x in Q2 2007, and negative 3% in a 2011 report, with different definitions and Peres disputing the 2007 information. These are point-in-time press interpretations, not final audited returns. mark-trajectory table

Lesson. Large diversified funds can create support capacity and simultaneously magnify vintage, pacing, and mark risk. Without final DPI/TVPI and an outlier-exclusion series, neither the favorable nor skeptical interpretation wins. [researcher inference · strong]

Lab 7 — Webbing, before 2021: governance through a failed IPO path

Freeze date: before the abandoned 2021 listing. Webbing, founded in 2010 by Noam Lando and Yaniv Elron, built global cellular/eSIM connectivity for enterprise mobility, automotive, and logistics. CTech later identified Pitango as primary investor and Peres as board chair. CTech 2023

• Known: technical/product category, global enterprise use cases, founder continuity, Pitango exposure, and Peres governance role.
• Unknown/disputed: entry date, check, security, ownership, revenue, margins, customer concentration, board interventions, IPO valuation, and dissent.
• Alternatives: continue private growth, sell strategically, pursue IPO, or recapitalize with a later-stage investor.
• Provisional underwriting — researcher exercise: verify carrier dependencies, roaming regulation, customer concentration, recurring gross margin, standards adoption, and strategic-buyer alternatives. [researcher inference · provisional]

STOP BEFORE REVEAL — analyst worksheet

1. Would you invest then? Only after contract, carrier, and regulatory diligence; the category's global reach does not eliminate concentration risk.
2. What would kill the deal? One-carrier dependency, adverse roaming rules, poor recurring margins, or an IPO valuation unsupported by cash generation.
3. What ownership/terms fit the risk? Unknown historically; require information, pro-rata, board, and exit-protection rights proportionate to infrastructure dependence.
4. What evidence unlocks the next check? Multi-carrier resilience, customer retention, recurring margin, regulatory approvals, and cash conversion.

Reveal — decision and outcome. Webbing abandoned a 2021 IPO over valuation discrepancies and was acquired by Wireless Logic in 2023 in a transaction reported above $200 million. Seller proceeds and Pitango return were not disclosed. CTech 2023

Lesson. Rejecting a public valuation and later taking a strategic exit can be disciplined governance, but the record does not reveal who made the decision or whether waiting improved fund economics. [researcher inference · moderate]

Portfolio-level forensics

What the observable portfolio supports

• Stage: Peres publicly endorsed balance across stages; later platform design separates First, Growth, and HealthTech. The 12-company set contains early co-investments, later growth rounds, and board/governance roles, but is too selected to estimate allocation. [investor-stated · strong] Globes 2007 · Pitango team
• Sector: the linked set is heavy in communications, internet/software, ad-tech, mobility, IoT, and video. That may reflect Peres's disclosed roles and public-company survival rather than the actual portfolio mix. [researcher inference · strong]
• Geography: Israeli technical roots plus US/global commercialization recur in Shany, the Palo Alto experiment, Taboola, Via, and Webbing. The record does not calculate performance by geography. [documented behavior · moderate]
• Check size and ownership: the Taboola $15 million and Via $27 million figures are total rounds, not Pitango checks. Fund III reporting supplies an interim company count but not a complete check ledger. No personal ownership is inferred. Taboola 2013 release · Via 2015 round
• Concentration: no supported company-by-company cost or proceeds schedule exists. A public residual stake can be large without having been the largest return; no literal power-law share is calculated.
• Reserves/follow-ons: multiple fund and continuation entities appear in Via; Taboola involved Fund VI entities. Exact reserve rules, cross-fund allocation, and transfers remain unknown. Via 424B4, n.12 · Taboola 13G/A
• Governance: board roles are best documented at Taboola, Via, and Webbing. No public record supplies board attendance, decisive interventions, founder replacement, or a controlled with/without-board outcome comparison.
• Exits and misses: five of the 12 scoped companies appear on the current page; that page marks VocalTec, AudioCodes, BackWeb, Taboola, and Via as IPOs. Historical omissions include two named failures, proving current outcome status is not a denominator. Pitango portfolio directory, 2026-08-22 snapshot · Globes 2001, failures

Ex-Taboola and ex-Via sensitivity

Removing Taboola and Via eliminates the two richest public cap-table cases and the only disclosed partial cash realization. What remains is AudioCodes' press-attributed prominent exit, Webbing's company transaction price, Shany's acquisition, several role-only names, and two failures. The thesis therefore weakens sharply from “documented long-duration economic record” to “historically important institution builder with selected company outcomes.” [researcher inference · strong]

Removing both does not erase the earlier institutional contribution—Mofet, the 1996 Polaris II combination, Shany/VocalTec cooperation, and AudioCodes—but it makes portfolio-level return claims untenable. That sensitivity is itself evidence against attributing platform greatness to a broad, measured Peres-only deal record. [researcher inference · strong] Globes 1997, Polaris II · Shany note · Globes 2001, AudioCodes

Team, role, privilege, and era sensitivity

Failure, miss, and anti-portfolio postmortems

Peres's 2025 direct voice says candid debriefing should expose and prevent repeated mistakes, but the public talks name no investment failure and show no case-linked before/after procedure. [investor-stated · strong; implementation evidence insufficient] IsraelTech note, 05:51–11:12

Terminal uncertainty and retrieval ledger

No company, partner, LP, founder, regulator, or archive was contacted. The record is strong enough to reconstruct seven bounded exercises and the institutional sequence—Mofet in 1992; separate Polaris I/Yozma in 1993; Peres joining/co-building Polaris II/Pitango in 1996—but not to certify a complete or audited personal track record. [researcher inference · strong] Wharton, Mofet chronology · Globes 2000, Polaris I · Globes 1997, Polaris II

Coverage and evidence base

The priority written corpus contains seven items: seven processed, seven accessible in full, zero partial, zero inaccessible, and seven deeply analyzed. Every ★★★ and ★★ written item in the source map has an exact-schema note and its own treatment below. [researcher inference · strong]

The corpus combines one decision retrospective (Taboola), one first-deal recollection (Shany), and five 2024–25 thesis essays on foresight, product-market fit, dual use, company scale, and government customers. It is direct voice but not a representative investment ledger: six items are hindsight or market doctrine, and none publishes rejected deals, fund cash flows, or a loss cohort. [researcher inference · strong] Taboola note · Shany note

Literal priority-source reconciliation

Mirrors, newsletter links, and the LinkedIn version of the Taboola essay are not counted as independent works. The CTech versions control authorship for Calibrate Bigger and B2G; the Pitango pages are publication mirrors, not an extra denominator. [contemporaneous record · strong] CTech Calibrate byline · CTech B2G byline

Source-by-source analysis

1. Three lessons I learned from Taboola's IPO — patient contact, team complementarity, and hindsight

Context, argument, and reasoning. Written when Taboola listed in 2021, the essay reconstructs three meetings: a large but unfinished “reverse search” idea, later business-model refinement, and a third meeting prompted by Zvi Limon. Peres's explicit lessons are dream scale, leadership, and people-first culture. The reasoning is that vision attracts attention, iteration makes it underwritable, a complementary leadership team converts it into execution, and transparent culture maintains coordination at scale. [investor-stated · strong] CTech, opening and lessons 1–3

Access boundary. The complete CTech body and byline were accessible; the piece remains an IPO-date retrospective rather than a contemporaneous IC record. CTech, byline and opening

Evidence, assumptions, and decision implications. The named team and repeated meetings are useful behavioral evidence, but the successful listing is the essay's selection condition. It assumes leadership and culture helped cause the outcome; it does not compare failed culturally strong companies or disclose valuation, ownership, decision rights, reserves, or proceeds. The disciplined rule is therefore narrower than “back dreamers”: maintain contact with a promising but premature founder, identify the evidence that must improve, and record who supplied the decisive independent reference. [researcher inference · moderate] note, takeaways and limits

Biography, evolution, and contradiction. The story shows Peres operating through a network rather than as a lone picker; he credits Limon, other investors/directors, and three operating leaders. Yet the chronology says his journey began in 2009, Pitango entered after the third meeting, and he joined in 2012. Until filing and IC records reconcile those dates, the piece supports a relationship arc but not a precise investment date. [retrospective witness · moderate] note, biography evidence

2. Looking Ahead to 2025 — foresight as signal discrimination

Context, argument, and reasoning. After the 2024 war year, Peres argues that venture timing requires noticing authentic future signals before consensus. He identifies trillion-dollar “giraffes,” AI, dual use, and Israel's movement from R&D toward scaling as linked changes. His causal chain is: new technological and geopolitical conditions change company ceilings; higher ceilings change talent, customer-adoption, capital, and time requirements; investors should update their reference set. [investor-stated · strong] Pitango, opening through dual-use sections

Access boundary. The complete signed Pitango essay was accessible; it offers observations and forecasts, not a tested portfolio cohort. Pitango, opening and “Where Do We Go From Here?”

Evidence, assumptions, and decision implications. The essay uses the Magnificent Seven, ChatGPT, drones, and Israeli sector testbeds as signals. These are selected examples, not a tested forecast or portfolio cohort. A usable rule is to specify the observation that would distinguish structural change from narrative noise before increasing market-size, duration, or valuation assumptions. [researcher inference · moderate] note, counterevidence

Evolution and contradiction. The aspiration shifts beyond the early Israeli acquisition model toward enduring global enterprises. But the essay's warning about noise sits uneasily beside an undefined animal category and no falsification threshold. A “giraffe” can be a useful ambition prompt only if the IC separately tests base rates, capital intensity, governance, and downside. [researcher inference · strong] Pitango, “The Rise of New Tech-Driven Empires” and “Where Do We Go From Here?”

3. Israel's Emerging Role in Global Product-Market Fit — separate innovation density from feedback scale

Context, argument, and reasoning. Peres defines product-market fit as discovery about users, needs, use models, and cost-for-value balance. His X/Y frame separates market scale and rapid feedback (X) from innovation density (Y); digital distribution and problem-specific testbeds can allow a small country to supply selected X-axis functions without becoming a large market. [investor-stated · strong] Pitango, opening and “Changing Global Footprint”

Access boundary. The complete signed Pitango essay was accessible; its X/Y framework is qualitative and publishes neither scoring thresholds nor a failure denominator. Pitango, “Market Fit is on the X-Axis” and “Innovation is on the Y-Axis”

Evidence, assumptions, and decision implications. Water, agriculture, cyber, defense, cloud distribution, and local adoption are offered as analogies. The unstated assumption is transferability: local validation must predict demand under foreign regulation, procurement, language, price, and distribution. The decision rule is to name which product-market-fit uncertainty a test geography actually resolves, then pre-register the evidence required before extrapolating. [researcher inference · strong] note, philosophy and limits

Biography and evolution. The framework reflects a participant who began when Israeli companies often sold technology outward and now emphasizes customer adoption and scale-up capability. That chronology is suggestive, not proof that experience caused the framework or that Pitango consistently used it. [researcher inference · moderate] PMF note, biography and access limits

4. A Two-Way Street — defense urgency and civilian economics must both work

Context, argument, and reasoning. Peres argues that dual use can combine civilian entrepreneurial speed and cost pressure with defense budgets, mission urgency, and field validation. Israel's dense movement among military units, defense companies, founders, and reservists can accelerate knowledge transfer; he extends the observation to wartime trauma and healthcare. [investor-stated · strong] Pitango, opening and “Investing in dual-use technologies”

Access boundary. The complete signed Pitango essay was accessible; it presents sector doctrine and selected examples, not a Pitango deal or return cohort. Pitango, “Challenges, limitations and opportunities”

Evidence, assumptions, and decision implications. Drones, autonomous systems, sensors, space, healthcare, and SpaceX illustrate possible transfer. SpaceX is an external example, not Peres's investment evidence. The essential diligence rule is two independent underwritings: defense mission, buyer, certification, and procurement; then civilian user, channel, liability, price, and unit economics. A larger combined TAM is not evidence that one product fits both. [researcher inference · strong] note, limits

Contradiction and historical boundary. Battlefield adoption can be strong technical evidence but weak evidence of civilian willingness to pay or ethical acceptability. Network density can improve access while increasing homophily and attribution ambiguity. The piece offers opportunity doctrine, not export-control policy, kill criteria, or a failed-deal denominator. [researcher inference · strong] dual-use note, counterevidence and access limits

5. Calibrate Bigger — update references without abandoning downside discipline

Authorship and context. The canonical CTech publication explicitly names Chemi Peres and contains the full essay. A full Pitango mirror also names Chemi, while a separate metadata-only duplicate shows a generic Pitango byline. These are one underlying work, and the generic duplicate is not used to override the two complete bylines. [contemporaneous record · strong]

Access boundary. The CTech body and June 4 Pitango mirror were accessible in full; the metadata-only Pitango duplicate contributes no content evidence. CTech, byline and full body

Argument and reasoning. Peres presents successive calibrations: 1990s IP/OEM exits, internet IPO ambition, post-crash efficiency, low-rate unicorn valuation, and post-Covid measurable durability. Comparable outcomes, macro conditions, technology, and market structure continually reset expectations. “Giraffes” represent companies with long horizons and influence across markets and governments. [investor-stated · strong] CTech, “How we got here” through “The empire builders”

Evidence, assumptions, and decision implications. The historical sketch is participant interpretation written in 2025, not contemporaneous proof of each regime. The useful rule is dynamic benchmarking: record the comparable set, why it applies, and what change would remove it. The dangerous version is upward-only calibration—using unprecedented upside to excuse price, duration, governance, or capital risk. [researcher inference · strong] note, contradictions

Evolution and contradiction. The essay explicitly criticizes economically unjustified unicorn valuations, yet advocates much larger ambitions without publishing valuation discipline. That tension should remain visible: bigger outcomes may justify greater duration and resources only after evidence, not as an input used to manufacture the valuation. [researcher inference · strong] CTech, “How we got here” and “The empire builders”

6. From Red Tape to Revenue — government as customer, platform, and concentration risk

Context, argument, and reasoning. Peres argues that digital procurement, public modernization, global challenges, sustainability mandates, and strategic infrastructure competition are lowering old B2G barriers. Governments can offer scale and contractual longevity; companies can supply innovation and become actors in national resilience and diplomacy. [investor-stated · strong] CTech, “The world beyond business” and five drivers

Access boundary. The canonical CTech essay was accessible in full; its author explicitly says the statistics are limited, so the examples cannot establish a base rate. A July 30 Pitango mirror adds a firm-portfolio subsection absent from CTech; that later insertion cannot establish Peres's authorship, individual deal role, or economics. CTech, “The fall of barriers to entry in B2G”

Evidence, assumptions, and decision implications. Estonia, Singapore, UK healthcare, PsiQuantum, Starlink, irrigation, and public-private partnerships illustrate the thesis, but the article itself says statistics are limited and leans on qualitative cases. A defensible IC rule is to map statutory buyer, appropriated budget, tender path, implementation authority, audit/ethics constraints, renewal exposure, working capital, and election sensitivity before treating contract size as quality. [researcher inference · strong] note, takeaways and philosophy

Biography, evolution, and contradiction. The essay brings investment and Peres Center public-purpose language closer together, but no evidence shows that peace or diplomacy goals caused a return. Government stickiness can mean durable revenue or dangerous concentration and capture; public influence raises governance duties rather than automatically adding moat. [researcher inference · strong] B2G note, biography and counterevidence

7. Shany first-investment reflection — preserve the two vehicles and two narrators

Context and argument. Peres retrospectively calls Shany his first venture investment, first fund deal, first co-investment with Rami Kalish, and first M&A exit. Founder Shlomo Touboul's embedded account says Shany bootstrapped, opened a US subsidiary in 1991, received 1993 investment from Mofet and Dash Polaris, and was acquired by Intel in October 1994. [retrospective witness · moderate] LinkedIn post, Peres and Touboul paragraphs

Access boundary. The complete public post and embedded founder text were accessible; both are 2025 recollections and neither supplies underlying transaction documents or economics. LinkedIn post, Peres and Touboul paragraphs

Attribution and chronology control. The source supports Peres/Mofet and Kalish/Dash Polaris as co-investors. It does not place Peres in Kalish's 1993 Yozma/Polaris I formation; Peres's Polaris II/Pitango institutional entry belongs in 1996. “Partner” in the 2025 post cannot silently replace the legal vehicle boundary. [researcher inference · strong] note, biography evidence

Evidence, assumptions, and implications. Touboul supplies an ex-ante company sketch—bootstrapped, already US-present, and challenging the belief that Israeli founders could only supply technology. That may explain appeal, but it is not a recovered memo. The post gives no check size, ownership, board rights, proceeds, or counterfactual to selling. Its “35 years” conflicts with the 1993 date and is treated as shorthand, not exact chronology. [retrospective witness · moderate] LinkedIn post, Peres and Touboul paragraphs

Cross-source synthesis

The recurring operating sequence

Across the corpus, a consistent sequence emerges: notice a structural change early; pair ambitious vision with concrete evidence; learn through repeated customer contact; use dense networks for references and validation; update calibration as macro and technology change; and build for institutions and markets beyond Israel. This is a reconstruction, not a published six-step Peres method. [researcher inference · moderate] Looking Ahead note · PMF note · Taboola note

What appears foundational versus occasional

• Foundational: people plus iterative clarity. Taboola's repeated meetings and named leadership complementarity are the corpus's most deal-specific evidence. [documented behavior · moderate] Taboola note, takeaways and biography
• Foundational: Israel as a launch point, not the final market. Shany's US subsidiary, the X/Y frame, and repeated global-scaling concern span early and current eras. [researcher inference · moderate] Shany note · PMF note
• Foundational: recalibration. Peres treats markets, technology, and company ceilings as nonstationary; the investor must update references. [investor-stated · strong] Calibrate Bigger, “How we got here”
• Newer and less tested: giraffes, dual use, and B2G. These dominate 2024–25 writing but lack a public Pitango outcome cohort, economics, or failure denominator. [researcher inference · strong] Looking Ahead note · dual-use note · B2G note

Rhetoric versus demonstrated decisions

Shany and Taboola show long relationship-building, co-investment, named collaborators, and successful exits/listings. The 2024–25 essays argue for company scale, dual use, and government customers, but do not reconstruct investments selected under those ideas. Repetition shows current conviction; it does not show efficacy. [researcher inference · strong] Shany note · Taboola note

Intellectual chronology, evolution, and contradictions

The stable thread is not a single sector but adaptation: maintain ambitious direction while changing the market, product, and institutional reference set. The tension is that the public writing is far richer on upside imagination than on valuation, ownership, reserves, loss control, or stopping. [researcher inference · strong] Calibrate note · Taboola note · B2G note

Decision playbook and modern translation

Rule 1 — turn “not yet” into an evidence contract

Historical rule. When vision is compelling but the model is unclear, specify what must change and keep the relationship alive; do not confuse continued contact with commitment. [researcher inference · moderate] Taboola note

Diagnostic questions. What was unclear? Who will produce the missing evidence? What deadline or market event changes the decision? Which independent reference can challenge the founder story?

Modern VC translation (as of 2026-08-22). In an analyst-created AI-native hypothetical, use the rule when product capability changes faster than buyer behavior; pre-register improving retention, buyer clarity, cost-to-serve, and reference quality. It fails when repeated meetings merely create familiarity bias. A likely misuse is retroactively calling any long relationship “patient conviction.” No documented current Pitango case in the corpus proves the transfer. [researcher inference · moderate] Taboola note, limits

Rule 2 — separate innovation density from market scale

Historical rule. Ask whether a geography supplies invention, representative users, rapid feedback, or scalable demand; do not collapse them into “ecosystem strength.” [researcher inference · strong] PMF essay

Diagnostic questions. Which uncertainty does the local market resolve? What foreign assumption remains? Is adoption voluntary, subsidized, mandated, or crisis-driven?

Modern VC translation (as of 2026-08-22). The essay's current cloud-and-digital-distribution premise supports testing remote feedback, but not assuming that regulation, procurement, localization, data rights, or willingness to pay transfer. Use a cross-market validation matrix. Misuse is to label a small friendly pilot global PMF; the corpus supplies no documented Pitango application. [researcher inference · strong] PMF essay, “Israel's Changing Global Footprint”

Rule 3 — underwrite dual use twice

Historical rule. A dual-use company must solve a real defense mission and a real civilian job; a larger combined market is not a substitute for either proof. [researcher inference · strong] dual-use essay

Diagnostic questions. Are product requirements compatible? Who owns certification, export approval, and liability? Can either market support the company alone? Which use is unacceptable?

Modern VC translation (as of 2026-08-22). Use autonomy, cyber, space, sensing, and medical response—the essay's current examples—as categories to test, not evidence that a company passes both underwritings. The rule fails when classified validation cannot convert into civilian evidence or export controls cap the market. Likely misuse: adding a civilian slide to enlarge TAM. No documented Pitango application is supplied. [researcher inference · strong] dual-use essay, investing and limitations sections

Rule 4 — make calibration bidirectional

Historical rule. Update comparables, duration, and scale assumptions as regimes change, but predefine evidence that forces calibration down as well as up. [researcher inference · strong] Calibrate Bigger

Diagnostic questions. Which reference companies truly match business model, margin, capital need, and regulation? What rate, liquidity, or adoption change invalidates them? What ownership and return follow from today's price?

Modern VC translation (as of 2026-08-22). Treat the essay's current hyperscale and GenAI examples as an analyst calibration exercise: model additional capital and time, then recalculate ownership and fund return rather than assuming an unprecedented ceiling erases fund math. Misuse is “giraffe” rhetoric as valuation permission. No documented current deal application is supplied. [researcher inference · strong] Calibrate Bigger, opening and “The empire builders”

Rule 5 — treat government revenue as a different risk system

Historical rule. Underwrite authority, appropriation, tender path, deployment, audit, ethics, renewal, and political concentration—not just contract size. [researcher inference · strong] B2G essay

Diagnostic questions. Who can sign and who can cancel? Is budget recurring? What working capital is required? Does one election, ministry, conflict, or prime contractor control the outcome?

Modern VC translation (as of 2026-08-22). For the essay's current quantum, digital-public-service, climate, infrastructure, and defense examples, test whether procurement modernization produces appropriated, renewable demand rather than merely announced policy. Likely misuse: equating stickiness with low risk while ignoring concentration, export controls, and accountability. The essay supplies cases, not a Pitango outcome cohort. [researcher inference · strong] B2G essay, five drivers

Rule 6 — map the people and vehicles before assigning skill

Historical rule. Record who sourced, referred, underwrote, approved, governed, and narrated, and through which vehicle. [researcher inference · strong] Shany note · Taboola note

Modern VC translation (as of 2026-08-22). Dense ecosystems improve access but make attribution easy to overstate. Use an influence map at IC and again at exit. Misuse is turning the most visible partner's retrospective into sole credit; Shany and Taboola are the current corpus examples, not proof of a portfolio-wide pattern. [researcher inference · strong] Shany note · Taboola note

Best writings and unresolved record

1. Three lessons I learned from Taboola's IPO — best deal-specific relationship, people, and decision narrative, with explicit hindsight limits.
2. Calibrate Bigger — clearest intellectual chronology of changing venture reference sets and the most important valuation tension.
3. Israel's Emerging Role in Global Product-Market Fit — most usable diagnostic framework once X/Y transfer limits are added.
4. A Two-Way Street — strongest current sector thesis and richest risk-countercase.
5. Shany first-investment reflection — indispensable chronology and entity-attribution control despite missing economics.

The B2G and Looking Ahead essays should follow as modern doctrine; each is useful but less deal-grounded than the top five. [researcher inference · strong] B2G note · Looking Ahead note

No writing in the priority corpus provides a complete contemporaneous IC memo, pass memo, ownership target, entry valuation rule, reserve policy, fund-duration adjustment, audited MOIC/IRR/DPI, full portfolio denominator, loss postmortem, dual-use ethics policy, B2G procurement cohort, or falsification record for “giraffe” outcomes. The highest-value retrievals are Mofet and Polaris II investment papers; Taboola 2009–12 IC and cap-table records; Pitango fund agreements and reserve models; rejected dual-use/B2G cases; and portfolio-wide outcomes that include write-offs. [researcher inference · strong] source-map coverage assessment

Coverage and evidence base

The priority spoken corpus contains 13 source-map items. All 13 are processed into exact-schema notes or an exact access-boundary note: 11 source items have full legitimately accessible text or media, TechCrunch is partial reporter-bounded coverage, and The Brothers route is metadata-only because its current official page exposes no recording or transcript. Because the two Wharton pages split one interview, the rows represent 12 underlying appearances: ten full, one partial, and one metadata-only. Eleven underlying appearances support content analysis; The Brothers item remains in the denominator but contributes no inferred speech. [contemporaneous record · strong] priority source map

The two Wharton pages are separate required source items but one underlying 2012 interview; the IsraelTech YouTube and Apple mirror are one work; the Leap landing page and Podscripts transcript route are one work. Mirrors and supplemental McKinsey or conference appearances are not used to inflate the 13-item source-map denominator. [contemporaneous record · strong] source-map boundary

The corpus runs from November 2007 through June 2026. It includes edited transcripts, adversarial financial-press interviews, long English and Hebrew media, and a printed transcript PDF. Direct voice establishes what Peres said at a date; it does not by itself prove causal skill, consistent implementation, sole deal credit, or fund returns. [researcher inference · strong]

Literal priority-source reconciliation

Dated source-by-source guide

1. Globes (2007) — diversification tested against an adverse mark

This is the earliest and most skeptical source. Peres gives a roughly 1,000-proposal/10–12-investment annual funnel, defines sector diversification and stage balance, and argues for a large partnership that can stay with companies. Globes simultaneously puts a negative interim CalPERS mark and IVC closure counts to the partners. [contemporaneous record · strong] 2007 note, takeaways and counterevidence

The interview’s discipline is speaker separation. Peres owns the diversification/balance wording and crisis-support stance; Kalish owns the adjacent hype-research and explicit stop-decision language. The unanswered final net-return question belongs in DEALS rather than being papered over with later platform scale. [researcher inference · strong] 2007 note, speaker boundary

2. Wharton part 1 (2012/2013) — export bridges and capital-source diversification

Peres frames Israel as a small export economy whose startups must bridge immediately to global markets. After the financial crisis, he says Pitango extended its company-level sector/stage diversification to LP geographies and investor types. [investor-stated · strong] Wharton part 1 note

Al Bawader appears as both a talent inclusion and market-creation project. The interview gives aspirations and selected counts, not audited vehicle outcomes; that boundary matters when converting a regional thesis into a venture rule. [investor-stated · moderate] Wharton part 1 note, access and open questions

3. Wharton part 2 (2012/2013) — the financing gap and vehicle fit

This is the best formation source. Peres distinguishes Yozma from Inbal, explains why Mofet chose a public-market route, and says quarterly disclosure was ill-suited to confidential, crisis-prone, long-duration startups. He dates the private partnership combination with Rami Kalish to 1996. [investor-stated · strong] Wharton part 2 note

His founder filter favors adaptable, persistent, honest domain experts over a fixed plan. It is a useful qualitative rubric, but the interview supplies neither a reference-check method nor false-positive and false-negative cases. [investor-stated · moderate] Wharton part 2 note, philosophy signals

4. Globes (2014) — durable companies, founder agency, and talent dilution

Peres combines ambition for enduring companies with a founder-first exit posture: the investor discusses, but the founder remains at the steering wheel. He also concedes only a small fraction of startups can become large and warns that startup proliferation disperses scarce technical talent. [investor-stated · strong] 2014 note

When pressed for relative fund performance, he declines false precision and says only that some funds are better and some worse. That refusal is a stronger evidence boundary than a marketing rank inferred from continued fundraising. [contemporaneous record · strong] 2014 note, final-performance exchange

5. TechCrunch (2016) — sector forecasts with a partial record

Reporter-bounded speech predicts that digitization will broaden Israel’s innovation beyond cybersecurity into health, drones, and autonomous driving; the report also attributes a billion-dollar-revenue scale filter to Peres. [investor-stated · moderate] TechCrunch note

The accessible page does not expose a full transcript, so the thesis cannot be scored without later deployment and outcome data. The Muslim-led-startup rationale is reporter paraphrase and remains weaker than the article’s direct quotations. [contemporaneous record · moderate] TechCrunch note, access boundary

6. Peace through Innovation (2020) — impact as a mainstream portfolio constraint

Peres describes Takwin’s capital mix and first cohort, then says Pitango chose to make the platform impact-oriented rather than isolate impact in a separate fund. He claims the aim was positive impact without return concessions; the statement is an objective, not an observed return comparison. [investor-stated · strong] Outcome support remains insufficient. Peace transcript note, printed pp. 9–10

The most decision-useful change is on printed p. 12: COVID-19 made him realize he had not previously tested whether companies were resilient to a global crisis. That is a specific claimed diligence update, still awaiting evidence of formal implementation. [investor-stated · strong] Peace transcript note, printed p. 12

7. Mizuho Conversation with Leaders (December 2022) — specialized vehicles and LPs as bridges

At 01:53–04:37, Peres describes parallel early, growth, and health vehicles and names digital transformation across industries as the first major trend. The reported raise, company, and LP figures are self-reported scale, not performance. [investor-stated · moderate] Mizuho note, 01:53–04:37

At 15:00–16:30 and 30:07–30:51, he describes regional normalization and LP relationships as routes to new markets, with Via in Japan as the selected example. The mechanism is plausible but selection-prone: one visible case does not show how often strategic LP value materializes. [researcher inference · moderate] Mizuho note, partnership evidence

8. Alain Elkann (2023) — technology married to values

Peres presents Israel’s innovation economy as a response to scarce resources, security burden, and small markets. He argues that technical capability is direction-neutral and needs values and governance. [investor-stated · strong] Alain note

The editorial introduction says he co-founded Pitango in 1993. That conflicts with stronger Mofet, SEC, and Globes records and cannot override the 1996 partnership date. The “empires of the mind” thesis also needs a counterweight: private technological power can create harms as well as public goods. [researcher inference · strong] Alain note, contradiction section

9. Human Founder (January 2025) — from unicorns to organizational “giraffes”

In Hebrew, Peres argues that Israel must widen its talent base, globalize early, and stop treating a unicorn valuation, IPO, or acquisition as the endpoint. At 47:42–51:27, a “giraffe” requires a compelling story, shared DNA, global access, and managers able to build sales, service, finance, and strategy at scale. [investor-stated · moderate] Human Founder note with timestamps and translations

The analogy is aspiration, not track record. Its value is as an organization-building checklist; its danger is using an extreme outcome target to rationalize unlimited capital or survivorship stories. [researcher inference · strong] Human Founder note, counterevidence

10. The Brothers Podcast (February 2025) — a documented access stop

Pitango’s page establishes a date and a broad global-impact theme but currently exposes no media, creator feed, runtime, transcript, hosts, or episode identifier. It therefore contributes no spoken rule, quote, or biography claim. [contemporaneous record · insufficient public record] Brothers access note

Keeping this row in the denominator prevents access failure from disappearing. Treating its promotional sentence as a processed transcript would contaminate the direct-voice record. [researcher inference · strong]

11. IsraelTech (April 2025) — debriefing, dual use, and the chronology trap

At 05:51–11:12, Peres treats honest debriefing as a survival practice and acknowledges failures without naming cases. At 27:40–30:23, he describes two-way civil/defense technology transfer. At 47:43–50:14, he reconstructs Mofet, Polaris, co-investment, and the 1996 move away from public-fund disclosure. [investor-stated · strong] IsraelTech note

The phrase “in 1993 we got together” is ambiguous and likely refers to co-investment, not a joint firm. The same answer distinguishes Mofet from his future co-founder’s Polaris, while stronger filings and contemporaneous press place the partnership in 1996. [researcher inference · strong] IsraelTech note, 48:57–49:17

12. Leap Academy (April 2025) — formation through professional and personal adversity

The chaptered interview connects Air Force debriefing, the Lavi program, job loss, software-company operating experience, and the turn to venture. Peres says he was more engaged by vision, strategy, and creation than by operation. [investor-stated · moderate] Leap note, 06:18–30:24

The third-party transcript is useful navigation but duplicates phrases and corrupts proper nouns. The note therefore uses timestamped paraphrase rather than unverified verbatim text, and it keeps the host’s “largest fund” framing out of the performance record. [contemporaneous record · strong] Leap note, access boundary

13. Calcalist (June 2026) — the chronology correction and mandate specialization

At 13:44–14:10, Peres explicitly dates the Kalish combination to 1996: Mofet and Polaris I records supported Polaris II, which he calls Pitango’s beginning. This late first-person account aligns with stronger chronology controls and corrects loose editorial 1993 co-founder language. [investor-stated · strong] Calcalist note, chronology evidence

At 17:56–18:31 and 28:24–31:44, he points toward AI, quantum, specialized funds, and a vehicle boundary between defensive dual-use and offensive defense. The third-party ASR is used only to navigate the canonical 46:56 Hebrew audio; noisy passages are paraphrased, and the actual mandate documents remain missing. [investor-stated · provisional] Calcalist note, timestamps and limits

Recurring stories, frameworks, and strongest counterarguments

People over plans—but make the people test falsifiable

Adaptability, persistence, honesty, storytelling, organizational DNA, and the ability to recruit recur from the 2012 interview through the 2025 media. [investor-stated · strong] Wharton founder filter · Human Founder scale doctrine

The counterargument is retrospective elasticity: a founder’s pivot can be called adaptability after success and lack of focus after failure. A usable IC version needs dated reference evidence, update behavior, and comparable passed or failed teams. [researcher inference · strong]

Diversify the portfolio and LP base—while concentrating institutional attention

The 2007 record defines sector and stage balance; 2012 adds geography and LP-type diversification. A large partnership is supposed to preserve enough company support despite that breadth. [investor-stated · strong] Globes 2007 · Wharton part 1

The tension is unresolved: wide exposure can reduce correlated risk or dilute conviction and partner time. No appearance provides allocation bands, ownership targets, reserve rules, or a with/without-outlier return test. [researcher inference · strong]

Match the vehicle to information cadence and duration

Mofet’s public structure and the later partnership provide the clearest operational lesson: early venture requires patient capital and enough confidentiality to work through crises. [investor-stated · strong] Wharton part 2 · IsraelTech, 47:43–50:14

The counterargument is agency: less public disclosure can enable patience, but it also reduces outside visibility into marks, fees, follow-ons, and failures. The 2007 CalPERS questioning shows why investor reporting still matters. [researcher inference · strong] Globes 2007 adverse record

Build durable companies without confiscating founder agency

The 2014 interview says exits are not the essence of industry-building, but the founder controls the steering wheel when a sale offer arrives. The 2025 “giraffe” language pushes the endpoint beyond IPO or unicorn status. [investor-stated · strong] Globes 2014 · Human Founder

This leaves a governance problem: LP liquidity, employee needs, founder concentration, and board fiduciary duties can diverge. No appearance supplies a difficult sale vote, contractual decision rights, or a founder-investor deadlock case. [researcher inference · strong]

Use optimism as an action engine, not a forecast score

Peres repeatedly turns future orientation into action: building cross-border markets, adapting after a crisis, and designing shared projects. [investor-stated · moderate] Peace transcript, printed pp. 5–8 · Alain interview

The counterargument is calibration. Optimism may expand search, but it can also underweight hostile regulation, financing scarcity, conflict, or execution. The record needs forecast scoring and named killed investments, not only inspirational consistency. [researcher inference · strong]

Innovation can bridge communities—but impact claims need dual measurement

Al Bawader and Takwin are presented as market-creation and inclusion vehicles; the Peace transcript adds an ESG/impact-platform doctrine. [investor-stated · strong] This label covers stated intent, not outcomes. Wharton part 1 · Peace transcript, printed pp. 9–10

The strongest alternative is that social purpose, fund economics, political legitimacy, and individual deal outcomes are different variables. The corpus supplies selected capital and company counts, not audited financial and impact results. [researcher inference · strong]

Debrief failures—but publish cases before calling it a repeatable edge

Military debriefing and candid mistake admission recur in IsraelTech and Leap, and COVID-19 produces a specific new resilience question in the Peace transcript. [investor-stated · strong] IsraelTech, 05:51–11:12 · Peace transcript, printed p. 12

Yet no talk discloses a full failed-investment postmortem, warning signs, reserve response, governance action, stop decision, or counterfactual. Process rhetoric without named losses cannot establish learning efficacy. [researcher inference · strong]

Biography, decisions, and attribution controls

[researcher inference · strong]

Evolution of spoken doctrine

[researcher inference · strong]

Five appearances to study first

1. Wharton part 2 — clearest formation, vehicle-fit, founder-filter, and partnership account.
2. Globes 2007 — only priority interview that pairs doctrine with direct adverse performance and closure questions.
3. IsraelTech 2025 — richest integrated account of debriefing, dual use, biography, and vehicle chronology; read with the 1993/1996 correction.
4. Calcalist 2026 — freshest formation correction and specialist/dual-use mandate discussion; Hebrew/ASR controls are essential.
5. Human Founder 2025 — strongest scale-company organization doctrine, with Hebrew originals and labeled translations.

Pair those five with Peace through Innovation for impact and crisis resilience, Globes 2014 for founder exit agency, and Mizuho for LP-network strategy. [researcher inference · strong]

Evidence still missing

No appearance supplies audited fund-level DPI, TVPI, or IRR; a complete Mofet or Peres-attributed deal denominator; ownership and reserve targets; IC voting; contractual exit rights; a full failed-investment postmortem; a rejected-deal sample; or a scorecard of earlier forecasts. The Brothers media remains unrecovered, the TechCrunch video lacks a full stable transcript, the Leap transcript is third-party ASR, and both Hebrew corpora require care with names and figures. [researcher inference · insufficient public record] literal reconciliation

The most valuable future retrieval would be Mofet annual reports and portfolio schedules, early partnership/LP records, named failed-deal discussions, stable creator media for The Brothers episode, and independent founder or partner accounts that can test Peres’s retrospective role narratives. [researcher inference · insufficient public record]

1. Convert “not yet” into an evidence contract

• Rule — When a founder's ambition is compelling but the business is not yet underwritable, record the exact missing evidence, owner, threshold, and recontact trigger; maintain the relationship without treating familiarity as commitment.
• Evidence status — [retrospective witness · moderate]. Taboola supplies one direct repeated-contact narrative, but neither the original IC record nor a pass cohort shows that Peres used a formal evidence contract or how often continued contact worked rather than prolonged indecision. Taboola note, takeaways and limits
• Provenance and origin — The earliest deal-specific articulation is Peres's 2021 IPO retrospective about meetings beginning in 2009, later business-model refinement, and a third meeting prompted by Zvi Limon. It is hindsight, not the original memo. CTech, opening and lessons
• Mechanism and reasoning — A precise deferral separates unresolved risk from rejection, lets new information change the decision, and reduces both premature commitment and story drift. The mechanism requires independent evidence and a deadline; repeated contact alone creates familiarity bias. [researcher inference · strong]
• Historical and vehicle context — The model appears most clearly in a mature private-fund period with the duration and network to revisit a company. It cannot be assumed for Mofet's public structure or for every later specialist vehicle because their decision records are absent. vehicle analysis [researcher inference · moderate]
• When to use — Use when product or market evidence can plausibly improve within the fund's horizon and the founder remains referenceable; do not use when integrity, legality, unit economics, or vehicle mandate already fails.
• How to apply it — Write: unresolved claim; baseline; requested independent proof; threshold; responsible person; expiry date; recontact trigger; and kill condition. At the next IC, compare only against that record, then decide invest, reset once with reason, or close.
• Examples —
  • Taboola — retrospective partial application, not a documented contract. Peres recounts three meetings, business-model refinement, a Limon referral, and eventual Pitango entry; the 2013 packet then showed named customers, product expansion, and reported growth. The dates conflict and no original threshold/owner/trigger record survives, so the evidence supports a relationship arc rather than the exact procedure reconstructed here. Taboola lab
  • Via — researcher-applied illustration. The 2015 packet contained rides, members, vehicles, and a constrained Manhattan corridor; an evidence contract would have made expansion economics, regulation, and retention the next-check gates. Nothing proves Peres used that exact procedure. Via lab [researcher inference · moderate]
  • GoRefer — failure/boundary illustration. When customers disappeared, the relevant evidence no longer improved; founders and co-investors closed while cash remained. The record does not show whether a pre-entry evidence contract existed. Globes 2002 [researcher inference · moderate]
• Evidence — The rule is grounded in Peres's Taboola retrospective, checked against the contemporaneous 2013 company release and the different outcome path at GoRefer. These are three underlying works and more than four claim-local placements across this section; none supplies a pass denominator. Taboola 2013 release · GoRefer postmortem
• Limits, failure modes, and exception ledger — The rule fails when uncertainty is irreducible, new evidence is founder-selected, the fund lacks time, or repeated meetings create commitment escalation. Taboola's IPO-selected narrative, unresolved 2009/2012/2013 dates, and absent pass memos are explicit exceptions; GoRefer shows that changing facts can require closure rather than another reset.
• Evolution and contradictions — Later writings elevate signal detection and recalibration, but publish no forecast ledger. The tension is between optimism that keeps a relationship alive and discipline that closes the file. Looking Ahead, signal passages [investor-stated · moderate]
• Modern VC translation (as of 2026-08-22) — [researcher inference · strong]. In an AI-native hypothetical, require retained paid use, named buyer authority, inference-cost trajectory, and reference quality rather than a better demo. Apply when those signals can change within the fund horizon; do not apply to integrity or legal failures. The likely misuse is retroactively naming every long courtship “patient conviction.” Taboola evidence boundary

2. Back adaptable truth-seekers and build the complementary team

• Rule — Prefer domain-credible founders who update honestly under pressure, persist without defending a frozen plan, and recruit complementary leaders before company complexity outruns them.
• Evidence status — [investor-stated · strong; application evidence moderate]. The founder filter recurs across Wharton, Taboola, and Human Founder, while the public record lacks a false-positive/negative cohort. Wharton note, philosophy signals
• Provenance and origin — The earliest accessible direct articulation is the 2012 Wharton interview: adaptable, persistent, honest domain experts matter more than plan fidelity. Taboola's 2021 retrospective adds complementary leadership; Human Founder in 2025 adds shared DNA and functional scale. Taboola note · Human Founder note, 47:42–51:27
• Mechanism and reasoning — Early plans face missing information. Honest updating improves search; domain competence constrains fantasy; persistence preserves learning; and complementary recruiting converts a founder's vision into product, sales, service, finance, and governance. The hidden assumption is that references can distinguish learning from opportunistic storytelling. [researcher inference · strong]
• Historical and vehicle context — Israel's small market and cross-border path increase the need for leaders who can add global commercial functions. A multi-partner private fund can help recruit them, but platform resources also make team contribution hard to attribute. Wharton part 1 note [researcher inference · moderate]
• When to use — Use from seed through scale whenever the plan must change but the problem and truth standard remain stable; do not excuse integrity failures, uncontrolled pivots, or a founder who cannot delegate as “adaptability.”
• How to apply it — Take dated examples of wrong calls, ask what evidence changed the founder's mind, check the same episode with former colleagues, map missing executive functions, set recruiting milestones, and predefine the behavior that would convert adaptation into drift.
• Examples —
  • Taboola — retrospective witness/documented team development. Peres describes an initially unfinished model, later clarity, and named complementary leaders; issuer records show his board tenure beginning in 2013. The IPO does not prove the founder rubric caused success. Taboola essay · Taboola proxy
  • Via — researcher-applied illustration. The 2015 evidence packet showed a functioning corridor and operating scale; the relevant team test was whether leaders could recruit and adapt across cities, regulators, and products. The record does not connect Peres's founder rule to the IC. Via lab [researcher inference · moderate]
  • GoRefer — counterexample/boundary. Customers disappeared and the shared decision was shutdown. An honest team can still face a broken market; the record does not score founder adaptation or assign the failure to people. GoRefer lab [contemporaneous record · moderate]
• Evidence — Independent underlying sources include the edited Wharton interview, Peres's signed Taboola essay, the 2013 Taboola release/issuer record, the long Hebrew Human Founder media, Via's contemporaneous round report, and Globes's GoRefer postmortem. They support the rule's articulation and boundaries more strongly than its predictive accuracy. Via 2015 report · Globes 2002
• Limits, failure modes, and exception ledger — Success makes pivots look wise and failure makes them look unfocused. Charisma can mimic honesty; complementary leaders can be added too late or lose authority; founder agency can conflict with boards, employees, and LP liquidity. No named passed-founder cohort, reference-check protocol, or difficult management-change case was found.
• Evolution and contradictions — The doctrine evolves from adaptable expert founder (2012) to durable-company agency (2014), complementary scale team (2021/2025), and “giraffe” organization. The contradiction is that founder-at-the-wheel rhetoric does not specify who decides when governance or liquidity interests diverge. Globes 2014 note [investor-stated · strong]
• Modern VC translation (as of 2026-08-22) — [researcher inference · strong]. Test AI-era founders on auditability, changing technical/economic assumptions, security/data constraints, and ability to recruit commercial and safety complements. Observable applicability is a dated reversal after disconfirming evidence plus stronger execution. Misuse is calling frequent narrative resets “learning” or charisma “domain expertise.” founder-filter evidence

3. Separate innovation density, test-market evidence, and scale-market proof

• Rule — State whether a geography supplies invention, representative users, fast feedback, voluntary economic demand, or scalable distribution; never use one as proof of the others.
• Evidence status — [investor-stated · strong; portfolio application moderate]. Peres's 2025 X/Y framework is explicit, while Shany, Taboola, and Via are supporting historical illustrations rather than documented applications of the later framework. PMF essay, X/Y sections
• Provenance and origin — The operating theme appears retrospectively in Shany's 1991 U.S. subsidiary and in the 2012 small-export-economy interview; the formal X/Y articulation arrives in 2025. Shany note · Wharton part 1 note
• Mechanism and reasoning — Dense technical networks improve invention and iteration; a small or urgent local test can reveal efficacy quickly; digital distribution can broaden reach. Transfer still fails when regulation, procurement, language, price, channel, liability, or willingness to pay differs. [researcher inference · strong]
• Historical and vehicle context — Early Israeli companies often needed U.S. commercialization before cloud distribution; later software can test remotely, while health, mobility, defense, and government remain locally authorized. Strategic LPs may bridge markets, but one Via/Japan account is selection-prone. Mizuho note, 30:07–30:51 [researcher inference · moderate]
• When to use — Use for every cross-border company and especially regulated, physical, or government-facing products; do not use the framework to dismiss valid home-market revenue or to assume all digital demand travels.
• How to apply it — Build a country-by-uncertainty grid: technical efficacy, user behavior, buyer authority, budget, price, channel, regulation/data, liability, retention, and repeatable acquisition. Pre-register which cell the next pilot must resolve and what would invalidate transfer.
• Examples —
  • Shany — researcher-applied historical illustration. The bootstrapped company had opened a U.S. subsidiary before Mofet/Dash Polaris invested, supplying action toward the scale market; no customer or economics data survives. Shany lab [retrospective witness · moderate]
  • Taboola — researcher-applied illustration. The 2013 packet named international publishers and expansion plans, stronger transfer evidence than Israeli technical density alone; audited cohorts and concentration remain missing. Taboola 2013 release [contemporaneous record · moderate]
  • Via — boundary/counterexample illustration. Manhattan corridor activity demonstrated one local operating system, not automatic transfer across cities, labor rules, transport authority, or pricing. Via lab [researcher inference · strong]
• Evidence — The rule draws on Peres's authored PMF essay, Wharton's edited global-market interview, the founder/Peres Shany retrospective, and contemporaneous Taboola/Via financing reports. More than two independent sources and four placements support the geography distinction; none reports a portfolio-wide transfer score. Via 2015 report
• Limits, failure modes, and exception ledger — Friendly pilots, subsidized demand, wartime urgency, strategic-LP introductions, and a famous foreign customer can all overstate transfer. The X/Y essay publishes no thresholds, failed countries, or company cohort. A local market can sometimes be the true scale market; the model should diagnose, not impose “global.”
• Evolution and contradictions — The doctrine moves from export bridges to a more granular geography-of-learning model. The newer claim that Israel can supply selected market-axis functions improves the early binary story, but still risks extrapolating crisis or government adoption. writings evolution [researcher inference · strong]
• Modern VC translation (as of 2026-08-22) — [researcher inference · strong]. Cloud and AI APIs reduce deployment friction but increase data-rights, model-governance, inference-cost, localization, and buyer-trust differences. Apply when foreign reference behavior repeats without subsidy or crisis compulsion; likely misuse is declaring global PMF from an Israeli or U.S. design-partner pilot. PMF framework and limits

4. Match vehicle duration, disclosure, and authority to the work

• Rule — Choose a capital vehicle whose time horizon, information cadence, liquidity, mandate, decision rights, and conflict controls match the uncertainty and repair cycle of the companies it owns.
• Evidence status — [documented behavior · strong]. The Mofet-to-Polaris II transition is the clearest behavior/rationale match; later specialist and continuation structures are documented, while their comparative outcomes and internal rights remain absent. Wharton formation source
• Provenance and origin — Mofet began in 1992 through a listed Inbal route. Peres later said public quarterly reporting was ill-suited to confidential startup crises, and in 1996 joined Kalish to build private Polaris II. IsraelTech note, 47:43–50:14 · Globes 1997
• Mechanism and reasoning — Long-duration private capital allows staged learning and repair without continuous public repricing; fees finance team capacity and profit participation aligns upside. The same opacity can hide marks, fees, allocation, and conflicts, so patience requires reporting and stop rights. [researcher inference · strong]
• Historical and vehicle context — Mofet/public shareholders, Polaris I/Yozma, Polaris II/private LPs, Pitango III, later Venture/Growth/Continuation entities, and impact/specialist initiatives had different capital and authority. No vehicle should be treated as Peres's personal account. vehicle table
• When to use — Use at fund design, stage/sector specialization, continuation transfer, cross-border mandate, and any company whose proof cycle exceeds the current fund's remaining life; do not solve a poor company by extending duration.
• How to apply it — Map capital source, legal owner, term, extensions, liquidity, disclosures, fee/carry, GP commitment, mandate, allocation, IC votes, reserves, valuation, conflicts, transfer approvals, key person, removal, and succession. Reject the design if a critical right is unknowable to its capital providers.
• Examples —
  • Mofet to Polaris II — documented application. Peres moved from a listed structure he described as mismatched to a private partnership beginning in 1996. This is the strongest model application. Wharton [documented behavior · strong]
  • Pitango III — counterexample/stress case. A $500 million vehicle increased support capacity and deployment/attention risk; the company count later exceeded the stated plan and interim marks were adverse. Pitango III lab [contemporaneous record · strong]
  • Via entities — researcher-applied later illustration. The 2025 filing identifies Venture, Growth, Principals, and Continuation entities around a retained stake. Extended ownership may fit duration, but transfer prices, approvals, fees, and allocations are unknown. Via 424B4, n.12 [researcher inference · moderate]
• Evidence — Independent supports include direct Wharton/IsraelTech formation accounts, contemporaneous Globes fund reporting, and primary SEC issuer ownership tables. They establish structure and stated mechanism, not comparative net performance. Taboola 13G/A, Item 2
• Limits, failure modes, and exception ledger — Private duration can become valuation opacity; specialist funds can fragment authority; continuation vehicles can shift conflicts and fees; public forms can offer accountability and liquidity. Fund terms, reserves, extensions, removal, succession, transfer approvals, and final net cash flows remain unavailable.
• Evolution and contradictions — The arc runs from listed general venture capital to private partnership scale and later specialist/continuation entities. Specialization may sharpen expertise while conflicting with Peres's earlier diversification doctrine and increasing overlap. Mizuho note, 01:53–04:37 [investor-stated · moderate]
• Modern VC translation (as of 2026-08-22) — [researcher inference · strong]. Apply to AI infrastructure, deep tech, health, defense, and secondary/continuation decisions by matching proof cycle to remaining term and making allocation/valuation approvals explicit. Observable fit is milestone runway plus transparent conflicts; misuse is calling any private or evergreen capital “patient.” vehicle evidence and limits

5. Diversify risks and capital sources without diversifying away attention

• Rule — Diversify genuinely different sector, stage, geography, and capital-source risks only to the level that partner capacity, reserves, ownership, and accountability can still support each company.
• Evidence status — [investor-stated · strong; outcome support insufficient]. Peres explicitly described sector/stage balance and later LP-geography/type diversification; adverse Fund III marks and missing allocation data prevent an efficacy claim. Globes 2007 note
• Provenance and origin — The earliest detailed source is the 2007 joint interview; the 2012 Wharton interview extends diversification from companies to LP geographies and types after the financial crisis. Wharton part 1 note
• Mechanism and reasoning — Different exposures can reduce correlated failure and strategic LPs can open markets; a larger partnership can distribute support work. Benefits disappear when hidden correlations, fund size, partner load, weak ownership, or reserve gaps overwhelm attention. [researcher inference · strong]
• Historical and vehicle context — Pitango III was planned for 30–35 companies with $6–12 million lifetime capital each, then was reported at 56 by 2007. Later separate early/growth/health and continuation entities added specialization and overlapping ownership questions. Globes 2000 · Via 424B4, n.12
• When to use — Use in fund construction and LP/customer-network design where exposures are measurably distinct and the team has capacity; do not use count, sectors, or LP logos as substitutes for correlation, ownership, and resource math.
• How to apply it — Pre-register sector/stage/geography correlations, company-count and ownership bands, reserve coverage, partner load, specialist authority, LP concentration, strategic-value evidence, conflict rules, and a closed-exit-market scenario. Re-run quarterly and after shocks.
• Examples —
  • Pitango III — documented plan and counterexample. The planned breadth/support model expanded beyond the reported company-count plan and met adverse interim marks; final return remains unknown. mark trajectory [contemporaneous record · strong]
  • Shany — researcher-applied syndication illustration. Mofet and Dash Polaris shared exposure and possibly complementary access, but no role, reserve, or proceeds record shows whether co-investment improved the outcome. Shany note [retrospective witness · provisional]
  • Via/Japan — investor-stated selected illustration. Peres said strategic LP relationships helped create market partnerships and named Via in Japan. One selected story does not show frequency, cost, or return effect. Mizuho note, 30:07–30:51 [investor-stated · moderate]
• Evidence — The model uses independent Globes interviews/fund reporting, Wharton's edited interview, a founder/Peres retrospective, Mizuho's canonical video, and Via's SEC filing. They establish doctrine, changing structures, and adverse controls, not a causal portfolio comparison.
• Limits, failure modes, and exception ledger — Sector labels can hide common financing/liquidity risk; stage balance can dilute specialist focus; strategic LPs can add conflicts or conditional access; a large fund can pressure deployment; more partners can blur accountability. No complete allocations, ownership, reserves, partner load, LP concentration, or ex-outlier cash-flow series was found.
• Evolution and contradictions — Company diversification (2007) expands to capital-source/geography diversification (2012) and specialist funds (2022). Specialization concentrates expertise while diversification spreads exposure; the reconciliation requires explicit mandate and capacity, which the public record does not provide. talks evolution [researcher inference · strong]
• Modern VC translation (as of 2026-08-22) — [researcher inference · strong]. Treat AI, cloud, geopolitics, rates, and exit liquidity as cross-sector common factors; stress partner time and reserves, not just company count. Apply when each exposure has a responsible owner and downside budget. Misuse is marketing broad sectors or LP nationalities as diversification. diversification evidence and tension

6. Recalibrate reference sets in both directions before scaling ambition

• Rule — Update comparable companies, market ceilings, capital needs, and duration when technology or macro regimes change, but predefine evidence that also forces expectations, price, and follow-on capital downward.
• Evidence status — [investor-stated · strong; decision application provisional]. Calibrate Bigger explicitly narrates successive reference shifts, but no deal-level valuation or forecast ledger proves bidirectional implementation. CTech, “How we got here” through “The empire builders”
• Provenance and origin — The retrospective history spans 1990s IP/OEM exits, internet IPOs, post-crash efficiency, low-rate unicorns, and post-COVID durability. The 2024 Looking Ahead essay and 2025 Human Founder media add “giraffe” scale. Looking Ahead note · Human Founder note
• Mechanism and reasoning — Old comparables can underprice a structural change; new references can reveal higher ceilings and longer investment needs. Upside-only calibration creates valuation permission, so the mechanism requires matching margins, capital intensity, regulation, distribution, and ownership math plus downward triggers. [researcher inference · strong]
• Historical and vehicle context — Peres participated across public-vehicle, Nasdaq, dot-com, low-rate, and later specialist/continuation regimes. Marks and exit windows changed dramatically, and large funds could finance longer ambitions while magnifying vintage risk. historical context
• When to use — Use after a demonstrable technology, distribution, regulation, financing, or buyer-authority change; do not use a fashionable category, one outlier, or a high private mark as the new base rate.
• How to apply it — Name the old and proposed comparable set; match business model, margin, capital, regulation, and channel; derive valuation and ownership; scenario time/capital up and down; define disconfirming metrics; and revisit before each follow-on.
• Examples —
  • Taboola — researcher-applied illustration. Product expansion, named publishers, and reported growth justified reassessing an earlier unfinished model; no entry valuation or fund-return math shows how far calibration moved. Taboola lab [researcher inference · moderate]
  • Via — researcher-applied illustration. A dense corridor could justify a larger mobility reference, but city transfer, capital, and regulatory assumptions had to be re-underwritten; the public packet does not disclose valuation discipline. Via lab [researcher inference · moderate]
  • Pitango III — failure/counterexample. A large dot-com-era vehicle and ambitious construction met sharply changing markets and adverse interim marks. It demonstrates why calibration must move down, while final performance remains unknown. Pitango III lab [contemporaneous record · strong]
• Evidence — Peres's authored CTech/Pitango essays and Hebrew media establish current doctrine; independent contemporaneous company releases, fund reporting, and SEC filings establish the examples and outcome boundaries. No source reports the actual valuation algorithm or comparable set used at IC.
• Limits, failure modes, and exception ledger — Narrative regimes can be identified only after the fact; one outlier can reset references incorrectly; bigger ceilings require more capital and dilute ownership; continuation marks can circularly validate price. Missing entry valuations, reserve decisions, rejected deals, and forecast scores keep implementation provisional.
• Evolution and contradictions — The public endpoint shifts from acquisition/IPO to unicorn and then enduring “giraffe.” Peres simultaneously criticizes unjustified unicorn valuations, creating the core productive tension: ambition must rise only after evidence, not serve as the evidence. writings, Calibrate analysis [researcher inference · strong]
• Modern VC translation (as of 2026-08-22) — [researcher inference · strong]. For an AI-infrastructure analyst exercise, double time and capital, model falling/volatile unit cost, recalculate ownership and fund return, and identify a lower-ceiling case. Apply when comparables match economics and authority. Misuse is using “unprecedented” to exempt price or duration from math. calibration evidence and contradiction

7. Underwrite public and dual-use markets through independent authorization and economics gates

• Rule — For government or dual-use companies, require each market to pass its own mission/user need, legal authority, budget, procurement/certification, delivery, renewal, unit-economics, export/end-use, liability, and ethics gates; never add TAMs before both paths work.
• Evidence status — [investor-stated · provisional]. The rule reconstructs two 2025 essays and 2026 mandate speech into an operational gate. The record contains investor-cited external cases but no complete Pitango cohort, rejected cases, terms, or outcomes under the doctrine. dual-use essay · B2G essay
• Provenance and origin — Peres's 2025 dual-use essay describes two-way civilian/defense learning; the B2G essay describes falling procurement barriers and durable public demand. In 2026 he distinguished defensive dual-use from a separate offensive-defense vehicle, but mandate documents are unavailable. Calcalist note, 28:24–31:44
• Mechanism and reasoning — Defense urgency and public budgets can fund costly validation; civilian competition can enforce productization and cost discipline. The paths can also conflict on specifications, secrecy, export, liability, ethics, procurement timing, and buyer concentration. Independent gates prevent one market's urgency from laundering the other's weakness. [researcher inference · strong]
• Historical and vehicle context — Israel's movement among military units, defense firms, founders, and reservists can accelerate knowledge flow, while state customers impose authorization and accountability. A separate offensive-defense mandate suggests risk/authority differences, not proof of execution. dual-use note [investor-stated · moderate]
• When to use — Use for defense, autonomy, cyber, space, sensing, health emergency, infrastructure, quantum, climate, and digital public services; do not treat a government logo, pilot, wartime use, or announced budget as appropriated recurring demand.
• How to apply it — Map end user, statutory buyer, budget owner, tender/prime, certification, export/end-use, data/security, prohibited uses, audit/appeal, renewal, working capital, concentration, product customization, civilian buyer/channel, price, liability, and kill criteria for each path.
• Examples —
  • SpaceX — investor-cited external illustration, not a Peres investment. Peres uses it to illustrate civilian/defense transfer. The example shows possible dual-market productization, not that every technology transfers or that Pitango captured the economics. dual-use essay, example passages [investor-stated · moderate]
  • PsiQuantum — investor-cited external B2G illustration, not a Peres investment. The essay invokes a government-linked project to show state participation in strategic infrastructure; it does not supply procurement economics or a portfolio attribution. B2G essay, examples [investor-stated · provisional]
  • Starlink — investor-cited external boundary case, not a Peres investment. Peres uses it around public/private infrastructure; the case highlights scale and strategic relevance while leaving authorization, concentration, accountability, and investor economics outside this record. B2G note, takeaways and limits [researcher inference · provisional]
• Evidence — The underlying evidence consists of two signed 2025 essays, a 2025 IsraelTech discussion of two-way transfer, and 2026 Hebrew audio on mandate separation. These are independent direct-voice works and supply more than four placements, but all three named cases are external illustrations rather than documented Pitango applications. IsraelTech note, 27:40–30:23
• Limits, failure modes, and exception ledger — Combined TAM double-counts incompatible products; crisis adoption distorts willingness to pay; public revenue can be slow, concentrated, political, or classified; civilian markets may reject cost, liability, or ethics. Attempted record routes: signed Pitango/CTech writings yielded doctrine but no cohort; current portfolio/team pages yielded no doctrine-tagged attribution; SEC and company filings yielded no Peres-linked dual-use/B2G case chain; IsraelTech and Calcalist media yielded categories/mandate language but no IC criteria; sources/notes contain no rejected case, return series, or end-use policy. The model remains provisional rather than padded.
• Evolution and contradictions — The 2025 two-way thesis emphasizes opportunity; B2G adds public customers and diplomacy; 2026 speech separates defensive from offensive mandates. The distinction could improve governance or merely rename risk; no implementation record decides. talks evolution [researcher inference · moderate]
• Modern VC translation (as of 2026-08-22) — [researcher inference · strong]. Apply only when authority, appropriated budget, repeatable procurement, productized delivery, compatible civilian economics, and enforceable prohibited-use controls are visible. A current example remains unavailable in the completed Peres record; treat any new candidate as a labeled IC exercise. Likely misuse is adding defense and civilian markets or equating stickiness with low risk. dual-use evidence and limits · B2G evidence and limits

8. Debrief losses, preserve residual value, and earn the next check

• Rule — After a miss or shock, compare the dated forecast with reality, name the controllable error, preserve cash and stakeholder value, change a specific control, and release more capital only against new evidence.
• Evidence status — [investor-stated · moderate; implementation evidence provisional]. Debriefing recurs in direct speech and GoRefer supplies one stop, but no portfolio-wide postmortem series or repeated-error test survives. IsraelTech note, 05:51–11:12
• Provenance and origin — Peres retrospectively traces candid debriefing to Air Force practice; the earliest public company-level stop in the researched record is GoRefer in 2001–2002. In 2020 he said COVID-19 made him add global-crisis resilience to diligence. Globes 2002 · Peace note, printed p. 12
• Mechanism and reasoning — A dated forecast prevents outcome rewriting; causal classification separates bad luck, thesis error, execution, governance, and integrity; residual-value planning limits loss; a changed control turns experience into future behavior. The next check is a fresh decision, not a reward for persistence. [researcher inference · strong]
• Historical and vehicle context — Venture duration allows repair, but finite fund life and reserves impose opportunity cost. Dot-com customer collapse, Pitango III marks, and later global shocks test whether support remains rational; less public disclosure makes internal debrief quality more important and less observable. historical context
• When to use — Use after missed milestones, lost customers, valuation breaks, regulatory shocks, failed financings, and portfolio-wide crises; do not confuse honest failure with integrity failure or extend capital without a new falsifiable plan.
• How to apply it — Record original thesis/forecast; actual result; earliest warning; root-cause class; avoidable decision; board/management response; residual cash/customer/employee/IP plan; changed control; follow-on threshold; owner; date; and a test for repeated error.
• Examples —
  • GoRefer — documented shared stop. Customers disappeared; founders and co-investors chose closure with cash remaining, and a Hebrew account said roughly one-third of the round was returned. No final fund loss or later process change is public. GoRefer lab [contemporaneous record · moderate]
  • Pitango III — portfolio-level counterexample/stress case. Interim press marks swung across definitions, closures occurred, and the planned company count expanded; absent final cash flows and decision logs, the public record cannot show which controls changed. mark trajectory [contemporaneous record · strong]
  • Webbing — researcher-applied governance illustration. An IPO path was abandoned over valuation differences and a later strategic sale occurred; that could preserve value, but the board vote, rejected price, counterfactual, and proceeds are missing. Webbing lab [researcher inference · moderate]
• Evidence — Independent works include Globes's contemporaneous GoRefer and Pitango reporting, Peres's IsraelTech/Leap debriefing testimony, the printed Peace transcript, and CTech's Webbing transaction report. They establish culture, one stop, one stated diligence change, and adverse cases—not a repeatable outcome advantage. Leap note, debriefing/formative passages · CTech Webbing
• Limits, failure modes, and exception ledger — Debriefs can become retrospective theater; optimism can dilute the stop; sunk cost can masquerade as patience; public postmortems select acceptable failures. Shopserve lacks formation and response evidence; GoRefer lacks the original memo and final fund loss; Pitango III lacks reconciled cash flows; no named pass/miss cohort or repeated-error rate was found.
• Evolution and contradictions — The record moves from one 2002 shutdown to 2007 crisis-support doctrine, 2020 resilience, and 2025 debriefing rhetoric. Support and stopping are complementary only with explicit milestone and opportunity-cost rules, which remain undisclosed. Globes 2007 note [researcher inference · strong]
• Modern VC translation (as of 2026-08-22) — [researcher inference · strong]. Add model/data incidents, cyber exposure, supplier concentration, geopolitical/export shocks, and capital-market closure to the debrief. Observable transfer is a changed control and lower repeated-error rate. Likely misuse is celebrating failure, hiding the named case, or funding another round because the relationship is long. debriefing evidence and limits

Evidence gaps and candidate models

The inventory reconciled every distinct rule in the seven writing notes, thirteen talk notes, writings synthesis, talks synthesis, investment forensics, and the model index in investment philosophy. Synonyms were merged: patient contact with evidence contracts; people-over-plans with complementary-team construction; export bridges with X/Y market transfer; diversification with LP/source breadth; and giraffe ambition with bidirectional calibration. [researcher inference · strong]

Plausible candidates excluded from model status:

• Impact without return concession. Peres states the objective and describes Takwin/Pitango impact structures, but the record has no matched financial/impact outcomes, denominator, or case-linked IC rule. Peace note, printed pp. 9–10 [investor-stated · insufficient public record]
• Founder controls the exit. Direct voice supports founder-at-the-wheel doctrine, but no difficult sale vote, contractual allocation of authority, dissent, or employee/LP conflict case demonstrates the reusable decision procedure. Globes 2014 note [investor-stated · insufficient public record]
• Strategic LPs create markets. Via/Japan is one selected account; no portfolio frequency, counterfactual, cost, conflict, or return effect establishes a model independent of global-market transfer and diversification. Mizuho note, 30:07–30:51 [investor-stated · provisional]
• Technology plus values. The moral direction claim is explicit, but no portfolio-level ethics screen, prohibited-use policy, governance control, or failure case was found. It remains a principle and a retrieval priority, not an evidenced investment model. Alain Elkann note [investor-stated · insufficient public record]
• “Giraffe” as a separate model. Compelling story, shared DNA, global access, and functional scale are useful diagnostics, but the animal label has no threshold, denominator, or independent outcome test. Its supported content is incorporated into team construction and bidirectional calibration. Human Founder note [researcher inference · strong]
• Map people and vehicles before assigning credit. This is an essential researcher control, not demonstrated as Peres's investment rule; it is applied throughout the biography, philosophy, and deal ledger rather than inflated into a personal model. people map [researcher inference · strong]

Decisive future evidence would include Mofet and Polaris II memos/pass logs; complete partnership and fund cash flows; deal-level role/IC/board records; a named founder false-positive/negative cohort; current dual-use/B2G mandate, ethics, rejected-case, and outcome records; impact scorecards; and successor/key-person terms. terminal uncertainty ledger [researcher inference · strong]

What the evidence supports

The strongest defensible account is narrower than a conventional “great picker” biography. Peres is well documented as an institution builder and translator of constraints: he founded the listed Mofet vehicle in 1992, concluded that public-market disclosure and liquidity were a poor fit for fragile startup work, and joined Rami Kalish in 1996 to build the private Polaris II successor platform. Wharton, Mofet and 1996 formation answers · Globes 1997, Polaris II managers and economics [documented behavior · strong]

The package's edge thesis—identify the interface blocking a technically strong company, then match the relationship, people, market bridge, and vehicle to that bottleneck—is falsifiable researcher synthesis, not a rule Peres published. It has moderate support from the Mofet-to-Polaris II change, the Shany cross-vehicle case, Taboola's repeated-contact narrative, and later market/vehicle doctrine. It has no complete pass cohort, intervention ledger, or audited outcome comparison. profile edge thesis · Taboola decision note [researcher inference · moderate]

Three findings survive the adversarial pass:

• Vehicle fit is the best-supported repeatable lesson. Mofet and Polaris II are distinct, and the rationale for changing capital form is explicit. Later specialist and continuation structures are visible, but their comparative outcomes and conflicts are not. vehicle chapter [documented behavior · strong]
• Patient relationships and complementary teams are supported behaviors, not proven alpha. Taboola supplies a named multi-person chain and board tenure from 2013; Via and Webbing show governance roles. None isolates Peres's causal contribution or personal economics. Taboola proxy, director and ownership n.10 · Via 424B4, director biography and n.12 [documented behavior · moderate]
• The failure record constrains the favorable story. GoRefer records a shared decision to stop while cash remained; Shopserve lacks a reconstructable formation or shutdown record; Pitango III has adverse, inconsistent interim marks but no final audited return. failure postmortems [contemporaneous record · moderate]

The chronology control is robust: Mofet in 1992; Kalish/Dovrat–Shrem/Yozma Polaris I in 1993; Peres joining the Polaris II successor in 1996. No reader-facing document now backdates Peres into Polaris I or presents a 1993 co-founding as fact. Globes 2000, Polaris I/II chronology · AudioCodes 2019 Form 20-F, printed p. 82

What could falsify the thesis

The institutional-translation thesis predicts records that are not yet public. It weakens or fails if any of the following evidence appears:

1. Original pass and IC logs show no dated uncertainty, evidence threshold, or re-entry trigger in deferred deals; Taboola would then be a successful retrospective relationship rather than evidence of a repeatable process. Taboola evidence boundary
2. Founder, partner, or board records assign the decisive sourcing, diligence, recruiting, market-opening, governance, or exit work to other people in the cases now associated with Peres. people and decision-rights map
3. Mofet, Polaris II, and later vehicle records show that capital form did not alter pacing, disclosure, reserves, intervention, or company outcomes—or that opacity and conflicts outweighed the claimed patience benefit. vehicle uncertainty
4. An ordinary-company and failed-company sample shows no repeatable bottleneck diagnosis or intervention advantage outside the public winners. The present 12-case research set is selected and only six cases are sufficiently reconstructed. denominator reconciliation
5. Complete, vintage-adjusted cash flows show ordinary or poor results after fees, failures, access, team, sector, era, and Taboola/Via outliers are controlled. No public source currently permits that test. ex-Taboola/ex-Via sensitivity

The favorable thesis strengthens only if the same predicted behavior appears in original memos, board records, founder accounts, pass cohorts, and complete cash flows across winners, ordinary companies, and failures. [researcher inference · strong]

Attribution and denominator audit

The attribution audit rejects a Peres-only track record. Shany belongs to Touboul's team plus Mofet/Peres and Dash Polaris/Kalish; AudioCodes includes founders, a DS Polaris/Chase/Genesis syndicate, and Eyal Kishon's board role; Taboola includes Singolda's team, Limon, other investors/directors, multiple Pitango entities, and six managers sharing power; Via includes founders, co-investors, multiple Pitango entities, and a Peres board seat; Webbing identifies Pitango as primary investor and Peres as board chair but does not expose sourcing, voting, or seller proceeds. Shany note · AudioCodes 2003 Form 20-F, Exhibit 10.4 and director biography · Taboola ownership control · Webbing transaction report

The denominator has two bounded objects and no career fraction:

All Pareto rankings are therefore directional. Taboola is first on the economic-proxy list because a 2025 issuer filing discloses about $23.4 million of cash paid for 6,387,648 shares sold by certain Pitango-affiliated funds; that is partial vehicle cash, not total proceeds, MOIC, fund return, or Peres's personal economics. Taboola 8-K, Item 1.01 [contemporaneous record · moderate]

Family-name, family-wealth, military-network, and public-service access are hypotheses or visibility confounders to test. No reviewed source shows that family wealth financed a vehicle or that family-name or military access improved deal introductions. Public board roles and one selected Via/Japan LP-bridge account are documented; a causal access advantage is not. access control [researcher inference · insufficient public record]

Transfer limits for modern VCs

The transferable product is a set of controls, not a claim that today's analyst can copy Peres's outcomes:

• Evidence contracts transfer. A modern analyst can date what blocks investment, who must produce independent evidence, and what reopens or kills the file. The public record does not show that Peres formalized this process or its hit rate. model 1 [researcher inference · moderate]
• Vehicle matching transfers only with transparency. Long proof cycles in health, deep tech, AI infrastructure, defense, and continuation structures make duration relevant, but private duration can hide valuation, fee, allocation, and conflict risk. model 4 [researcher inference · strong]
• Geography must be decomposed. Innovation density, fast feedback, voluntary demand, procurement authority, and global distribution are different variables. Cloud/API reach does not erase data, liability, localization, price, or authorization. PMF writing [investor-stated framework; application support moderate]
• Dual use and B2G require independent gates. Civilian and government/defense demand, authority, economics, end-use, export, ethics, and renewal must each stand alone. Peres's current doctrine has no complete Pitango cohort or outcome record, so model 7 remains provisional. dual-use essay · B2G essay [investor-stated · provisional]
• Debriefing transfers only when behavior changes. A useful postmortem names the original forecast, warning, root cause, residual-value action, changed control, owner, and repeated-error test. Public speech establishes the aspiration and one GoRefer stop, not a portfolio-wide implementation record. model 8 [researcher inference · moderate]

The completed-peer controls are dimension-matched, not return comparisons: Yigal Erlich for Israeli vehicle architecture, Jim Breyer for institutional network/board/team attribution, and Trae Stephens only for current defense/government authorization and role-conflict questions. peer comparison

Unresolved evidence

The assessment would change materially with:

• complete Mofet prospectuses, holdings, write-offs, reserves, and liquidation cash flows;
• Polaris II and Pitango vintage LP reports, final net DPI/TVPI/IRR, fees, distributions, and manager/carry allocation;
• pass logs and original IC records, especially for Taboola and ordinary or failed companies;
• board minutes and founder/co-investor accounts allocating sourcing, diligence, recruiting, follow-on, stop, and exit work;
• a complete Pitango/Peres vehicle-company ledger with aliases, follow-ons, removed companies, passes, and write-offs;
• current dual-use/B2G mandates, prohibited-use rules, allocation/conflict policy, rejected cases, and outcomes;
• a stable creator recording or transcript for The Brothers episode; and
• exact Mofet identifiers in TASE archives and legitimate archived copies of still-indexed but inaccessible low-priority event/filing routes. source-map coverage assessment

The direct-link review found that core SEC, CTech, Pitango, PR Newswire, issuer, and Globes evidence used for central claims remained retrievable or returned a documented access restriction. A former Peres Center PDF was replaced with the live 2018 report; the profile was corrected from an unsupported exact 2017 chair-start claim to “documented chair by 2018.” The Webbing source date was corrected to 2023-08-09. The old Sachs conference-guide route now records its 2026-08-22 404 barrier instead of implying current access. Peres Center 2018 report, printed p. 14 · Webbing report

Twenty-point acceptance evidence

Central-claim audit trail

Quote, number, and economic-figure audit

The note corpus contains 45 admitted short quotations and two explicit no-quote records. Every admitted quote is 25 words or fewer, appears in an exact-schema note, and has a heading, page, or timestamp locator; translations are labeled. Core documents either paraphrase or reuse these bounded quotations with local/source links.

Ordinary-claim audit trail

The sample was selected reproducibly before opening sources: two early, two middle, and two late line regions across the seven pre-review core documents, then one claim from this review's own denominator section. This spreads the test across all eight core documents instead of clustering around the favorable thesis.

Mechanical reconciliation record

• Core package: 8 substantive root documents after this review; 20 notes = 7 writing + 13 talk source-item notes.
• Openings: every root document's next nonblank line after H1 is a substantive H2; investments.md opens with the prescribed Pareto heading.
• Source map: 100 checkbox rows; 100 unique primary row URLs; 62 processed and 38 retained leads/barriers. The map contains every external URL cited across the seven pre-review core documents and notes.
• Priority parity: writings 7/7; spoken source items 13/13, representing 12 underlying appearances and preserving one partial plus one metadata-only boundary.
• Decision labs: 7/7, each with four no-hindsight worksheet questions and separate reveal; includes GoRefer failure and Pitango III adverse fund case.
• Mental models: 8 models; each has all 12 exact labels and 3 case bullets; model 7 contains the required attempted-route ledger.
• Quotes: 45 admitted short quotations, all at most 25 words; two explicit no-quote records; Hebrew translations labeled.
• Links: all 27 pre-review Markdown files' relative file and anchor targets resolve mechanically; review links were checked after drafting. External central links were reopened or assigned an explicit barrier.
• Chronology: no unsupported 1993 Peres/Pitango co-founding claim; supported sequence is preserved everywhere.
• Hygiene: no placeholder, full transcript, orphan temporary file, or status/run preamble appears in the investor package; git diff --check is required again at handoff.

Reader conclusion

The completed package supports studying Chemi Peres as a builder of vehicles, cross-border interfaces, patient relationships, and partnership capacity under Israel's changing venture constraints. It does not support an audited personal-return ranking, a lone-picker legend, a family-access mechanism, or a claim that current dual-use/B2G doctrine has produced measured outcomes. The useful lesson is conditional: match people, evidence, market authorization, capital duration, and decision rights to the actual bottleneck—then preserve enough contemporaneous evidence to prove whether that diagnosis helped. [researcher inference · strong]

Source-map boundary and retrieval priorities

This map separates Peres's direct voice, third-party biographies, issuer filings, firm and vehicle records, portfolio-company outcomes, nonprofit/impact entities, and adverse context. A checkbox means processed into the completed package, not endorsement. Platform history beginning in 1993 is not evidence that Peres joined before the 1996 Polaris II combination; issuer or fund holdings are not personal ownership or personal returns.

Priority written corpus

• ★★★ | Three lessons I learned from Taboola's IPO | https://www.calcalistech.com/articles/0%2C7340%2CL-3911460%2C00.html | 2021-07-01 | signed essay | Peres's three-meeting investment recollection, named contributors, and retrospective founder/culture lessons; canonical article body was recovered through CTech and its access boundary remains explicit.
• ★★★ | Looking Ahead to 2025 With Optimism and Hope | https://www.pitango.com/resources/looking-ahead-to-2025-with-optimism-and-hope/ | 2024-12-26 | signed firm essay | Direct voice on signal versus noise, technology, optimism, and future-oriented reasoning.
• ★★★ | Israel's Emerging Role in Global Product-Market Fit | https://www.pitango.com/resources/israel-role-product-market-fit/ | 2025-02-13 | signed firm essay | Thesis that Israel can extend from R&D into product-market-fit and growth work; test policy and market assumptions.
• ★★★ | A Two-Way Street: The Future of Dual-Use Technology | https://www.pitango.com/resources/dual-use-technology/ | 2025-03-27 | signed firm essay | Direct doctrine on military/civilian technology transfer, markets, mandate limits, and the governance risks of dual use.
• ★★★ | Calibrate Bigger: Rapidly Adapting to a Paradigm Shift in Tech Investment | https://www.calcalistech.com/ctechnews/article/sjuhapnmlx | 2025-05-28 | signed CTech essay | Canonical Chemi byline for hyperscale, activist-CEO, GenAI, and large-outcome framing; a full Chemi-byline Pitango mirror and a separate metadata-only Pitango duplicate are retrieval controls, not independent works.
• ★★★ | From Red Tape to Revenue: The Growth of Business-to-Government Partnerships | https://www.calcalistech.com/ctechnews/article/rk66rfjdex | 2025-07-27 | signed CTech essay | Canonical Chemi byline on B2G markets, government sales, policy, and strategic power; the later Pitango mirror adds an editorial portfolio subsection and is not independent evidence.
• ★★ | First Investment as a Venture Capitalist: Shany Reflection | https://www.linkedin.com/posts/chemiperes_it-was-my-first-investment-as-a-venture-capitalist-activity-7388680205738881024-ZKs8 | 2025-10 | LinkedIn post with founder account | Retrospective claim tying Mofet and Kalish's Polaris vehicle to Shany; spelling and “first” assertions require controls.

Official identity, firm, archive, and impact routes

• ★★★ | Nechemia (Chemi) J. Peres — Pitango biography | https://www.pitango.com/people/nechemia-chemi-j-peres/ | live 2026 | official biography | Current title and firm affiliation; the page is marketing and not a chronology or performance audit.
• ★★★ | Pitango team directory | https://www.pitango.com/team/ | live 2026 | official team page | Current partner, stage, sector, finance, and operating-role map.
• ★★★ | Pitango portfolio directory | https://www.pitango.com/portfolio/ | live 2026 | official portfolio page | Reproducible 2026-08-22 platform snapshot supplied the 140-card marketing-directory denominator; it is not a historical fund ledger or Peres-only record.
• ★★ | Pitango resources archive | https://www.pitango.com/resources/ | live 2026 | official content index | Enumerates firm writings, reports, interviews, and transcript routes; author and date require page-level verification.
• ★★ | Pitango Impact & Sustainability Report 2023 | https://www.pitango.com/wp-content/uploads/Pitango-Impact-Sustainability-Report-2023.pdf | 2024 | firm report PDF | Current impact framing and joint leadership statements; firm self-report, not audited fund performance.
• ★★ | Pitango ESG & Impact Report 2020–2022 | https://www.pitango.com/wp-content/uploads/Pitango-ESG-Impact-Report-2020-2022-PUBLIC.pdf | 2023 | firm report PDF | Impact Committee, diligence, governance, LP reporting, and team/vehicle distinctions.
• ★★ | Pitango Growth Playbook: Navigating the Growth Curve | https://www.pitango.com/wp-content/uploads/Pitango-Growth_Annual-Event_Play-Book-3.pdf | 2025 | firm/McKinsey report PDF | Firm operating doctrine derived from founder interviews; not necessarily Peres-authored.
• ★★ | Peres Center 2018 Impact Report | https://www.peres-center.org/media/1711/peres-center-2018-impact-report.pdf | 2019 | nonprofit report PDF | Printed p. 14 documents Peres as board chair by 2018 and separates nonprofit governance from Pitango; it does not establish the exact appointment date.
• ★★ | Peres Center official site | https://www.peres-center.org/en/ | live 2026 | nonprofit website | Current mission, leadership, programs, and archive route; no inference to venture returns.
• ★★ | Reinhard Mohn Prize 2020 announcement and program | https://www.bertelsmann-stiftung.de/en/our-projects/reinhard-mohn-prize/announcement-rmp-2020 | 2020–2021 | foundation archive | Direct quote, award rationale, and discussion route.
• ★★★ | Reinhard Mohn Prize bilingual biography dossier | https://doku.bertelsmann-stiftung.de/fileadmin/files/BSt/Publikationen/Infomaterialien/Preistraegerbroschuere_zum_Reinhard_Mohn_Preis_2020.pdf | 2021 | foundation dossier PDF | Detailed formation, education, operating chronology, Mofet history, and social initiatives.
• ★★ | Saban Forum 2005 proceedings | https://www.brookings.edu/wp-content/uploads/2012/04/sabanforum2005_proceedings.pdf | 2005 | conference proceedings PDF | Contemporaneous biography naming Mofet 1992 and Pitango partnership 1996.
• ★ | Pitango April 2025 newsletter | https://www.pitango.com/resources/newsletter-april-2025/ | 2025-04 | official newsletter | Confirms contemporaneous written-corpus publication sequence and source links.
• ★ | Pitango 2026 March newsletter | https://www.pitango.com/resources/pitango-march-2026-newsletter/ | 2026-03-31 | official newsletter | Current platform activities and Peres appearance lead; marketing, not deal evidence.

Priority talks, interviews, podcasts, and transcript routes

• ★★★ | Venture Capitalist Chemi Peres: Opening Israeli Ventures to New Markets | https://knowledge.wharton.upenn.edu/article/venture-capitalist-chemi-peres-opening-israeli-ventures-to-new-markets/ | interview conducted 2012; published 2013-05-10 | edited transcript, part 1 | Direct macro, export, capital-source, and ecosystem account; live fetch can return 403, so indexed-body access remains explicit. Note: notes/t-wharton-opening-new-markets.md.
• ★★★ | Venture Capitalist Chemi Peres: Building a Vehicle to Finance Entrepreneurial Dreams | https://knowledge.wharton.upenn.edu/article/venture-capitalist-chemi-peres-building-a-vehicle-to-finance-entrepreneurial-dreams/ | interview conducted 2012; published 2013-05-10 | edited transcript, part 2 | Highest-value direct chronology for Mofet, the financing gap, the 1996 Kalish partnership, founder filters, and reported scale; same underlying interview as part 1. Note: notes/t-wharton-building-vehicle.md.
• ★★★ | Balance and diversification at Pitango | https://en.globes.co.il/en/article-1000280726 | 2007-11-29 | edited joint interview | Peres/Kalish direct voice plus adverse questioning on Pitango III, closures, exits, and interim CalPERS marks; every rule keeps its speaker boundary. Note: notes/t-globes-balance-diversification-2007.md.
• ★★★ | Chemi Peres: Exits aren't enough for a real industry | https://en.globes.co.il/en/article-exits-cant-make-a-real-industry-1000984714 | 2014-11-09 | edited interview | Founder control of exits, local institutional LPs, large-company formation, and refusal of false relative-return precision. Note: notes/t-globes-exits-industry-2014.md.
• ★★ | Chemi Peres makes predictions for Israel's tech scene | https://techcrunch.com/2016/08/29/chemi-peres-makes-predictions-for-israels-tech-scene/ | 2016-08-29 | video with report | Reporter text supports bounded direct quotations and sector predictions; it is not a complete video transcript. Note: notes/t-techcrunch-predictions-2016.md.
• ★★ | Chemi Peres — Alain Elkann Interviews | https://www.alainelkanninterviews.com/chemi-peres/ | 2023-03-12 | edited interview and podcast | Direct biography, innovation/peace framework, and Israel ecosystem claims; the editorial 1993 co-founder line conflicts with stronger chronology controls. Note: notes/t-alain-elkann-2023.md.
• ★★★ | With Eyes Closed, You See Everything — The Human Founder #156 | https://www.galiblochliran.com/podcast/with-eyes-closed-you-see-everything-chemi-peres-on-humility-vision-and-optimism-in-entrepreneurship-and-beyond | 2025-01-06 | 90-minute Hebrew podcast | Hebrew speech is timestamped with original language and researcher-labeled translations; ASR names, dates, and numbers remain controlled. Note: notes/t-human-founder-156.md.
• ★★★ | How to Build the Next Big Thing — The Brothers Podcast | https://www.pitango.com/resources/chemi-peres-brothers-podcast/ | 2025-02-27 | official episode route | The current page exposes only a description and no stable creator recording or transcript route; exact metadata-only access boundary recorded. Note: notes/t-brothers-next-big-thing.md.
• ★★★ | Legacy, Leadership & the Future of Israeli Innovation | https://www.youtube.com/watch?v=jVDUpY7x_xE | 2025-04-05 | 67-minute IsraelTech video | Full English automatic captions processed with names, dates, numbers, and ambiguous 1993 co-investment recollection controlled against stronger records. Note: notes/t-israeltech-legacy-leadership.md.
• ★★★ | Overcoming Personal Challenges to Build Pitango — Leap Academy #98 | https://www.leapacademy.com/episodes/chemi-peres | 2025-04-22 | one-hour podcast/show notes | Chapters and a third-party automatic transcript provide navigation; unverified transcript wording is not quoted. Note: notes/t-leap-academy-98.md.
• ★★ | Mizuho Conversation with Leaders — Chemi Peres | https://www.mizuhogroup.com/beyond-the-obvious/conversations-with-leaders-a-sit-down-with-chemi-peres-managing-partner-and-co-founder-of-pitango | page 2022-12-13; Brightcove asset 2022-12-20 | 37:54 institutional video | Locally generated ASR timestamps vehicle specialization, transformation, regional scale, ESG, and LP-network claims; Brightcove asset 6317562867112 has no captions. Note: notes/t-mizuho-conversation-leaders.md.
• ★★ | 30 Minutes or Less — Chemi Peres Continues to Dream | https://www.calcalist.co.il/podcast/article/bylrycmwzg | 2026-06-07 | 46:56 Hebrew podcast | Timed third-party ASR navigates formation, partnership, AI/quantum, and dual-use mandate evidence; Hebrew original, labeled translation, and ASR limits explicit. Note: notes/t-calcalist-continues-to-dream.md.
• ★★ | Peace through Innovation | https://www.filepicker.io/api/file/VF9lElwhTA2zSuEyFZH4 | 2020-12-21 | 14-page Lockdown University interview transcript PDF | Direct impact/peace framework with printed-page and speaker-turn locators; host prompts remain separate from Peres answers. Note: notes/t-peace-through-innovation.md.
• ★ | Common Good: Shimon Peres memoir discussion | https://podcasts.apple.com/us/podcast/chemi-peres-on-memoir-written-by-his-father-shimon/id1577634432?i=1000531221733 | 2017-09-18 | podcast episode | Family/intellectual influence; low priority for venture process.
• ★ | Reinhard Mohn Prize laureate discussion | https://www.bertelsmann-stiftung.de/en/media-center/media/mid/reinhard-mohn-prize-2020-highlights-from-the-award-ceremony | 2021 | official award-video highlight | Live official route confirms the appearance, but a complete stable transcript/caption record was not recovered; do not infer unseen speech.
• ★ | WeTechBerlin conference agenda | https://newmedia.calcalist.co.il/conferences/2019/berlin/assets/agenda.pdf | 2019-02 | conference agenda PDF | Event/date verification for reported remarks; no content evidence by itself.
• ★ | Global Financial Innovation conference guide | https://www.sachsforum.com/uploads/5/1/9/6/51964431/gfi-conference_guide.pdf | 2019-02-26 | conference guide PDF | Search-indexed appearance lead, but the direct PDF returned 404 during REVIEW on 2026-08-22; acquire a legitimate archive copy and recording before deriving content.
• ★ | Herzliya Conference 2017 speaker dossier | https://www.runi.ac.il/media/aj4n5wi5/speakers2017a.pdf | 2017 | conference speaker PDF | Appearance and title lead; not substantive direct voice.

Chronology, vehicle formation, and historical context

• ★★★ | Kalish, Peres Graduate from Shrem-Fudim-Kelner Incubator | https://en.globes.co.il/en/article-383138 | 2000-02-29 | contemporaneous financial press | Essential distinction: Polaris I raised $20 million in 1993; Peres joined Polaris II management, raised $125 million in 1996.
• ★★★ | Polaris 2 Fund Raises $40 Million More | https://en.globes.co.il/en/article-367952 | 1997-03-31 | contemporaneous financial press | $124 million fund size, management fees, Dovrat–Shrem rights, and Kalish/Peres manager attribution.
• ★★★ | Significant Share: Polaris III and Eucalyptus | https://en.globes.co.il/en/article-384678 | 2000-03-20 | contemporaneous financial press | Kalish founded Polaris in 1993; legacy-fund and team combination; direct Peres comments on cooperation.
• ★★★ | Chemi Peres, Pitango | https://en.globes.co.il/en/article-522702 | 2001-09-16 | long-form profile | Explicit Mofet 1992–1996 and Pitango/former Polaris from 1996 chronology plus AudioCodes attribution lead.
• ★★ | Palarus: Making the Connections | https://www.forbes.com/1997/07/10/palarus.html | 1997-07-10 | contemporaneous profile | Early Polaris II formation and $125 million claim; title/transliteration and details need corroboration.
• ★★ | Pitango Closes $270 Million Sixth Fund | https://en.globes.co.il/en/article-1000882135 | 2013-10-03 | financial press | Fund VI amount and reported platform AUM; not net-performance evidence.
• ★★ | Pitango: Many IPOs Ahead | https://en.globes.co.il/en/article-1000884570 | 2013-10-09 | press-conference report | Direct market-expansion comments, fund cadence, and public forecast for later calibration.
• ★★ | Venture Capital—Start-up Co-evolution and Israel's High-tech Cluster | https://citeseerx.ist.psu.edu/document?doi=1da68ab4bf9d726160e7266925bebfc0b6b607de&repid=rep1&type=pdf | 2001 | academic working paper PDF | Contemporaneous industry-phase model informed by interviews including Peres; context, not personal performance proof.
• ★★ | Palestinian Technology Ecosystem report | https://www.belfercenter.org/sites/default/files/pantheon_files/files/publication/Report%20-%20Palestinian%20Technology%20Ecosystem%20vF2.pdf | 2022 | university policy report PDF | Takwin structure and regional ecosystem context; distinguish partners and mandates.
• ★ | Pitango Raises $300 Million Across Two New Early-stage Funds | https://www.calcalistech.com/ctechnews/article/b14dscfjzg | 2025-11-06 | financial press | Current stage/sector platform and reported assets; quotes are from current team leaders, not Peres doctrine.
• ★ | D&B Israel: Pitango Is Top VC Fund | https://en.globes.co.il/en/article-1000423862 | 2009-02-09 | financial press | Historical scale ranking and reported capital; not a return audit.

Filings, holdings, boards, deal attribution, and economics controls

• ★★★ | AudioCodes 2019 Form 20-F | https://www.sec.gov/Archives/edgar/data/1086434/000110465920024465/tm206469-1_20f.htm | 2020-02-25 | SEC annual filing | Printed p. 82 names Dovrat–Shrem/Yozma–Polaris 1993–1996 role; critical chronology control.
• ★★★ | Orckit registration statement | https://www.sec.gov/Archives/edgar/data/1401678/000095012307009751/y35652fv1.htm | 2007-09-24 | SEC registration statement | Peres biography, board appointment, fund entities, ownership, US office, and shared principal power.
• ★★★ | Taboola 2025 proxy statement | https://www.sec.gov/Archives/edgar/data/1840502/000114036125015810/ny20045512x3_def14a.htm | 2025-04-11 | SEC proxy | Board since 2013, current biography, committee service, and collective investment/voting authority.
• ★★★ | Taboola Schedule 13G/A | https://www.sec.gov/Archives/edgar/data/1840502/000121390025010716/xslSCHEDULE_13G_X01/primary_doc.xml | 2025-02-06 | SEC ownership filing | Four fund entities and six principals; shared power prevents sole-ownership claims.
• ★★★ | Via Transportation S-1/A | https://www.sec.gov/Archives/edgar/data/1603015/000160301525000012/viatransportationinc-sx1a.htm | 2025-09-03 | SEC registration statement | Board, capital, principal stockholders, selling holders, company metrics, and IPO facts.
• ★★★ | Via Transportation final 424B4 prospectus | https://www.sec.gov/Archives/edgar/data/1603015/000160301525000031/viatransportationinc-424b4.htm | 2025-09-15 | SEC final prospectus | Final offering and selling-stockholder table quantifies Pitango-affiliated shares sold and retained while preserving the multi-entity ownership boundary.
• ★★★ | Via Transportation Schedule 13G | https://www.sec.gov/Archives/edgar/data/1603015/000121390025102489/xslSCHEDULE_13G_X01/primary_doc.xml | 2025-10-27 | SEC ownership filing | Post-IPO check on the retained 5,132,117-share / 6.8% Pitango-affiliated position across nine reporting entities; not evidence of Peres's personal economics.
• ★★★ | Taboola 2025 Form 8-K | https://www.sec.gov/Archives/edgar/data/1840502/000114036125041711/ef20058675_8k.htm | 2025-11-12 | SEC current report | Records the repurchase of 6,387,648 shares from certain Pitango-affiliated funds for about $23.4 million; a partial realization, not a complete fund return.
• ★★★ | AudioCodes 2003 Form 20-F | https://www.sec.gov/Archives/edgar/data/1086434/000108643404000022/f20fyear2003.htm | 2004-06-30 | SEC annual filing | Issuer history and board record identify the DS Polaris/Chase/Genesis investment agreement and Eyal Kishon's governance role, constraining Peres-only attribution.
• ★★★ | Riskified 2025 Form 20-F | https://www.sec.gov/Archives/edgar/data/1851112/000185111226000007/rskd-20251231.htm | 2026-03-06 | SEC annual filing | Current Pitango entity holdings and nine-principal shared-power boundary.
• ★★★ | TScan Therapeutics Schedule 13G/A | https://www.sec.gov/Archives/edgar/data/1783328/000110465924011824/tm245444d1_sc13ga.htm | 2024-02-09 | SEC ownership filing | HealthTech entities and nine-principal shared voting/dispositive power.
• ★★ | Radware Form 3 | https://www.sec.gov/Archives/edgar/data/1145057/000162828017010298/xslF345X02/wf-form3_150906715804124.xml | 2017-10-26 | SEC ownership filing | Fund III entities and eight principals with shared power; individual beneficial ownership disclaimer.
• ★★ | Koor directors and management filing | https://mayafiles.tase.co.il/RPdf/72001-73000/P72857-00.pdf | 2003 | TASE/issuer filing PDF | Koor board role, Mofet 1992, Pitango/former Polaris, education, and other directorships.
• ★★ | Teva 2020 proxy statement | https://mayafiles.tase.co.il/rpdf/1291001-1292000/P1291508-00.pdf | 2020 | issuer proxy PDF | Teva board start, biography, 1996 platform framing, ecosystem and nonprofit roles.
• ★★ | Israel Amplify Program Corp. S-1 | https://www.sec.gov/Archives/edgar/data/1851833/000119312521096611/d125700ds1.htm | 2021-03-30 | SEC registration statement | False-positive control: the SPAC filing's Idomoo outside-directorship belongs to Rami Kalish, not Chemi Peres; it is neither an Idomoo offering nor evidence of a Peres-linked deal.
• ★★ | Rocket Pharmaceuticals prospectus | https://ir.rocketpharma.com/static-files/da2c53ed-fdc1-4c48-9b69-d9460f6b414c | 2026 | issuer prospectus PDF | Fund IV entities and indirect ownership/shared-power structure; Ittai Harel board attribution.
• ★★ | AudioCodes AGM notice 2022 | https://www.audiocodes.com/media/y2qmupt2/audiocodes-agm-notice-2022.pdf | 2022 | issuer proxy PDF | Eyal Kishon's 1993–1996 Dovrat–Shrem/Yozma–Polaris role; independent chronology control.
• ★★ | ComSovereign/Virtual NetCom merger agreement | https://www.sec.gov/Archives/edgar/data/1178727/000121390021012369/ea136683ex10-1_comsovereign.htm | 2021-02-26 | SEC transaction exhibit | Pitango Art Partners signature by Peres; entity evidence, not personal economics.
• ★★ | Retalix S-1 ownership filing | https://www.sec.gov/Archives/edgar/data/1058057/000089161803005063/f91283b3e424b3.htm | 2003-11-07 | SEC offering filing | Fund III entities and six-principal shared voting/dispositive power.
• ★★ | Varonis S-1 | https://www.sec.gov/Archives/edgar/data/1361113/000119312513405885/d587800ds1.htm | 2013-10-22 | SEC registration statement | Fund IV holdings and six-principal shared power; no individual deal-role proof.
• ★★ | Cerus registration statement | https://www.sec.gov/Archives/edgar/data/1619250/000119312516718758/d220271ds1a.htm | 2016-09-16 | SEC registration statement | Fund V selling holders and six-principal shared investment power.
• ★ | TopSpin registration statement | https://www.sec.gov/Archives/edgar/data/1211805/000089322007002982/w33351a2sbv2za.htm | 2007 | SEC filing | Fund III trusts, family/CEO funds, and eight-person control structure.
• ★ | 2026 AudioCodes Form 20-F filing index | https://www.sec.gov/Archives/edgar/data/1086434/000118518526001124/0001185185-26-001124-index-headers.html | 2026-03-30 | SEC filing index | Current issuer archive route; later-period business context, not early deal attribution.
• ★ | Via notice of effectiveness | https://www.sec.gov/Archives/edgar/data/0001603015/999999999525002928/xslEFFECTX01/primary_doc.xml | 2025-09-11 | SEC effectiveness notice | Exact IPO registration effectiveness date; no investment-return evidence.

Portfolio outcomes, skeptical evidence, press, and exclusions

• ★★★ | Start-Up AudioCodes to Issue on NASDAQ in 3 Months | https://en.globes.co.il/en/article-370514 | 1999-02-07 | contemporaneous financial press | Pre-IPO report places a 14% position with the Polaris/Chase/Genesis syndicate; it is a position snapshot, not realized fund proceeds.
• ★★★ | AudioCodes Sent to Wall Street | https://en.globes.co.il/en/article-375269 | 1999-05-31 | contemporaneous financial press | Post-IPO report values the Polaris position at $16.5 million; a paper mark rather than a cash-flow or Peres return.
• ★★★ | Reports of Chiaro's Death Are Greatly Exaggerated | https://en.globes.co.il/en/article-617689 | 2002-08-27 | edited Chemi Peres interview | Contains Peres's account of the GoRefer shutdown, vanished customers, shared stop decision, and cash-preservation logic; retrospective and not a fund-level loss schedule.
• ★★★ | מרבה כסף מרבה דאגה | https://www.globes.co.il/news/article.aspx?did=514475 | 2001-08-16 | contemporaneous Hebrew financial press | GoRefer financing and shutdown account reporting a roughly $10 million round and about one-third returned; timing and vehicle allocations remain inconsistent or undisclosed.
• ★★★ | Taboola Raises $15M Following Dramatic Growth in 2012 | https://www.prnewswire.com/news-releases/taboola-raises-15m-following-dramatic-growth-in-2012-191798941.html | 2013-02-19 | company financing release | Ex-ante Series D packet: Pitango-led round, growth, product, publisher, geography, and management signals; company-reported and not a valuation or ownership disclosure.
• ★★★ | Via Closes $27M Series B Funding | https://www.finsmes.com/2015/04/via-closes-27m-series-b-funding.html | 2015-04 | financing report | Contemporary operating packet and syndicate for the Pitango-led Via round; total round size is not Pitango's check or ownership.
• ★★★ | Tomorrow.io Expands Investment from Pitango and Harel Insurance | https://www.prnewswire.com/news-releases/tomorrowio-expands-investment-from-pitango-and-harel-insurance-increasing-series-f-to-210m-302774218.html | 2026-05-18 | company financing release | Current firm boilerplate claims 13 funds, $3 billion-plus AUM, 300-plus companies, and 91 exits/IPOs; self-report used only as a denominator boundary.
• ★★ | The Great Close | https://en.globes.co.il/en/article-803740 | 2004-06-09 | financial press / CalPERS analysis report | Early point-in-time Pitango III write-off and reported-IRR snapshot; neither the underlying LP schedule nor final DPI/TVPI.
• ★★ | Israeli VCs Wrote Off $425m by End 2005 | https://en.globes.co.il/en/article-1000101409 | 2006-06-12 | financial press / CalPERS analysis report | End-2005 Pitango III paper write-off and reported negative IRR; commitment, called capital, and portfolio marks use different denominators.
• ★★ | CalPERS Shows Carmel Ventures Vintage 2000 Top VC Fund | https://en.globes.co.il/en/article-1000543086 | 2010-03-01 | financial press / CalPERS analysis report | Contains a later reported Pitango III paper IRR; an as-of mark, not audited final performance.
• ★★ | CalPERS Report Shows Most Israeli Funds Still Losing | https://en.globes.co.il/en/article-1000710146 | 2011-12-26 | financial press / CalPERS analysis report | Later Pitango III paper write-off and return snapshot used only to show mark instability across dates and definitions.
• ★★ | CalPERS Again Shows Disappointing Returns on Israeli Funds | https://en.globes.co.il/en/article-1000762175 | 2012-07-02 | financial press / CalPERS analysis report | Subsequent paper write-off snapshot; cannot be reconciled into a final fund return without the underlying cash-flow schedule.
• ★★ | Polaris Sets Up $15 Mln General Managers Fund | https://en.globes.co.il/en/article-385270 | 2000-05-22 | contemporaneous financial press | Describes Pitango III's manager subfund and its relationship/loyalty rationale; allocation, conflicts, fees, and outcomes remain undisclosed.
• ★★★ | Wireless Logic Acquiring Webbing for Over $200 Million | https://www.calcalistech.com/ctechnews/article/bkrt6e11nn | 2023-08-09 | transaction press | Reports Pitango as primary investor and Peres as chair; company price is not investor proceeds or return.
• ★★ | Content Recommendation Firm Taboola to Acquire Perfect Market | https://www.wsj.com/articles/content-recommendation-firm-taboola-to-acquire-perfect-market-1407151801 | 2014-08-04 | transaction press | Taboola growth/acquisition context; access may be partial and does not prove investor causality.
• ★★ | Partner at Pitango Steps Down Amid Sexual Harassment Claims | https://www.timesofisrael.com/partner-at-israeli-vc-firm-pitango-steps-down-amid-sexual-harrassment-claims/ | 2018-03-15 | adverse press | Firm-governance context involving Rami Beracha; allegations are not findings against Peres.
• ★★ | Potential Conflict Questions Around the Peace Technology Fund | https://israelbehindthenews.com/2004/12/24/is-there-a-potential-conflict-of-interests-between-shimon-peres-and-the-operations-and-investments-of-the-peace-technology-fund/ | 2004-12-24 | advocacy/adverse article | Speculative lead only; do not state its insinuations as fact without filings or official findings.
• ★★ | Israeli Tech Back to Business, Cybersecurity Among Most Resilient | https://www.reuters.com/world/middle-east/israeli-tech-back-business-cyber-security-among-most-resilient-chemi-peres-2023-11-13/ | 2023-11-13 | Reuters interview | Current conflict-period industry views; separate ecosystem forecast from performance evidence.
• ★ | Israel's Venture-Capital Risk Calculation — 2007 conference program | https://www.innovationjournalism.org/blog/Globes%20Israel%20Business%20Conference%202007.pdf | 2007 | conference program PDF | Peres chaired risk panel; content cannot be inferred without recording/transcript.
• ★ | Congressional seminar speaker booklet | https://icpro.ch/hubfs/seminar-booklet-israel-2018.pdf | 2018 | event booklet PDF | Biography and appearance lead; no independent investment evidence.
• ★ | SEC EDGAR full-text search | https://www.sec.gov/edgar/search/ | live 2026 | regulatory database | Canonical route for Chemi/Nechemia/Peres, Pitango/Polaris/Mofet and issuer/entity variants.
• ★ | TASE Mayafiles search | https://maya.tase.co.il/reports/fund | live 2026 | exchange database | Mofet and Israeli filing route; exact Hebrew issuer/security identifiers remain a retrieval dependency.
• ★ | Internet Archive CDX search for Pitango and Chemi sites | https://web.archive.org/cdx/search/cdx?url=pitango.com/*&output=json&filter=statuscode:200&filter=mimetype:text/html | live 2026 | web archive index | Historical team, portfolio, and article snapshots; archived content needs snapshot-specific locators.
• ★ | Jim Breyer completed peer | ../030-jim-breyer/investmentphilosophy.md | updated 2026-08-21 | canon package plus underlying evidence | Institutional-platform, shared-credit, board-work, and public-filing comparator; Forbes 2011 controls Facebook/Accel credit and economics; geography/era differences limit equivalence.
• ★ | Yigal Erlich completed peer | ../035-yigal-erlich/investmentphilosophy.md | updated 2026-08-22 | canon package plus underlying evidence | Closest Israeli vehicle/ecosystem comparator; Japan Cabinet presentation, pp. 17–18 controls Yozma design and timing; no company-return inference.
• ★ | Trae Stephens completed peer | ../099-trae-stephens/investmentphilosophy.md | updated 2026-08-19 | canon package plus underlying evidence | Dual-use, public-mission, sponsor-led IC, and attribution comparator; 20VC preserves the direct-process boundary; current US defense context differs materially.

Coverage assessment

Discovery covered “Chemi,” “Hemi,” “Nechemia,” and “Nechemia Jacob/J. Peres,” plus Pitango, Polaris, Mofet/Moret misspelling, Yozma, Inbal, Dovrat–Shrem, biographies, writings, interviews, transcript, podcasts, YouTube, boards, filings, fund returns, losses, criticism, lawsuit, conflict, privilege, and portfolio terms. Direct routes attempted were Pitango, Chemi's LinkedIn articles/posts, YouTube, Apple Podcasts, McKinsey, Wharton, the Peres Center, Bertelsmann, SEC EDGAR, TASE Mayafiles, issuer investor-relations pages, academic/context archives, and the Internet Archive index.

The priority written denominator is seven unique works: six ★★★ essays and one ★★ first-person post. The CTech originals are canonical for “Calibrate Bigger” and “From Red Tape to Revenue”; mirrors are retrieval controls, not independent works. The priority spoken denominator is thirteen source-map items: eight ★★★ and five ★★ rows, representing twelve underlying appearances because the two Wharton pages split one 2012 interview. Ten underlying appearances were accessible in full, TechCrunch remained partial, and The Brothers remained metadata-only; Apple/YouTube or firm/podcast mirrors were not counted separately. Additional ★ event and archive leads require content acquisition before analysis.

SCOUT found no public LP statements, audited fund cash-flow schedule, portfolio-wide write-off ledger, career hit rate, reserve schedule, carry allocation, or personal economics record. Company valuations, acquisition prices, issuer holdings, AUM, committed capital, and “exits” cannot substitute. Candidate denominator routes are the live/historical Pitango portfolio, Mofet/TASE disclosures, each Pitango fund's issuer filings, archived team/portfolio snapshots, and company-level investment announcements; none is represented as a complete Peres-only career ledger.

Highest-value retrieval is exact Mofet security/issuer identifiers and prospectuses on TASE; Dovrat–Shrem/Yozma–Polaris I and Polaris II partnership materials; LP reports by vintage; complete investment/write-off/reserve ledgers; IC and board records for Shany, AudioCodes, Taboola, Via, Webbing, and ordinary or failed companies; full captions for the 2025 IsraelTech, Leap, Human Founder, Brothers, and McKinsey appearances; and archive snapshots of historical Pitango portfolios and biographies. No person, company, LP, or archive was contacted.

Source notes