Tommy Davis

1. Snapshot and testable edge thesis

Tommy Davis moved from Kern County Land corporate-development capital to the finite 1961–1968 Davis & Rock partnership and then to Mayfield, founded with the unrelated Wallace “Wally” Davis in 1969. The public record is concentrated in two inspectable direct-voice bodies—the 1979 congressional hearing and Davis-attributed passages in a 1981 Electronics feature—plus participant histories. Biography · written-corpus denominator · spoken-corpus denominator

Testable edge thesis. [researcher inference · moderate] In U.S. technology-company formation from the late 1950s through the early 1980s, Davis's repeatable advantage was relationship-based access to technical teams and patient capital, followed by selection for adaptable, customer-facing people and help connecting founders to managers, investors, and corporate partners. The proposed mechanism joins four steps: networks enlarge the opportunity set; repeated people assessment tests adaptability; equity recruits scarce builders; locked capital and selected introductions/governance give uncertain companies time and resources. 1981 Electronics, Davis passages, printed pp. 94, 98, 104 · 1979 hearing, printed pp. 134, 139–143, 157

The scope is narrow: capital-scarce technical companies inside Davis’s U.S. networks, not every sector, geography, vintage, or vehicle. Catalytica supports team sponsorship; Genentech supports a narrower Corning bridge; Tandem supports governance participation. Fund I’s weak reported result, the Maxfield initial-plan pass, unnamed failures, and succession risk are disconfirming controls. Cusumano, printed pp. 25–29 · Swanson, printed pp. 52–53 · Myers, printed pp. 15–22, 45–47 · hearing, pp. 141–143 · Maxfield, printed pp. 16–17

The denominator prevents stronger confidence: twelve named cases were searched and researched, 0/12 is sufficiently reconstructed end to end, all 12 retain a material unresolved field, and career coverage is not computable. DEALS reconciliation Unseen records should show repeated team references, customer-learning evidence, network actions that changed resource access, and equity/duration matched to milestones. The thesis fails if those mechanisms are absent from decisions, ordinary cases do not improve relative to comparable cases, or performance disappears after controlling for SDS/Teledyne, partner credit, access, vehicle, vintage, and luck. [researcher inference · strong]

2. Core philosophy

Davis's own public words establish a compact core. In 1981 he put people before the product, required founders to sell and remain close to customers, and linked new-company formation to teams leaving large employers. [investor-stated · strong] Electronics, Davis passages, printed pp. 94, 98, 104 In 1979 he described equity as the near-universal financing form of his partnership, employee ownership as a recruiting mechanism, non-withdrawable LP capital, and selection authority delegated to general partners. [investor-stated · moderate] Joint hearing, printed pp. 134–140, 157

Synthesis. [researcher inference · moderate] Back adaptable, market-facing builders; align them through meaningful ownership; finance uncertain formation with patient risk capital; and preserve accountable decision authority. The word “accountable” is an analyst safeguard, not Davis's reported language: the same record lacks a scoring rubric, pricing rule, reserve policy, and complete outcomes, while Myers describes early process as informal. Myers oral history, printed pp. 10–19

3. Sourcing

Operational rule. [researcher inference · moderate] Map every introduction from source to resource gained, and record who converted it into financing, hiring, diligence, or governance. Kill the “network edge” claim when no action or information advantage follows.

4. Picking

Davis's strongest inspectable selection rule puts people ahead of a fixed product, but the same article requires direct selling and customer contact. [investor-stated · strong] Electronics, printed pp. 94, 98 The original 1966 six-part manager screen is Nicholas-mediated and unavailable; it cannot be reconstructed as Davis's verified checklist. Nicholas, printed pp. 198–200 · Crisp archive route

A defensible modernized historical screen asks:

1. What prior behavior shows learning, judgment, truthfulness, recruiting, and adaptation?
2. What has direct customer contact changed in the plan?
3. Which technical, market, or economic fact remains fatal even for an excellent team?
4. Who investigated, sponsored, approved, and can reverse the decision?
5. What next milestone would update the people assessment?

[researcher inference · strong] Maxfield is the essential red-team case: four capable engineers did not override a services-heavy proposal and limited operating experience, while later product success cannot be leaked backward. Maxfield, printed pp. 16–17 Applied Biosystems suggests Davis could still ask basic product-utility questions under a tiny recalled market forecast, but the evidence does not reveal his final vote or terms. Alafi, printed pp. 65–66

5. Deal and portfolio mechanics

No IRR, TVPI, DPI, loss rate, hit rate, or Davis-person multiple is computable. [researcher inference · strong] The correct portfolio mechanic is therefore evidentiary before prescriptive: separate vehicles, rank economics directionally, and reopen selection claims against ordinary cases and failures. Pareto tables

6. Vehicle and incentive fit

Table-wide evidence boundary. [researcher inference · moderate] Each row synthesizes the cited investor or witness record; row-specific stronger or weaker labels govern where shown.

Vehicle fit. [researcher inference · moderate] Long-duration equity, GP discretion, and employee ownership fit capital-scarce company formation, while institutional fundraising and personal judgment made partner continuity consequential. Davis's illness exposes key-person risk; the absence of a governing agreement prevents claims about the actual response. Myers, printed p. 22 No modern ten-year term, “2-and-20,” reserve model, reporting standard, or key-person clause is projected backward.

7. Team attribution and work with founders

Table-wide attribution boundary. [researcher inference · strong] Roles are separated from company outcomes; row-specific investor or witness labels govern where shown.

[researcher inference · strong] Davis's best-supported founder work is mentoring, remembering people, connecting resources, persuading partners, and serving in at least one governance role. The record does not show a general CEO-replacement rule, recruiting system, board cadence, follow-on authority, or exit influence.

8. Views on the craft

• Power law. [researcher inference · strong] Davis did not publish a processed power-law theory; the likely dominance of SDS and absent winner-free record require outlier testing anyway. Rock, pp. 40–41
• Fund size. [investor-stated · moderate] The hearing reports $16m deployed across three partnerships and a new $20m pool, but supplies no optimal-size doctrine. hearing, pp. 133–140
• Access versus selection. [researcher inference · strong] Networks widened capital and company access; no comparative funnel shows Davis selected better than equally connected peers. Sourcing evidence
• Governance. [investor-stated · moderate] GP authority and the Tandem board/audit role show delegated selection and company oversight; no general board doctrine survives. hearing, pp. 139–143
• Exits. [researcher inference · insufficient public record] Acquisitions, IPOs, stock distributions, and later marks appear case by case, but no sell/hold rule or cash ledger exists.
• Great investor. [researcher inference · moderate] The supportable standard is not fame: assemble access and patient capital, test people against market behavior, allocate roles honestly, and leave an institution that can decide without the founder.

9. Signature bets and decision process

The no-hindsight record contains five bounded decision exercises:

1. Watkins-Johnson — corporate capital before the independent fund.
2. Scientific Data Systems — an experienced team leaving a capital-constrained incumbent division.
3. Teledyne — strong operators and a broad acquisition plan, with Rock-centered attribution.
4. Catalytica — team sponsorship and analogy under missing technical/market evidence.
5. Maxfield initial proposal — the failure/miss lab; a capable team with a services-heavy plan and a later product pivot.

Each exercise freezes what was known, asks invest/pass, kill criteria, ownership/terms, and next-check evidence, then reveals the sourced decision/outcome. [researcher inference · strong] No case is presented as fully reconstructed: the denominator remains 0/12 end to end.

10. Mental models and maxims

The evidence supports five distinct rules; synonymous “people first” formulations are merged, while research hygiene is not misattributed to Davis.

1. Back adaptable people, but make the judgment falsifiable — direct Davis people priority, constrained by plan, customer, and counterexample evidence. [researcher inference · strong]
2. Keep founders in direct contact with customers — direct selling and learning rather than laboratory retreat. [researcher inference · moderate]
3. Match patient risk-bearing equity to uncertain company formation — equity, duration, and delegated GP judgment. [researcher inference · moderate]
4. Use meaningful ownership to recruit builders — employee ownership as a talent and alignment mechanism. [researcher inference · moderate]
5. Treat relationships as operating infrastructure, then prove the conversion — a researcher-reconstructed network rule with explicit attribution limits. [researcher inference · moderate]

The inaccessible 1966 factor list, “report the denominator,” and “allocate credit by role” are cataloged as excluded candidates in the evidence-gaps section.

11. Evolution and contradictions

Table-wide status. [researcher inference · moderate] This chronology interprets dated evidence without promoting sequence into causation.

The central contradiction is productive: Davis publicly simplified selection to people, while his vehicle testimony and case fragments require technology, market, financing, incentive, and governance conditions. [researcher inference · strong] A second contradiction is institutional: personal judgment helped create the firm, but illness made continuity dependent on partners and processes the public record cannot inspect.

12. Failures and limits

• Mayfield Fund I. [retrospective witness · moderate] Reported weak outcome and informal process; no company loss ledger or defined return measure. Full postmortem
• Maxfield initial services proposal. [retrospective witness · strong] A real bounded decline, but not evidence that Davis saw and rejected the later ROLM product plan. Decision lab
• Unnamed non-prosperers and failures. [investor-stated · strong] Failure frequency, capital response, governance, stop decisions, and lessons cannot be reconstructed. Hearing, p. 133
• Key-person risk. [retrospective witness · moderate] Davis's illness and Myers succession expose continuity risk, but no clause, portfolio handoff, or LP response is public. Myers, p. 22
• Selection and survivorship. [researcher inference · strong] Winners dominate the named record; 0/12 is end-to-end reconstructed and career coverage is not computable. DEALS denominator
• Era dependence. [researcher inference · strong] Scarce technical capital, close regional networks, and less formal early partnerships may have amplified relationship judgment in ways that do not transfer automatically. Historical context

13. Historical operating environment

Davis's career crossed three financing environments. Watkins-Johnson used corporate diversification cash when the alternative was not a mature independent VC fund. D&R pooled private LP capital for capital-constrained technical teams such as SDS. By 1979 Davis described institutional and individual LPs, locked capital, almost-all-equity investing, high bank borrowing costs, and public-company compliance burdens. Rock, pp. 36–41 · hearing, pp. 134, 139–143, 157

[researcher inference · moderate] Scarce risk capital increased the value of trusted intermediation; labor mobility made team formation possible; employee ownership could compensate recruits for leaving incumbents; and finite partnerships created both patience and an eventual liquidity boundary. The surviving sources do not provide contemporaneous startup survival, loss, ownership, or return base rates, so no benchmark is manufactured. DEALS historical context

14. Modern VC translation (as of 2026-08-21)

Analyst-created 2026 hypothetical. [researcher inference · provisional] A deep-tech team has strong technical references, no repeated customer learning, and a certification cycle longer than the next fund-liquidity pressure. The Davis-derived screen says “not yet”: specify paid design-partner evidence, an independent technical milestone, ownership/terms for staged risk, and a kill condition; approve only if the fund’s reserves and duration can carry the validated plan. No current-company example is claimed.

15. Comparison with completed peers

Comparison status. [researcher inference · moderate] Conclusions below are dimension-matched synthesis; the cited underlying sources, not canon completion alone, support the factual premises.

Georges Doriot is the relevant institutional antecedent: Davis reportedly encountered his influence through a brother who took Doriot's HBS course, although the underlying 1986 Davis interview is unavailable. Kenney/Florida route, discussion surrounding cited interview Arthur Rock is the closest control because he shared D&R, while his independent public record assigns roles that cannot be credited to Davis. Both peers are complete in the canon: Doriot profile · Rock profile.

Comparison conclusion. [researcher inference · moderate] Davis sits between Doriot's institution-building/company-development model and Rock's explicit longitudinal people diligence. His distinct surviving emphasis is the combination of team priority, founder-customer contact, employee ownership, and LP patience. But his evidence is thinner than either peer's, and the shared D&R record makes Davis-versus-Rock selection comparison especially confounded.

16. The strongest case against greatness

[researcher inference · strong] The skeptical thesis is that Davis's reputation is an artifact of privileged access, partner and firm outcomes, a few famous survivors, and an era in which scarce capital and close networks made introductions unusually valuable—not proof of repeatable superior judgment. SDS and Teledyne may dominate visible D&R economics, yet the strongest role evidence centers Rock; Genentech's surviving Davis evidence is a relationship/Corning bridge, not board or founding credit; Catalytica is one hindsight-exposed founder account. Rock, pp. 30–41 · Swanson, pp. 52–53 · Cusumano, pp. 25–29

The denominator and economics strengthen that challenge. Every observed case is materially incomplete; failure companies are mostly unnamed; Fund I was reportedly weak; aggregate D&R capital/distribution claims conflict; no Davis personal return exists. Selection success may therefore be survivorship, team contribution, vehicle design, or luck. [researcher inference · strong] DEALS denominator · vehicle table

[researcher inference · moderate] Era effects are also plausible: personal networks could dominate in a capital-scarce regional ecosystem, while relationship memory might encode exclusion as easily as informational advantage. The people maxim itself is under-specified, and Maxfield shows that strong engineers could still be declined. Davis's illness then exposed key-person fragility. Maxfield, pp. 16–17 · Myers, p. 22

Discriminating evidence. [researcher inference · strong] A full deal/pass funnel, contemporaneous IC records, partner-role logs, governing agreements, board minutes, cap tables, and cash flows could decide between the favorable and skeptical theses. The favorable thesis predicts repeatable people/customer/network mechanisms and better ordinary-case outcomes after controls; the skeptical thesis predicts credit concentrated in partners/access/outliers, no mechanism in losing cases, and no performance persistence without famous survivors.

17. Analyst study guide, glossary, and go deeper

Recommended reading path

1. Start with the biography and its falsifiable edge thesis.
2. Read the two direct-voice bodies through writings.md, then the participant countervoices in talks.md.
3. Use the three Pareto views and denominator before any famous-deal narrative.
4. Work all five no-hindsight decision labs, including the Maxfield failure/miss lab.
5. Apply the five checklists in mental-models.md, then return to the skeptical case.

Investor-specific glossary

Reusable checklists

Checklist status. [researcher inference · strong] These controls are analyst reconstructions from the Davis record, especially the 0/12 denominator, vehicle separation, direct-voice rules, and attribution map. DEALS denominator · Direct-voice evidence

Sourcing

• Name the introducer, prior relationship, incentive, and counterfactual access route.
• Record what information or resource changed because of the connection.
• Reject a “network edge” claim when the connection produced no observable conversion.

Diligence

• Test team learning, truthfulness, recruiting, and direct customer evidence separately.
• State fatal product, market, technical, legal, and economic facts.
• Preserve what was knowable then; keep later outcomes outside the decision freeze.

Investment committee

• Name source, sponsor, investigator, approver, dissent, and conflicts.
• Specify invest/pass, kill criteria, ownership/security, and next-check milestones.
• Reconcile the vehicle, reserves, duration, and key-person coverage before approval.

Governance

• Define board, audit, recruiting, follow-on, CEO, and exit authority by person.
• Build independent bad-news channels and pre-agree escalation/stop rules.
• Log introductions and mentoring without turning them into sole outcome credit.

Portfolio and review

• Keep corporate, fund, personal, board, and syndicate roles separate.
• Report ordinary cases and failures beside winners; recompute without the dominant outlier.
• Do not calculate return without cost, ownership/dilution, timing, and distributions.

Go deeper, ranked

1. 1979 hearing, printed pp. 133–143, 157, 159 — fullest Davis direct body on vehicle, incentives, policy, governance, and acknowledged failure.
2. 1981 Electronics, Davis passages pp. 94, 98, 104 — direct people/customer/mobility doctrine with strict speaker control.
3. Myers oral history, printed pp. 10–26, 36–38, 45–47 — strongest internal countervoice on roles, weak Fund I, process, fundraising, Genentech, and succession.
4. Rock Bancroft oral history, printed pp. 30–47 — D&R formation, roles, outliers, and incompatible economics.
5. Maxfield oral history, printed pp. 16–17 — sharpest bounded miss and people-rule counterexample.
6. Cusumano, printed pp. 25–29 — Catalytica application evidence with survivor and attribution limits.
7. Sources and retrieval leads — original 1966 speech, 1986 interview, vehicle documents, and complete portfolio evidence that could overturn the synthesis.

Pareto 80/20 — what drove the record

The public record cannot support a literal 80/20 split. The tables below rank only the observed evidence: fame is not economic contribution, company value is not fund profit, and neither a firm logo nor a partner's recollection proves Davis-person credit. [researcher inference · strong] The reported economic ordering is directional because no complete vehicle ledger, ownership history, dilution schedule, cash-flow series, or Davis capital account is public. Davis hearing, printed pp. 133–135 · D&R capital conflict, Rock HBS profile and transcript pp. 4–7

Ranked view 1 — economic contribution

Dominant-outlier test. Removing SDS does not yield a defensible winner-free ranking. Teledyne has a reported mark but disputed cost basis; Catalytica has only an unquantified “significant” return; Genentech has no public Mayfield cash flow. The record therefore cannot show whether the remaining portfolio was attractive or whether another unseen company drove distributions. [researcher inference · strong] Rock, pp. 40–41 · Cusumano, pp. 25–29 · Davis hearing, pp. 133–135

Ranked view 2 — historical importance

Ranked view 3 — decision-learning value

Evidence boundary and searched denominator

The search separated seven universes: Kern corporate-development capital; the finite 1961–1968 Davis & Rock partnership; Mayfield Fund I; Mayfield II; the exact-identity-unknown $20 million partnership Davis described in 1979; Davis personal positions; and board/syndicate relationships without proven investment. A company can appear in more than one narrative, but follow-ons within one company are deduplicated as one observed case. [researcher inference · strong] A later reference entry claims D&R invested in fifteen companies; separately, Davis reported 23 companies acquired for $8,343,607 in his then-current partnership over five years; those universes overlap neither transparently nor necessarily at all. later reference entry, D&R company-list paragraph · Davis hearing, pp. 133–135

Searches covered Davis/D&R/Mayfield name variants, the public hearing, SEC historical News Digests and EDGAR routes, company names, firm and obituary accounts, participant oral histories, Princeton/Stanford/HBS finding aids, criticism, losses, lawsuits, and namesake controls. No complete partnership portfolio, capital-call ledger, quarterly-report series, cap tables, board-minute set, or audited cash-flow schedule surfaced. annotated search frame

Coverage fraction: not computable. 12 is an observed-case numerator, not a defensible career denominator. D&R's reported ≈15 and the 1979 partnership's 23 are incomplete, differently scoped, and omit Mayfield years after the hearing; personal, board-only, and syndicate activity are also unresolved. Selection and survivorship bias are severe because named winners dominate the accessible record, while Davis acknowledged unnamed non-prosperers and failures. [investor-stated · strong] hearing, p. 133

Historical market, regulatory, and base-rate context

What was knowable at the decision dates

• Corporate capital before an independent fund. At Watkins-Johnson, Davis deployed Kern County Land's oil-royalty cash and Rock's employer handled financing. The relevant alternative was corporate diversification, not a modern ten-year VC fund. [retrospective witness · moderate] Rock, pp. 36–37
• Capital-constrained technical spinouts. Rock remembers SDS arriving through an offering circular after Packard Bell would no longer fund Palevsky's computing division; the team had operating history but needed new risk capital. That source is retrospective, yet it preserves decision-date facts distinct from Xerox's later acquisition. [retrospective witness · moderate] Rock, pp. 40–41
• Illiquid equity and delegated selection. In 1979 Davis described a seven-year partnership whose LPs could not withdraw capital or choose investments, and said GPs invested almost entirely through equity. He also cited 19–20% bank borrowing and public-company compliance costs as constraints on young firms. These are contemporaneous advocacy claims, not a neutral market survey. [investor-stated · moderate] hearing, pp. 134, 140–143, 157
• Emerging markets with weak forecasts. Alafi remembers a study projecting only about ninety sequencers by year five for Applied Biosystems. The study was not recovered, so the number is evidence of perceived uncertainty, not a validated base rate. [retrospective witness · provisional] Alafi, pp. 65–66

No contemporaneous dataset supplies a comparable startup survival, loss, ownership, or return base rate for these vehicles. Later exit values and firm histories are outcome evidence only; they are not inserted into the ex-ante information set. [researcher inference · strong]

Vehicle economics and incentives

Modern comparison (as of 2026-08-21). [researcher inference · moderate] The supported transferable dimensions are closed-end illiquid capital, delegated GP selection, equity incentives, and key-person risk. It is unsafe to project a modern ten-year term, 2-and-20 fee schedule, reserve model, institutional reporting standard, or key-person clause onto any Davis vehicle because none of those exact terms was inspected.

People, capital, and influence map

Across the remaining ledger, source, sponsor, diligence, approval, negotiation, board work, recruiting, follow-on authorization, and exit influence are unknown unless explicitly stated below. That blankness is a finding, not permission to credit the best-known partner. [researcher inference · strong]

Comprehensive deal ledger

Ex-ante forensic case studies and decision labs

Case 1: Watkins-Johnson — corporate capital before the fund

Information freeze (1958). Davis said he was a vice president of a wealthy California corporation and made his first technology-startup investment in 1958. Rock later connects that corporation, Kern County Land, to Watkins-Johnson and says its capital came from oil royalties; Hayden Stone handled public financing. [investor-stated · moderate] hearing, p. 133 · [retrospective witness · moderate] Rock, pp. 36–37

Known then: a technology company needed financing and Kern had diversification capital. Unknown then/publicly now: technical diligence, market size, alternatives, price, security, check, ownership, kill criteria, corporate dissent, milestone plan, board rights, reserves, and disposition policy. The later NYSE listing and sales cannot justify the initial decision. [researcher inference · strong]

Reader exercise. Would you invest then with corporate cash? What evidence would kill the deal? What ownership and governance fit an unproven technology company? What technical/customer milestone unlocks the next check?

Actual decision and outcome. Kern invested with Davis managing the activity; Davis later cited Watkins-Johnson's listing and >$100m sales. No inspected record establishes Kern's cash return or Davis economics. Rock, pp. 36–37 · hearing, pp. 133, 135

Case 2: Scientific Data Systems — capital-constrained team spinout

Information freeze (c. 1961). An intermediary brought D&R an offering circular. Rock says Max Palevsky had run Packard Bell's computer division successfully enough to form a relevant team, while trouble in Packard Bell's television business stopped further division funding. Palevsky then left with colleagues to form SDS. These are the strongest sourced ex-ante facts; they come from Rock decades later, not the offering circular itself. [retrospective witness · moderate] Rock, pp. 40–41

Known then: experienced technical leadership, a team, an incumbent funding constraint, and a reviewed circular. Unknown: product specification, customers, competing computers, price, check, ownership, partner debate, downside case, follow-on plan, and Palevsky succession risk. The 1964 SEC digest later identifies Rock as board chair, which constrains Davis governance credit but is not ex-ante evidence. [contemporaneous record · strong] SEC digest, p. 1

Reader exercise. Would you invest before product/customer evidence? What would kill the deal: team departures, no design advantage, or no customer commitment? What staged security/ownership fits the capital risk? What independent product and buyer evidence unlocks the next check?

Actual decision and outcome. D&R invested; SDS later went public and Xerox acquired it. Rock calls it probably the partnership's largest return, but the public record lacks investment cash flow, ownership, dilution, distributions, and Davis allocation. Rock, pp. 40–41 · hearing, p. 133

Case 3: Teledyne — strong operators with a broad strategy

Information freeze (1960–1961). Rock met Henry Singleton and George Kozmetsky before D&R. He remembers strong Litton reputations, Singleton's technical accomplishments, and a plan to acquire companies and improve operations rather than commercialize one fixed invention. Singleton and Kozmetsky committed to become D&R LPs; interests were later exchanged so D&R held Teledyne stock. [retrospective witness · moderate] Rock, pp. 30–33

Known then: operator history, a broad acquisition strategy, and alignment through the partnership exchange. Unknown: acquisition pipeline, valuation discipline, leverage limits, governance protections, D&R exchange ratio, Davis diligence, and succession plan. The later value and company success are held outside the freeze. Rock's account explicitly makes him—not Davis—the initial financier and board member. [retrospective witness · strong]

Reader exercise. Would you finance a team without a fixed product? What acquisition/leverage or key-person condition kills the deal? What ownership and board protections fit a roll-up? What first acquisition evidence unlocks more capital?

Actual decision and outcome. Rock financed Teledyne and served on its board; D&R acquired an interest. The obituary's $200k-to-nearly-$3m observation conflicts with Rock's denial of that initial figure and is not a distribution schedule. Rock, pp. 30–33 · LAT, paragraphs 3, 7

Case 4: Catalytica — people-first analogy under technical risk

Information freeze (reported 1983). Cusumano remembers a technical founding team seeking its first venture financing. He says Davis favored the team and persuaded partners by comparing the opportunity with Genentech; Mayfield reportedly supplied $3m. The contemporaneous memo, market analysis, cap table, and partner vote are missing. [retrospective witness · moderate] Cusumano, p. 25

Known then: founder account of a capable team, a proposed technical company, internal skepticism, and Davis's analogy. Unknown: what dimensions were actually analogous, customer evidence, technical milestones, valuation, security, ownership, reserves, dissent, and kill criteria. An analogy can organize judgment but cannot substitute for falsification. [researcher inference · moderate]

Reader exercise. Would you invest on team evidence plus analogy? What technical or market mismatch kills it? What milestone-based terms fit the uncertainty? What independent customer and process evidence unlocks the next check?

Actual decision and outcome. Mayfield reportedly financed Catalytica; Cusumano reports a 1993 IPO, later dispositions, and a significant but unquantified return. No Davis-person economics follow. Cusumano, pp. 25–29

Case 5: Maxfield team — a real decline with a false-negative boundary

Information freeze (fall 1968 and later fundraising). Maxfield says four engineers proposed customized military/systems engineering services. They lacked a repeatable product and, by their own later assessment, operating-management experience. He names Davis at Mayfield among several investors who declined. [retrospective witness · strong] Maxfield, pp. 16–17

Known then: technical founders, services-heavy plan, no identified repeatable product, limited management experience. Unknown: exact meeting date, Davis's words, internal sponsor/vote, requested terms, customer commitments, milestones, and whether Mayfield ever saw the later product proposal. The evidence therefore supports a pass on the initial plan, not “Davis passed on ROLM.” [researcher inference · strong]

Reader exercise. Would you fund the services plan? What evidence would kill it? Could milestone or founder-vesting terms bound management and product risk? What repeatable-product/customer proof unlocks the next check?

Actual decision and outcome. Davis/Mayfield declined the initial proposal; the team learned investors wanted products and searched for one. Later company history is counterfactual evidence, not proof the original decision was wrong. No documented process change is attributable specifically to this pass. Maxfield, pp. 16–17

Portfolio-level forensics

The observed 12-case sample spans one Kern case, four D&R cases, and seven Mayfield-era cases. It clusters in electronics, computing, telecommunications, instrumentation, and technology-enabled industrial businesses, with later biotechnology/life-science exposure. That pattern is descriptive of named public survivors and filings, not a complete sector allocation. [researcher inference · moderate] hearing, pp. 133–136 · Mayfield history, paragraphs 3–9

No IRR, TVPI, DPI, gross/net return, loss ratio, hit rate, check-size distribution, or Davis-person multiple is computed. Entry/exit valuations without ownership, dilution, timing, and cash flows are deliberately kept separate. [researcher inference · strong]

Operating rules from the observed record

1. Do not merge vehicles to manufacture a track record. [researcher inference · strong] Keep Kern, D&R, Mayfield I, Mayfield II, the unnamed 1979 $20m partnership, personal positions, boards, and syndicates separate until legal documents establish identity and economics. hearing, pp. 133–140, 157 · Myers, pp. 14–17
2. Test people through a falsifiable plan. [researcher inference · strong] The Maxfield decline shows that a capable technical team did not override a services-heavy model and missing management experience; it does not establish a pass on the later product company. Maxfield, pp. 16–17
3. Allocate credit by role, not fame. [researcher inference · strong] Record source, sponsor, diligence, vote, terms, board, follow-ons, and exit separately: Rock’s Teledyne/SDS work, Davis’s Tandem board role, and the narrower Genentech Corning bridge demonstrate why. Rock, pp. 30–33, 40–41 · hearing, pp. 141–143 · Swanson, pp. 52–53
4. Report the denominator before the winner. [researcher inference · strong] Davis’s own admission of failures and the weak Fund I recollection require unnamed losses and missing cash flows to remain beside SDS, Teledyne, and Catalytica. hearing, p. 133 · Myers, pp. 15–19

Failure, miss, and anti-portfolio postmortems

Mayfield Fund I — weak vehicle evidence, not a defined return

Original thesis and evidence. Myers describes early Mayfield as people- and intuition-centered, with Tommy and Wally making decisions while he investigated management. The fund was reportedly $3.5m. [retrospective witness · moderate] Myers, pp. 10–19

Warning, response, outcome. Myers calls Fund I unsuccessful and remembers only “a couple of percent,” but does not say whether that is total, annual, gross, net, realized, or marked. He says later partner meetings and process became more structured as younger partners gained influence. That sequence does not prove Fund I caused the change or that process alone improved results. [retrospective witness · moderate] Myers, pp. 15–19

Unresolved lesson. The evidence falsifies “people-first intuition guaranteed performance”; it does not identify losing companies, escalation decisions, write-offs, reserves, governance responses, or a documented Davis lesson. [researcher inference · strong]

Maxfield initial services plan — pass, learning, and counterfactual

The full decision lab reconstructs the only named Davis/Mayfield decline with a clear proposal boundary. Early warning signals were a services model and missing management experience; Mayfield declined rather than escalating capital. The founders later changed direction. Exact Davis rationale and any Mayfield process change remain unknown, while calling it a ROLM pass would leak the later product into the original decision. [retrospective witness · strong] Maxfield, pp. 16–17

Unnamed non-prosperers and failures — denominator evidence without invented stories

Davis contemporaneously acknowledged that quite a few financed companies had not prospered and some failed. He supplied no names, original theses, warning signals, follow-ons, board responses, stop decisions, losses, or lessons. [investor-stated · strong] The correct conclusion is that failure existed but cannot be reconstructed company by company from this record. hearing, p. 133

Key-person and succession risk

Myers recalls Davis becoming ill and inactive around 1983–1984 and Myers becoming managing partner around 1984–1985. No agreement reveals a key-person clause, carry transfer, portfolio reassignment, or LP response. [retrospective witness · moderate] This is vehicle-governance risk, not a medical diagnosis and not proof of a portfolio loss. Myers, p. 22

Uncertainty and retrieval ledger

Until those records surface, the strongest conclusion is narrow: Davis operated across corporate, partnership, board, and network roles; SDS and Teledyne likely drove much of the visible D&R reputation; Catalytica ranks above Genentech in the observable Mayfield economic record because it at least has an explicit—though unquantified—return claim; and every one of the 12 observed cases still lacks a material field. [researcher inference · strong]

Coverage and evidence base

The priority written/direct-voice denominator contains five source families: two full public bodies, two secondary-trace-only bodies, and one catalog-only body. Two families are deeply analyzable: the 1979 prepared statement/testimony and the 1981 edited magazine interview. The 1966 speech and 1986 interview are preserved only through archival/scholarly traces; the 1989 memoir is catalog-only. Participant oral histories, firm histories, obituaries, filings, and scholarship are supporting controls, not extra Davis writings.

Two additional full ★★ participant-authored writings are processed as behavioral controls: Cusumano on Catalytica and Blank on a MIPS-era encounter. They are excluded from the five-family Davis-authored/direct-voice denominator and from the count of two deeply analyzable Davis bodies. Cusumano note · Blank note

The public record is unusually thin for a canon investor. Depth therefore comes from keeping exact speaker, vehicle, and access boundaries—not from treating later Mayfield language or partner memories as additional Davis prose.

Source-by-source analysis

1979 — statement on innovation, startup growth, and survival

Context and access. Davis appeared with Greylock's Daniel Gregory before joint congressional hearings on innovation and small technology firms on 1 November 1979. His prepared statement occupies printed pp. 133–138, oral testimony continues through pp. 138–143, and later Q&A on fund sources and mechanics appears at pp. 157 and 159. Gregory's printed pp. 144–154 are a separate speaker. Joint hearing, printed pp. 133–143, 157, 159

Central argument. [investor-stated · moderate] Davis argued that private equity capital for young technology companies created innovation, employment, exports, and productivity, while tax and securities rules could weaken capital supply, employee ownership incentives, and long holding periods. This was public-policy advocacy supported by his own portfolio snapshot, not a neutral performance study. Prepared statement, printed pp. 133–138

Reasoning and evidence. He moved from personal experience—first technology investment in 1958 and roughly $16 million deployed with partners across three partnerships—to named company examples and a current-portfolio snapshot. He reported 23 companies acquired for $8,343,607 over five years, $132 million in aggregate company sales, about 3,100 employees, and several non-prosperers or failures. [investor-stated · moderate] The figures show what Davis chose to measure; they do not disclose company list, ownership, realized value, write-offs, fees, carry, cash flows, or Davis-person economics. Prepared statement, printed pp. 133–135

Assumptions and decision implications. Davis assumed that ownership incentives could recruit managers away from incumbent-company security and that long-duration equity was the normal instrument for fragile young firms. [investor-stated · moderate] The prepared statement and oral testimony place investment selection with the general partners and deny limited partners a selection vote or withdrawal during the seven-year term; later Q&A says there were three general partners and that the partnership invested almost entirely through equity. Prepared statement, printed p. 134; oral testimony p. 140; Q&A p. 157

Biography and intellectual development. The statement is the strongest contemporaneous bridge from Harvard/war service and Kern-era exposure to Davis's self-conception as a twenty-year venture practitioner. It also dates institutional LP participation and succession beyond the original D&R individual-capital network. [investor-stated · moderate] The testimony does not prove that the sequence caused his investment philosophy. Prepared statement, printed p. 133; Q&A, pp. 157 and 159

Contradictions and limits. The admission of failures qualifies celebratory firm histories. The employment/sales case favors surviving measurable companies, and Davis's policy position gives him an incentive to emphasize public benefits. [researcher inference · strong] No audited return comparison tests whether the structure delivered superior investor outcomes. Prepared statement, printed pp. 133–138

1981 — “Time is ripe for entrepreneurs”

Context and authorship. Martin Marshall wrote a multi-investor Electronics feature. Davis is one quoted source, not the author of its statistics, referral discussion, success rates, market framing, or other investors' views. Feature, printed pp. 93–104

Central argument. [investor-stated · strong] Davis stated that venture capital's principle was to back the right people and that products followed people. The two short quotations establish selection priority; they do not provide the mechanism for assessing people or evidence that the rule produced returns. Feature, Davis quotations, printed p. 94

Reasoning, evidence, and assumptions. The article supplies no Davis case study or denominator tied to the maxim. [researcher inference · strong] A defensible operational reading is that product uncertainty may be more adaptable than weak execution, but that mechanism is researcher reconstruction unless the 1966 speech or 1986 interview supplies Davis’s own reasoning. Feature, printed p. 94

Market and ecosystem claims. [investor-stated · moderate] On printed p. 98, Davis rejects product-builds-itself thinking and favors founders who stay directly with customers; on p. 104, he links new-team formation to employees leaving large companies. These passages add market proximity and labor mobility to the people rule, but the article gives no outcome test and Marshall’s surrounding generalizations remain separate. Feature, Davis passages, printed pp. 98, 104

Decision implications and contradictions. “People first” cannot mean “people only”: Davis's 1979 testimony repeatedly situates teams inside technology, market, financing, incentive, and regulatory conditions. [researcher inference · moderate] Maxfield's account of an initial Mayfield pass also shows that a people-oriented investor could reject a capable team when its first plan appeared services-heavy or diffuse. 1979 statement, pp. 133–138 · Maxfield oral history, printed pp. 16–17

Supporting witness writing — Cusumano on Catalytica (2013)

Context and access. James Cusumano’s founder memoir is not Davis authorship; its full author-posted chapter was inspected on 2026-08-20; that URL later failed, so the current linked licensed-book route is partial. Printed pp. 25–29 cover the 1983 transition from consulting to venture-backed products, the Mayfield financing, Davis’s remembered maxim, IPO, and later disposition. Balance, printed pp. 25–29

Argument, reasoning, and evidence. [retrospective witness · moderate] Cusumano reports that Mayfield supplied Catalytica’s first venture financing—reported as $3 million in 1983—and remembers Davis preferring an A team with a B idea. He says Davis persuaded partners through a Genentech/Catalytica analogy. This supports founder backing and partner persuasion in one survivor case; it does not prove the precise contemporaneous words, Mayfield entity, vote, ownership, or Davis-person economics. Balance, printed p. 25

Outcome, implications, and contradictions. Cusumano reports a 1993 IPO and calls the eventual Mayfield return significant without a cash-flow measure. [retrospective witness · moderate] The celebratory three-decade lookback is consistent with people-first behavior but is exposed to hindsight and survivor bias; named cofounders and directors also defeat lone-investor attribution. Balance, printed pp. 26, 29

Supporting witness writing — Blank on the investor–entrepreneur relationship (2011)

Context and access. Steve Blank’s first-person article, edited as told to Kevin Ohannessian, is a retrospective founder/employee account rather than Davis prose. The relevant passage sits under “The Investor/Entrepreneur Relationship.” Blank, paragraphs 1–5 under the heading

Argument, evidence, and decision implication. [retrospective witness · moderate] Blank remembers Davis recognizing him about six years after he had silently assisted an earlier Mayfield pitch, knowing his intervening company history, and showing personal attention during a MIPS financing visit. The evidence supports relationship memory, not sourcing, approval, pricing, board work, or outcome attribution. Blank’s broader claim that VCs serve founders belongs to Blank. Blank, paragraphs 1–5 under the heading

Assumptions and contradiction. [researcher inference · moderate] The episode is consistent with Davis treating people longitudinally, but Blank himself preserves the alternative that Davis used recognition strategically to win the deal. Approximate relative dates and missing meeting records prevent a stronger inference. Blank, paragraphs 1–3 under the heading

1966 — “How to Pick a Winner in the Electronics Industry”

Tom Nicholas reports a six-part manager screen in Davis's September 1966 speech, held through the Peter O. Crisp papers. The original body, exact wording, examples, and qualifications were not publicly accessible. [researcher inference · provisional] It is a priority retrieval lead rather than a content-analyzed primary source. Nicholas, VC: An American History, Google Books record and printed pp. 198–200 · Crisp papers

The speech’s date matters: it would predate the Mayfield brand and could test whether people-first selection was a D&R-era doctrine rather than later institutional memory. [researcher inference · moderate] Until retrieval, no checklist item is quoted, reordered, or called investor-stated. Nicholas, printed pp. 198–200 · Crisp papers

1986 — Kenney/Florida interview

Later scholarship bibliographically cites a 16 December 1986 Davis interview, but no transcript, recording, accession, exact wording, or pinpointable excerpt was inspected. [researcher inference · provisional] No substantive claim is therefore derived from the missing body. The family remains inventoried because recovering it could add late-career direct voice. Kenney/Florida scholarly route; exact interview locator absent

1989 — One Man's War

The privately published memoir is cataloged but its body was not inspected. It may resolve Davis's military unit, theater, duties, and chronology. It cannot support an investment rule or life-to-philosophy causal link merely from its title. WorldCat search route

The only independently inspected compact biographical control says Davis was an Army captain; a later Mayfield transcript's Vietnam reference is chronologically wrong. Los Angeles Times, paragraph 9 · Mayfield/Silicon Catalyst transcript, history exchange

Cross-source synthesis

The durable core: people, equity, and patient control

Across the two full bodies, three elements cohere. First, Davis publicly prioritized people in selection. Second, he described equity rather than loans as the normal financing instrument. Third, he located investment choice with general partners while keeping LP capital illiquid and non-controlling. [researcher inference · moderate] 1981 feature, p. 94 · 1979 statement/oral testimony, pp. 134, 140; equity answer p. 157

That combination suggests an operating logic: select adaptable teams, use ownership to align and recruit, and preserve a vehicle capable of waiting through technical/company formation. [researcher inference · moderate] It remains a reconstructed mechanism, not a demonstrated comparative edge. 1981 feature, p. 94 · 1979 statement, pp. 134–138

What the writings do not establish

The corpus does not disclose sourcing funnel, diligence sequence, pricing discipline, target ownership, reserves, concentration, board cadence, firing thresholds, exit rules, or company-by-company attribution. It also does not establish that people-first selection beat a market-first or product-first alternative. Firm and witness records must supply behavior; DEALS must test outcomes and misses.

Rhetoric versus observed boundaries

Davis acknowledged failures in 1979. Myers later recalled Mayfield I as producing “a couple percent,” without defining whether total or annual, and described the early process as informal; Maxfield remembered a capable team's first plan being declined. [retrospective witness · moderate] These controls prevent the 1981 maxim from becoming a universal success claim. Myers oral history, printed pp. 15–19 · Maxfield, printed pp. 16–17

Intellectual chronology, evolution, and contradictions

The apparent continuity from a 1966 manager screen to a 1981 people-first slogan is plausible but not yet proved from the original 1966 text. [researcher inference · provisional] Nicholas, printed pp. 198–200 · 1981 feature, p. 94 The 1979 statement complicates slogan-only readings by making incentives, capital duration, governance rights, regulation, technology, and markets part of the operating environment. 1979 statement, pp. 133–138

The sharpest tension is between selection confidence and evidence scarcity. Davis stated a clean people priority, yet the public writing record supplies no repeatable scoring method and no comparative denominator. [researcher inference · strong] 1981 feature, p. 94 Later phases must test behavior rather than fill that gap with admiration.

Decision playbook and modern translation

Rule 1 — evaluate the people, but specify reversible evidence

Historical rule. [investor-stated · strong] Back the right people; people make products. [researcher inference · moderate] Product adaptability is an analyst-supplied mechanism, not Davis’s stated causal explanation. 1981 feature, p. 94

Diagnostic questions. What has this team learned faster than peers? Which reference or work sample proves judgment, recruiting, truth-seeking, and adaptation? What evidence would reverse the assessment? Which product/market facts remain fatal even for an excellent team?

Boundary/failure mode. [researcher inference · strong] Charisma, reputation, or a warm introduction can masquerade as people quality; a people screen without explicit disconfirming evidence invites bias and hindsight.

Modern VC translation (as of 2026-08-21). [researcher inference · moderate] What transfers is testing learning speed, judgment, recruiting, and truthful revision. What does not transfer is using reputation or a warm introduction as the screen, or ignoring product, market, price, and compliance evidence. Likely misuse: paying any price because the founder is compelling. Analyst-created hypothetical: approve a pre-seed team only after three reference checks and an eight-week customer test; kill the next check if the team conceals contrary evidence or misses the agreed learning milestone.

Rule 2 — match long-duration equity to fragile company formation

Historical rule. [investor-stated · strong] Use equity almost entirely and preserve non-withdrawable LP capital while general partners select investments. 1979 statement/oral testimony, pp. 134, 140; equity answer p. 157

Diagnostic questions. Does the company need risk-bearing capital rather than debt? What milestone justifies the next check? What ownership and governance compensate for uncertainty? Does the fund have enough duration and reserves?

Boundary/failure mode. [researcher inference · moderate] Patient capital can become unaccountable capital without portfolio limits, reserve rules, valuation discipline, and stop criteria.

Modern VC translation (as of 2026-08-21). [researcher inference · moderate] What transfers is matching risk-bearing capital and fund duration to milestone uncertainty. What does not transfer is assuming illiquidity itself creates discipline or that every young company needs the same capital form. Observable signals are long validation cycles and staged technical risk; likely misuse is calling weak learning patient. Analyst-created hypothetical: fund a materials startup in tranches tied to an independent yield test, with the next check stopped if the test misses the pre-agreed threshold.

Rule 3 — use ownership incentives to recruit builders

Historical rule. [investor-stated · moderate] Davis argued that options could attract managers from secure incumbents and align them with young-company outcomes. 1979 prepared statement, pp. 135–138

Diagnostic questions. Which hires change the outcome? Is equity material after dilution and exercise cost? Do vesting, governance, and information rights reward durable value rather than short-term marks?

Boundary/failure mode. [researcher inference · moderate] Options do not repair bad role design, weak management, or impossible economics; they can also concentrate risk for employees.

Modern VC translation (as of 2026-08-21). [researcher inference · moderate] What transfers is giving outcome-changing recruits legible, meaningful ownership. What does not transfer is assuming option count alone creates alignment; dilution, exercise cost, liquidity, governance, and employee concentration risk must be shown. Likely misuse is substituting a grant for role design. Analyst-created hypothetical: compare a critical engineering hire’s fully diluted stake under three financing paths and reject the plan if management cannot explain value, vesting, and downside in plain language.

Rule 4 — report failures inside the denominator

Historical rule. [investor-stated · strong] Venture is high risk; some financed companies do not prosper or fail. 1979 prepared statement, p. 133

Diagnostic questions. What is the actual-deal/pass/write-off denominator? Which failure invalidates the people thesis? What did the partnership do after warning signals? Are company growth measures being confused with investor returns?

Modern VC translation (as of 2026-08-21). [researcher inference · strong] What transfers is reporting ordinary and failed cases beside winners; what does not transfer is substituting company sales, valuation marks, or power-law rhetoric for cash outcomes and actor attribution. Likely misuse is treating loss frequency as proof of a sound process. Analyst-created hypothetical: require each quarterly review to reconcile invested capital, ownership, write-offs, distributions, responsible partner, and next-check decision; reopen the thesis if the winner-free record fails its stated base case.

Rule 5 — require founders to meet the market directly

Historical rule. [investor-stated · moderate] Founders should recognize that products must be sold and stay close enough to customers to learn what they need. 1981 feature, printed p. 98

Diagnostic questions. Which founder conversations changed the product? What evidence distinguishes customer need from founder preference? Who owns selling before a repeatable motion exists?

Boundary/failure mode. [researcher inference · moderate] Founder contact can become anecdote-chasing; customer proximity does not repair a structurally small market, weak unit economics, or an infeasible product.

Modern VC translation (as of 2026-08-21). [researcher inference · moderate] What transfers is direct evidence that founders learn and sell with customers; what does not transfer is treating enthusiasm or raw meeting count as demand. Observable signals are changed priorities, paid commitment, repeat use, and honest loss analysis. Likely misuse is mistaking responsiveness for an unbounded custom-services roadmap. Analyst-created hypothetical: fund the next tranche only if three design partners convert to paid use under one product boundary; stop if every customer requires a different build.

Best writings and unresolved record

1. 1979 congressional statement — best body for vehicle mechanics, portfolio snapshot, public-policy beliefs, explicit failures, and Davis's own chronology; its figures remain self-reported.
2. 1981 Electronics feature — earliest accessible public anchor for the people-first maxim plus customer-facing and mobility claims; only Marshall-attributed Davis passages count.
3. 1966 “How to Pick a Winner” speech — potentially the earliest and most operational selection document; ranked for retrieval, not for uninspected content.
4. 1986 Kenney/Florida interview — bibliographically verified late-career interview route; content and transcript/audio unavailable.
5. 1989 One Man's War — useful primarily for biography and any explicit formative linkage; body unavailable and not presumed to be investment writing.

Unresolved work that could materially change this synthesis: the original Crisp-paper speech with exact box/folder; the Kenney/Florida recording/transcript; a lawful memoir copy; D&R and Mayfield letters, memos, annual/LP reports, and partnership agreements; and any Davis-authored board or company correspondence. Until those bodies surface, the written record supports a compact people/equity/patience doctrine with explicit risk—not a complete, validated investing system.

Coverage denominator and archive result

The public spoken record is small enough that the denominator needs two layers. The direct-voice/appearance layer contains six included priority routes: four independently confirmed Davis speech/interview families and two Western Association of Venture Capitalists roster entries whose delivery is not independently confirmed. Only 2/6 routes preserve inspectable Davis words: the 1979 hearing transcript and Davis-attributed passages in the published 1981 Electronics feature. The 1966 speech and 1986 interview are confirmed families whose bodies remain unavailable; the two WAVC items are metadata leads, not verified delivery.

The spoken-countervoice layer contains ten accessible participant oral histories. Those witnesses can test biography, deal roles, vehicle mechanics, failures, and succession, but they are not Davis appearances and do not enlarge the direct-voice numerator. Together the two layers contain 16 included priority spoken-corpus routes: 12 with inspectable content and four partial/metadata-only routes. Cusumano and Blank are written participant controls and remain in writings.md, outside both spoken denominators. Davis's One Man's War memoir and the Portraits of Success profile are likewise written/catalog routes, not spoken-source families.

Reasonable searches covered C-SPAN, YouTube, Internet Archive, American Archive of Public Broadcasting, Vanderbilt Television News Archive, Charlie Rose, Paley, Computer History Museum, Bancroft, HBS, Stanford, Georgia Tech, WAVC, podcast directories, and exact name/date/title variants. No authentic public Davis audio or film surfaced; results for the athlete, baseball player, politician, and Scientology namesakes were rejected. The absence of a public recording is a search result, not proof that no recording exists.

Dated guide to the Davis direct-voice routes

September 1966 — “How to Pick a Winner in the Electronics Industry”

Tom Nicholas identifies a Davis address to the Western Electronic Manufacturers Association in Palo Alto and reports a six-part manager screen. Nicholas, printed pp. 198–200 The HBS Peter O. Crisp collection is a general custody/request route, but its public landing page does not index Davis, this title, or an exact item. The original speech, box/folder locator, wording, examples, and qualifications were not publicly inspected. [researcher inference · provisional] Its pre-Mayfield date makes it the best test of whether people-centered selection was already a D&R-era doctrine; until retrieval, the reported checklist cannot be presented as a verified Davis quotation or reconstructed in detail. Crisp papers collection · source note

The 1985 book The New Venturers is an additional retrieval path, not a separate appearance. Later quotation trails attribute six selection dimensions to Davis, but the licensed page and provenance have not been inspected; the trail appears to relay the 1966 speech family. WorldCat OCLC 14954671 · trace note

1 November 1979 — congressional oral testimony and Q&A

This is the strongest surviving Davis spoken body. His oral testimony occupies printed pp. 138–143 and later answers appear on printed pp. 157 and 159; the prepared statement at pp. 133–138 is written evidence, and Daniel Gregory's pp. 144–154 belong to a different speaker. Davis described long-duration equity, general-partner selection authority, limited-partner non-control and illiquidity, employee ownership, regulation, and a newly raised $20 million partnership with three general partners. Those vehicle details are [investor-stated · moderate] and unaudited rather than an inspected partnership agreement. Joint hearing, printed pp. 138–143, 157, 159 · source note

The hearing also resists a pure success story. Davis acknowledged that some portfolio companies had not prospered or had failed, while using a selected current-company sales and employment snapshot to advocate policy changes. [researcher inference · strong] That makes the speech useful for understanding what he measured, but not for calculating fund returns, loss rates, or a Davis-personal record. Prepared statement and oral exchange, printed pp. 133–143

May 1980 — Atlanta high-technology venture conference lead

A Georgia Tech retrospective says Davis was the first California venture capitalist who agreed to come to a two-day Atlanta conference, after which other investors followed and the event drew 60 venture capitalists and 15 firms. It does not expressly confirm his attendance or preserve a session, remarks, recording, transcript, or investment follow-up. [researcher inference · provisional] This is evidence of a scheduled appearance and possible network-signaling role, not a verified Davis talk. Georgia Tech history, printed p. 34 · source note

25 August 1981 — edited Electronics interview feature

Martin Marshall's feature spans printed pp. 93–104 and contains three distinct Davis passage clusters on pp. 94, 98, and 104. Davis prioritizes selecting people, insists that founders must sell and stay close to customers, and connects entrepreneurial formation to employees leaving large organizations to form teams. [investor-stated · moderate] The feature is a complete published body but not a raw interview: no question sequence, recording, interview date, or full Davis transcript survives publicly. Electronics, printed pp. 93–104; Davis passages pp. 94, 98, 104 · source note

Speaker boundaries are essential. Referral sourcing comments belong to George Quist, John T. Neises, and Burton J. McMurtry; return categories and figures belong to Pitch Johnson or Marshall's reporting. Repetition of the people maxim supports continuity of expression, not comparative efficacy or sole-Davis application to every Mayfield company. Electronics, printed pp. 94, 102–104

1970s and 1980s — WAVC roster leads

WAVC's current historical page lists “Thomas Davis, Partner, Mayfield Fund” under 1970–1979 and separately groups Thomas Davis with Franklin Johnson Jr. and Frank Chambers under 1980–1989. The page supplies no exact dates, titles, programs, recordings, or transcripts, and the grouped entry does not establish a panel format. [researcher inference · provisional] Both remain priority retrieval leads with delivery unconfirmed. WAVC past-speaker roster · 1970s note · 1980s note

16 December 1986 — Kenney/Florida interview

Later scholarship bibliographically cites a personal interview with Davis on this exact date and includes it in broader source pools used for accounts of venture networks, limited partnerships, and Doriot's influence; without claim-specific footnotes, those accounts cannot all be attributed to Davis. The recording, transcript, notes, accession, exact wording, and permissions did not surface. [researcher inference · provisional] The derivative papers verify an interview family, not the contents of the missing body. Kenney/Florida scholarly route · source note

Participant countervoices — what they add and constrain

The countervoices create a less heroic and more useful account. Myers describes complementary partner roles—Davis as intuitive about chemistry and people, Wally Davis as the engineer, and Myers between them—and says later partners added process to an initially informal practice. He also calls the first Mayfield fund unsuccessful and remembers only a few-percent result without defining total versus annual return. Myers oral history, printed pp. 10–19

For Genentech, Myers and Swanson support a narrower contribution: Davis connected the company to Corning through Ned Olivier, while Myers says Mayfield had no board seat and acknowledges date uncertainty. [retrospective witness · moderate] That is valuable network help, not founding, sole sourcing, or governance credit. Myers, printed pp. 36–38, 45–47 · Swanson, printed p. 52

Recurring frameworks, evolution, and contradictions

People first, but not people only

The durable direct-voice signal is that Davis ranked the team ahead of a fixed product. The 1981 market passage adds an important constraint: the selected founder must sell, remain in customer contact, and resist retreating into the laboratory when sales disappoint. [researcher inference · moderate] The rule is therefore better stated as “back adaptable, market-facing people” than as unconditional founder worship. Electronics, Davis passages, printed pp. 94 and 98

The Maxfield account is a real boundary case. Maxfield recalls that he spoke with Davis at Mayfield and was turned down on an accomplished team's initial services/systems plan; the team later reformulated around products. It shows that people-first selection could still reject a team-plan combination, but it does not prove that Davis rejected ROLM as ultimately pitched. Maxfield oral history, printed pp. 16–17

Equity, delegated judgment, and patient capital

In 1979 Davis joined ownership incentives to a seven-year vehicle in which LPs did not select companies or withdraw capital and the GPs invested almost entirely through equity. [investor-stated · moderate] The implied mechanism is that ownership recruits talent, GP discretion enables uncertain bets, and illiquidity gives companies time. The public record does not establish pricing discipline, target ownership, reserves, carry economics, or company-level returns. Hearing, printed pp. 134, 140, 157

Networks as operating infrastructure

The Atlanta invitation story, Draper's syndication account, the Swanson/Myers Corning bridge, and Bowes's fundraising account suggest that Davis operated through reciprocal investor, founder, corporate, and LP networks. [researcher inference · moderate] Those links may expand access and help companies, but the evidence does not convert every introduction into sourcing, approval, board work, or return attribution. Georgia Tech, printed p. 34 · Draper, printed pp. 67–68 · Swanson, printed p. 52

Evolution and disagreement across time

The surviving sequence suggests continuity from a reported 1966 manager screen to the 1981 people maxim, but only the later statement is inspectable direct voice. The 1979 hearing broadens the philosophy beyond selection to incentives, capital duration, governance, and public policy. Myers then supplies the organizational counterpoint: an intuitive early method was later made more systematic, and a weak first Mayfield vehicle plus Davis's illness exposed process and key-person risk. [researcher inference · moderate] 1966 custody route · 1979 hearing, printed pp. 133–143, 157, 159 · Myers, printed pp. 15–22

Biography and decision evidence

• Kern to D&R. Rock recalls that Kern hired Davis for diversification, that Davis helped form Watkins-Johnson with corporate capital, and that their disagreement with Kern preceded D&R. The capital initially belonged to Kern, not Davis personally. Rock CHM, printed pp. 7–8
• D&R to Mayfield. The 1966 speech is located inside the D&R period; the 1979 and 1981 sources belong to Mayfield. The sequence supports continuity of attention to managers but does not prove unchanged diligence or vehicle mechanics.
• Fundraising and institutionalization. Myers says Davis was the fundraiser and that University of Rochester connections helped unlock Fund II; Bowes separately remembers the difficulty of raising an early institutional fund. Neither is a signed close record. Myers, printed pp. 15–16, 26 · Bowes, printed pp. 12–14
• Illness and succession. Myers remembers Davis becoming ill and inactive around 1983–84, after which Myers became managing partner. This is participant evidence of key-person dependence, not a medical diagnosis or exact incapacity date. Myers, printed p. 22

Decision playbook and modern translation

Historical rules and diagnostics

1. Test the team before freezing the product. [researcher inference · moderate] Davis directly prioritizes people, while the detailed 1966 factor list remains mediated. Test adaptability and execution with evidence rather than charisma, and kill the deal if “people first” becomes an excuse to ignore truthfulness or performance. Electronics, p. 94
2. Require direct market contact. [investor-stated · moderate] Ask who sells, what founders learned from recent customers, and whether weak demand produces market learning or only more laboratory work. Do not use this rule where the buyer is structurally remote without adapting the evidence test. Electronics, p. 98
3. Match ownership and duration to company formation. [researcher inference · moderate] Verify that equity incentives can recruit the needed team and that capital cannot be withdrawn before technical and market milestones. Reject the translation if fund duration, reserves, or stakeholder liquidity needs are incompatible. Hearing, pp. 134, 140, 157
4. Separate help from credit. [researcher inference · strong] For every introduction or syndication, record who sourced, approved, negotiated, governed, and financed. A Corning bridge is not a Genentech founding or board role. Swanson, p. 52 · Myers, pp. 36–38, 45–47

Modern VC translation (as of 2026-08-21)

[researcher inference · moderate] The transferable core is to test adaptable founders with observable customer behavior and to align the financing horizon with milestone uncertainty. A modern IC should ask whether its fund duration, reserve policy, liquidity promises, partnership process, and succession plan actually fit the company rather than assuming Davis’s seven-year 1979 structure transfers unchanged. Observable signals include founder-led customer discovery, learning after lost sales, explicit decision ownership, reserve math, and named key-person coverage. The likely misuse is turning “people first” into charisma-first underwriting or treating a famous introduction as evidence of investment skill.

Hypothetical illustration. A 2026 deep-tech team has strong technical references but no repeat customer conversations. Applying the supported rule would make customer-learning milestones—not founder prestige—the next diligence gate; capital duration and reserves would then be tested against the technical certification schedule. This is an analyst-created example, not a Davis deal.

Read these five first

No public Davis audio is available, so this ranks five distinct readable spoken-record bodies rather than pretending that inaccessible leads can be heard.

1. 1979 congressional hearing, printed pp. 133–143, 157, 159 — the fullest Davis body; separate prepared text, oral testimony, Q&A, and Gregory’s testimony. Transcript
2. 1981 Electronics, printed pp. 93–104 — edited direct voice on people, customers, and labor mobility; read with strict speaker control. Feature
3. Gib Myers oral history, printed pp. 10–22, 26, 36–38, 45–47 — the strongest internal countervoice on team roles, process, Fund I, fundraising, Genentech, and succession. Transcript
4. Arthur Rock Bancroft oral history, printed pp. 36–45 — the fullest D&R partner narrative; read skeptically on partner allocation and economics. Transcript
5. Bob Maxfield oral history, printed pp. 16–17 — the cleanest false-negative and boundary case for people-first selection. Transcript

Unresolved evidence and retrieval priorities

• Request the 1966 speech from HBS Baker Library with exact box/folder, access, and quotation-permission terms.
• Locate Kenney/Florida's 16 December 1986 recording, notes, transcript, accession, and reuse terms through the authors' papers or institutional archives.
• Obtain dated WAVC programs/newsletters and a Georgia Tech agenda or attendee list before upgrading those three leads to verified appearances.
• Search Mayfield, Stanford, Corning, University of Rochester, HBS, NARA, and company archives for Davis correspondence, filmed events, board materials, and vehicle records; log catalog access separately from body access.
• Preserve the public-record limit: the two inspectable direct-voice routes establish consistency of themes, not a complete career corpus, a causal biography, or investing efficacy.

Back adaptable people, but make the judgment falsifiable

• Rule — Prefer a capable, adaptable team to a fixed product thesis, but require observable execution, customer learning, truthfulness, and a plan-specific kill condition. [researcher inference · strong]
• Evidence status — The people priority is [investor-stated · strong]; the falsification procedure is [researcher inference · moderate], because Davis's public words do not supply a scoring rubric or reversal threshold. Electronics, Davis passage, printed p. 94
• Provenance and origin — A September 1966 speech may contain the earliest manager screen, but only Nicholas's mediated account is inspectable; the earliest accessible direct formulation is the 1981 feature. [researcher inference · provisional] Nicholas, printed pp. 198–200 · Crisp archive route
• Mechanism and reasoning — [researcher inference · moderate] Under product uncertainty, learning speed and judgment can allow a team to revise; customer and execution evidence constrain the risk that charisma or reputation masquerades as adaptability. Davis states the preference, not this complete causal mechanism. Electronics, printed pp. 94, 98
• Historical and vehicle context — Capital-scarce electronics and computing formation rewarded teams able to leave incumbents and recruit through equity; locked partnership capital could wait through formation. Those conditions are material, while a specific modern fund term is not inferred. [researcher inference · moderate] Electronics, p. 104 · hearing, pp. 134–140
• When to use — Use at pre-product or pivot-prone stages where team learning can change the plan; do not use it to excuse infeasible technology, absent demand, unlawful conduct, bad economics, or a price that destroys the investment case. [researcher inference · strong]
• How to apply it — Document consequential revisions, obtain multiple independent references on truthfulness and recruiting, and inspect direct customer-learning evidence and identify the fatal product/market fact. Invest only with a named next milestone; kill if evidence is concealed, the team cannot recruit a missing capability, or the agreed learning test fails. [researcher inference · moderate]
• Examples —
  • Catalytica — documented witness application. Cusumano remembers Davis backing the technical team and persuading partners with a Genentech analogy; the company later went public, but the account is survivor- and hindsight-exposed. [retrospective witness · moderate] Cusumano, printed pp. 25–29
  • Maxfield initial proposal — documented boundary/counterexample. Four capable engineers did not override a services-heavy plan and limited operating experience; Davis/Mayfield declined that proposal, not necessarily the later ROLM product. [retrospective witness · strong] Maxfield, printed pp. 16–17
  • SDS — researcher-applied illustration. Rock remembers an experienced team leaving a capital-constrained incumbent division; the facts fit the rule, but no source says Davis applied it or assigns him the decision. [researcher inference · moderate] Rock, printed pp. 40–41
• Evidence — Direct formulation comes from Electronics p. 94; market constraint from p. 98; Catalytica supplies application testimony; Maxfield supplies a pass; Myers independently describes Davis as intuitive about chemistry and people while exposing informal process and weak Fund I. Myers, printed pp. 10–19
• Limits, failure modes, and exception ledger — Charisma, familiarity, and hindsight can all be relabeled “people quality”; Fund I's weak reported result and Davis's admitted failures defeat determinism. [researcher inference · strong] Myers, pp. 15–19 · hearing, p. 133 Application evidence is one named survivor plus inferred/counterexample cases; the original 1966 speech, 1986 interview, Mayfield IC/LP reports, company memos, and named-failure files were routed through HBS Crisp, Kenney/Florida scholarship, firm/LP archives, filings, and exact name/date searches, but no additional public Davis-linked applications surfaced. Retrieval ledger
• Evolution and contradictions — The reported 1966 manager screen and 1981 maxim suggest continuity, but only the later body is direct. The 1979 emphasis on technology, markets, capital, incentives, and regulation and the Maxfield pass prevent “people only”; later partner process formalization suggests intuition did not remain sufficient. [researcher inference · moderate] hearing, pp. 133–143 · Myers, pp. 15–19
• Modern VC translation (as of 2026-08-21) — [researcher inference · moderate] What survives is evidence-based assessment of learning, integrity, and recruiting; what fails is charisma-first underwriting or ignoring price, market, technology, and compliance. Applicable signals are truthful revision and recruited complements; likely misuse is “great founder at any price.” Adapt by pre-registering falsifiers and next-check evidence. Analyst-created hypothetical: a pre-seed team advances only after references plus an eight-week customer test; concealment or no learning kills the check.

Keep founders in direct contact with customers

• Rule — Before a repeatable sales motion exists, founders must sell directly and convert customer contact into bounded product learning. [researcher inference · moderate]
• Evidence status — The direct-contact rule is [investor-stated · moderate] from one edited Davis passage; its use across named deals is mostly [researcher inference · provisional] because company decision records do not connect the maxim to multiple investments. Electronics, Davis passage, printed p. 98
• Provenance and origin — The earliest inspectable articulation is 1981: Davis warned against assuming the product sells itself and against retreating to the laboratory when sales disappoint. The feature is edited, not a raw interview transcript. Electronics, printed pp. 93–104; Davis p. 98
• Mechanism and reasoning — Direct contact shortens the loop from buyer objection to product and positioning change; bounded commitments distinguish need from politeness. [researcher inference · moderate] Contact alone fails if each buyer pulls the company into unrelated custom work.
• Historical and vehicle context — Young electronics and instrumentation companies faced uncertain uses and scarce distribution; the rule assumes founders can reach users and still change product. It is less direct where procurement, regulation, or long validation separates user, buyer, and approver. [researcher inference · moderate] Alafi's recalled tiny sequencer forecast shows perceived market uncertainty, not a verified base rate. Alafi, printed pp. 65–66
• When to use — Use during discovery, founder-led sales, product definition, or a stalled launch; do not use raw conversations as proof of willingness to pay or require technical founders to ignore regulated buying roles. [researcher inference · strong]
• How to apply it — Name user, economic buyer, and blocker; log changed beliefs after each meeting; demand paid or behaviorally costly commitment; define one product boundary; kill or reframe if repeated conversations produce neither convergent need nor commitment by the next milestone. [researcher inference · moderate]
• Examples —
  • Maxfield initial proposal — researcher-applied boundary. Investors declined a customized services/systems plan; the team later concluded investors wanted products. This illustrates why contact needs a repeatable boundary, but no source says Davis invoked the 1981 rule then. [researcher inference · moderate] Maxfield, pp. 16–17
  • Applied Biosystems — researcher-applied illustration. Davis reportedly asked what the machine did amid a very small recalled market forecast; product utility was still a basic diligence question, but founder-selling behavior is not documented. [researcher inference · provisional] Alafi, pp. 65–66
  • Catalytica — researcher-applied exception. Team sponsorship and analogy are documented retrospectively, while customer evidence is absent from the public account; this is a missing application test, not proof the rule was ignored. [researcher inference · provisional] Cusumano, p. 25
• Evidence — The model rests primarily on Electronics p. 98, bounded by the Maxfield and Alafi witness records and by Davis's 1979 description of young firms operating amid finance and policy constraints. hearing, pp. 133–138
• Limits, failure modes, and exception ledger — Three deal illustrations are grounded but none is a source-connected Davis application; support therefore remains provisional. The 1966 Crisp speech, 1986 Kenney/Florida interview, WAVC routes, Mayfield IC/board/LP reports, founder interviews, and exact customer/sales query variants were attempted or cataloged; no further public Davis-linked case emerged. Spoken retrieval priorities Founder contact can also overfit anecdotes, obscure structurally small markets, and turn products into consulting.
• Evolution and contradictions — The customer rule appears in 1981 after the broader 1979 vehicle/policy account. It qualifies the people maxim but no later direct body shows whether Davis formalized customer thresholds. [researcher inference · moderate] writings chronology
• Modern VC translation (as of 2026-08-21) — [researcher inference · moderate] Direct learning survives; raw meeting counts and enthusiasm do not. Signals are paid conversion, repeat use, convergent roadmap, and honest loss analysis; misuse is anecdote-chasing or unlimited custom work. Adapt by separating user/buyer/approver and predefining a product boundary. Analyst-created hypothetical: release the next tranche only if three design partners become paid users of one bounded product; stop if every customer requires a different build.

Match patient risk-bearing equity to uncertain company formation

• Rule — Finance uncertain company formation with risk-bearing equity and capital that cannot be withdrawn before agreed milestones, while giving identified GPs responsibility for selection. [researcher inference · moderate]
• Evidence status — Vehicle form and equity use are [investor-stated · moderate] from Davis's unaudited 1979 testimony; the claim that this design improved outcomes is [researcher inference · insufficient public record]. hearing, printed pp. 134, 139–140, 157
• Provenance and origin — D&R's 1961 seven-year LP is the earliest observed use, supported by partner/LP recollection; the fullest Davis articulation is the 1979 hearing's seven-year, non-withdrawable, GP-selected, equity-heavy pool. Rock, pp. 38–39 · Fialkov, pp. 88–91
• Mechanism and reasoning — [researcher inference · moderate] Equity absorbs technical/company risk; capital lockup prevents a withdrawal run; delegated GPs can act under uncertainty. The mechanism breaks without milestone discipline, reserves, truthful valuation, partner accountability, or a duration matched to company needs.
• Historical and vehicle context — Kern corporate cash, D&R, Mayfield I, Mayfield II, the unnamed 1979 $20m partnership, personal positions, and board/syndicate roles are separate. High borrowing costs and compliance burdens described in 1979 made equity more relevant, but Davis was an advocate, not a neutral market survey. hearing, pp. 140–143, 157
• When to use — Use when outcomes are uncertain, debt service is mismatched, milestones are staged, and the fund can remain committed; do not use equity by reflex where cash flows support cheaper capital or where fund liquidity precedes company validation. [researcher inference · strong]
• How to apply it — Identify the legal vehicle, capital-call and withdrawal rights, actual decision makers, check/security/ownership, milestone budget, reserves, and extension/key-person provisions. Kill if duration cannot cover the base case, decision rights are ambiguous, or no independent next-check test exists. [researcher inference · strong]
• Examples —
  • D&R/SDS — researcher-applied illustration. A finite pool financed a capital-constrained technical spinout; Rock, not Davis, is the documented board chair, and cash flows remain incomplete. [researcher inference · moderate] Rock, pp. 38–41 · SEC digest, p. 1
  • Catalytica — researcher-applied illustration. Mayfield reportedly supplied $3m to a technical team, but security, reserves, duration, and exact vehicle are absent. [researcher inference · provisional] Cusumano, p. 25
  • Mayfield I — documented failure boundary. A pooled vehicle with people-centered judgment was reportedly unsuccessful; patient equity did not guarantee performance. [retrospective witness · moderate] Myers, pp. 14–19
• Evidence — The hearing provides direct form/authority evidence; Rock and Fialkov independently support D&R term and reporting recollections; Myers supplies the weak Fund I control. Exact reported D&R capital/distribution measures conflict, so no multiple is calculated. vehicle forensics
• Limits, failure modes, and exception ledger — Patient capital can become unaccountable capital; illiquidity is not diligence, reserves, governance, or returns. Governing agreements, calls, quarterly reports, partner accounts, cash flows, and key-person terms were sought through D&R/Mayfield/LP archives, Princeton AC233, SEC routes, oral histories, and hearing files; no complete public package surfaced. Uncertainty ledger
• Evolution and contradictions — Davis moved from corporate cash to a finite small partnership and then later institutional pools; reported capital scale increased while Fund I underperformance and illness exposed process and continuity risk. [researcher inference · moderate] Myers, pp. 14–22
• Modern VC translation (as of 2026-08-21) — [researcher inference · moderate] Matching risk capital, duration, and milestones transfers; Davis's seven-year structure does not establish a modern standard. Signals are validation cycles longer than debt tolerance and explicit staged tests; misuse is calling weak learning “patient.” Adapt by inspecting current governing documents, reserve math, liquidity promises, and key-person coverage. Analyst-created hypothetical: finance a materials company in tranches tied to independent yield tests; stop when the pre-agreed threshold fails.

Use meaningful ownership to recruit builders

• Rule — Give outcome-changing employees ownership whose fully diluted economics are legible enough to justify leaving secure incumbents. [researcher inference · moderate]
• Evidence status — Davis's recruiting/alignment claim is [investor-stated · moderate]; company-level application and performance evidence is [researcher inference · provisional] because grants, dilution, exercise cost, vesting, and employee outcomes are not public. hearing, printed pp. 135–138
• Provenance and origin — The earliest inspectable articulation is Davis's 1979 prepared statement, where options help young firms recruit managers from established companies. SDS's broader employee-option experiment is a retrospective contextual case, not proof Davis originated or administered it. hearing, pp. 135–138 · Rock, pp. 46–47
• Mechanism and reasoning — Ownership may compensate for salary/security risk and align builders with durable enterprise value. [researcher inference · moderate] It requires understandable dilution, fair vesting, credible governance, and a plausible value path; otherwise it is opaque risk transfer.
• Historical and vehicle context — Davis linked employee ownership to labor mobility out of large companies in an era of young electronics firms and illiquid equity. Electronics, Davis passage, p. 104 Tax/policy advocacy in the hearing is period-specific and not a complete causal proof. [researcher inference · moderate]
• When to use — Use for scarce hires whose decisions affect enterprise value and who can understand the risk; do not substitute ownership for role clarity, fair cash pay, lawful disclosure, or management quality. [researcher inference · strong]
• How to apply it — Name the hire and outcome-changing responsibility; show fully diluted stake under base/downside financings, vesting, exercise/liquidity constraints, governance, and downside; kill or redesign if the candidate cannot explain the stake or if dilution makes the alignment claim immaterial. [researcher inference · moderate]
• Examples —
  • SDS — researcher-applied contextual case. Rock remembers broad employee participation in option ownership, but Davis's design/administration role and individual outcomes are unknown. [researcher inference · provisional] Rock, pp. 46–47
  • Entrepreneurial mobility — direct doctrine, not a company application. Davis linked new teams to employees leaving large organizations. [investor-stated · moderate] Electronics, p. 104
  • Mayfield I — failure boundary. The vehicle's reportedly weak outcome shows that equity incentives plus a people screen do not establish portfolio success; company grant evidence is missing. [researcher inference · moderate] Myers, pp. 14–19
• Evidence — Two independent Davis bodies establish the recruiting/mobility proposition; Rock provides SDS context; Myers supplies the weak-vehicle boundary. No grant-level primary record was found. writings synthesis
• Limits, failure modes, and exception ledger — This model falls short of three documented company applications and remains provisional. Searches/routing covered the hearing, Electronics, SEC issuer filings, SDS/Mayfield company and oral-history routes, the Crisp speech, Kenney/Florida interview, partnership/LP reports, and exact options/equity/recruiting queries; no Davis-linked grant, cap-table, or employee-outcome ledger surfaced. Sources coverage Options can also concentrate employee risk or reward short-term marks.
• Evolution and contradictions — The 1979 policy argument and 1981 mobility passage are consistent, but no later direct record shows how Davis adjusted for dilution, taxation, or failed companies. [researcher inference · moderate]
• Modern VC translation (as of 2026-08-21) — [researcher inference · moderate] Meaningful, legible ownership transfers; raw option count does not. Signals are role materiality, understandable fully diluted value, and aligned vesting; misuse is replacing job design or cash fairness with an opaque grant. Adapt by modeling multiple financing and liquidity paths. Analyst-created hypothetical: reject a critical-hire plan until management can explain stake, dilution, vesting, exercise burden, governance, and downside in plain language.

Treat relationships as operating infrastructure, then prove the conversion

• Rule — Build reciprocal founder, investor, LP, and corporate relationships, but credit a network edge only when a connection changes information, talent, financing, governance, or market access. [researcher inference · moderate]

• Evidence status — Network activity is supported by [retrospective witness · moderate] across independent witnesses; the rule and conversion test are researcher reconstruction, not Davis's stated maxim.

• Provenance and origin — Watkins-Johnson connected Davis and Rock before D&R; Rock says relationship capital supplied many LPs. [retrospective witness · moderate] Rock CHM, printed pp. 7–8 The later Georgia Tech invitation story and Mayfield syndication accounts show the pattern continuing, not its origin. Georgia Tech history, printed p. 34

• Mechanism and reasoning — Trusted connections may surface teams earlier, speed capital formation, improve reference quality, recruit scarce people, or connect companies to strategic partners. [researcher inference · moderate] The edge is causal only if the connection produces a resource or information change beyond prestige.

• Historical and vehicle context — A geographically concentrated, capital-scarce Peninsula ecosystem and small partnerships raised the value of personal trust and reciprocal syndication. Necessary conditions are scarce access and informative relationships; fame and proximity are incidental unless converted. [researcher inference · moderate] Draper, pp. 67–68 · Rock, pp. 36–41 · hearing, pp. 140–143

• When to use — Use in sourcing, references, recruiting, syndication, corporate development, and follow-on support; do not use relationship status as a proxy for independent diligence or allocate sole credit for collective work. [researcher inference · strong]

• How to apply it — Log source, prior relationship, incentives, information/resource transferred, responsible converter, and counterfactual path; require independent diligence; kill the “network edge” claim when no conversion occurs or conflicts are undisclosed. [researcher inference · strong]

• Examples —

  • Genentech/Corning — documented witness application. Swanson and Myers remember Davis as mentor and bridge through Ned Olivier; Myers also says Mayfield had no board seat. The conversion is a corporate relationship, not sole investment credit. [retrospective witness · moderate] Swanson, pp. 52–53 · Myers, pp. 45–47
  • Mayfield Fund II fundraising — documented witness application. Myers credits Davis as fundraiser and a University of Rochester connection with helping institutional fundraising; signed close evidence is absent. [retrospective witness · moderate] Myers, pp. 15–17, 26
  • MIPS encounter — documented witness boundary. Blank remembers Davis recognizing his history during fundraising; he also preserves the alternative that recognition was strategic. Investment entity, vote, and outcome remain unresolved. [retrospective witness · moderate] Blank, “Investor/Entrepreneur Relationship,” paras. 1–5
  • Atlanta invitation — network-signaling boundary. Davis's acceptance reportedly helped attract others, but attendance and speech delivery are unconfirmed. [researcher inference · provisional] Georgia Tech, p. 34

• Evidence — Independent sources support different conversions: Rock on relationship/capital origins Rock CHM, pp. 7–8; Draper on reciprocal syndication Draper, pp. 67–68; Swanson and Myers on Corning Swanson, pp. 52–53 · Myers, pp. 45–47; Myers on Rochester Myers, pp. 15–17, 26; Blank on relationship memory Blank, paras. 1–5; and Georgia Tech on invitation signaling Georgia Tech, p. 34. None supplies a comparative return study.

• Limits, failure modes, and exception ledger — Networks can reproduce privilege, create groupthink, hide conflicts, and turn introductions into exaggerated credit. Genentech's no-board boundary and Atlanta's unconfirmed attendance are explicit exceptions. Firm portfolios, partner papers, board/IC minutes, LP closes, Stanford/Corning/Rochester archives, and exact introduction/syndication queries were attempted or routed; economic conversion and rejected-network cases remain absent. Retrieval priorities

• Evolution and contradictions — The pattern runs from Kern/Rock through D&R capital, Mayfield syndication, Fund II fundraising, and later company bridges. [researcher inference · moderate] Increasing institutionalization may have made personal relationships more scalable, while Davis's illness made their key-person dependence visible. Myers, p. 22

• Modern VC translation (as of 2026-08-21) — [researcher inference · moderate] Relationship mapping and conversion logging transfer; insider prestige without incremental information does not. Signals are faster trusted diligence, an accepted critical hire, closed financing, or a strategic customer; misuse is equating proximity with skill. Adapt with conflict disclosure, independent references, and counterfactual access tracking. Analyst-created hypothetical: a warm corporate introduction counts as value only after a paid pilot or decision-useful diligence result, not when a famous name joins a meeting.

Evidence gaps and candidate models

• The inaccessible 1966 six-factor manager screen is a plausible model family, but the original body, exact wording, order, examples, and qualifications were not inspected. Nicholas's mediated account and the general Crisp collection route are insufficient to promote six separate rules. Nicholas, printed pp. 198–200 · Crisp collection
• “Report the denominator before the winner” and “allocate credit by role” are essential analyst controls derived from the incomplete record, not Davis mental models. They remain in portfolio operating rules.
• A universal pricing, ownership, reserve, governance, follow-on, or exit rule is unsupported. The hearing and case fragments do not provide enough repeated decisions; D&R/Mayfield agreements, IC/board records, cap tables, quarterly reports, and cash flows remain the necessary proof. Uncertainty ledger
• A Doriot-derived doctrine remains plausible but unpromoted: later scholarship says Davis credited Doriot through an unnamed brother, while the 1986 interview body is unavailable. Kenney/Florida scholarly route

What the evidence supports

The defensible thesis is narrower than a greatness narrative. [researcher inference · moderate] From the late 1950s through the early 1980s, Davis combined relationship access to technical teams and patient capital with an expressed preference for adaptable, customer-facing people; some introductions and governance work then helped companies acquire resources. The record supports pieces of that mechanism—Kern/Watkins-Johnson, Davis & Rock's finite vehicle, the Genentech/Corning bridge, Catalytica sponsorship, and Davis's Tandem board role—but no complete portfolio or cash-flow record proves superior repeatable selection. Rock, printed pp. 36–41 · Davis hearing, printed pp. 133–143, 157 and 159 · Swanson, printed pp. 52–53 · Cusumano, printed pp. 25–29

The direct record supports four bounded propositions. Davis publicly put people before a fixed product and separately required founders to sell and stay close to customers. Electronics, Davis passages, printed pp. 94 and 98 He described employee ownership as a recruiting instrument and a seven-year pool in which limited partners could not withdraw capital or select companies, while three general partners invested almost entirely through equity. Hearing, printed pp. 134–140 and 157 He also acknowledged unnamed non-prosperers and failures, so a winners-only interpretation contradicts his own testimony. Hearing, printed p. 133

Central-claim audit

This table covers every high-risk claim family required by the review gate. Pass means the adjacent source supports the bounded wording and locator; it does not upgrade a participant memory into an audited record.

What could falsify the thesis

The favorable thesis predicts more than famous winners. If Davis had a repeatable edge, contemporaneous files should show the same people/customer/network mechanism across ordinary investments and passes, with explicit changes in access, information, recruiting, financing, or governance. After controlling for vehicle, vintage, co-investor, and initial access, cases exhibiting the mechanism should have better milestone and cash outcomes than comparable cases without it. [researcher inference · moderate] The existing record cannot run that test because the searched set is not a career denominator and no case is reconstructed from thesis through distribution. Denominator reconciliation

The skeptical thesis predicts that partner roles, privileged regional access, and one or two outliers explain most of the visible reputation. That account gains support if D&R records assign decisive sourcing and governance mainly to Rock, if Mayfield's ordinary and losing cases show no repeatable Davis process, if the record excluding SDS is weak, or if role-adjusted results disappear after Davis's illness. Rock's documented Teledyne and SDS work, the weak Fund I recollection, and the incomplete winner-free record make this alternative credible now. Rock, printed pp. 30–41 · Myers, printed pp. 14–22

The review therefore rejects both a claim of proven greatness and a claim of proven failure. [researcher inference · strong] The supported conclusion is that Davis articulated and sometimes appears to have applied a useful company-formation philosophy, while the public record remains too incomplete to estimate personal or vehicle performance, repeatability, or causal superiority.

Attribution and denominator audit

The package keeps seven universes separate: Kern corporate capital; Davis & Rock; Mayfield Fund I; Mayfield II; the exact-identity-unknown $20 million pool described in 1979; Davis personal positions; and board/syndicate relationships. The reconciliation is mechanically consistent: two incompatible reported universes (approximately fifteen D&R companies and 23 current-pool companies) are never summed; twelve named cases are searched and researched; zero of twelve is sufficiently reconstructed end to end; six have bounded multi-fact evidence; all twelve retain a material missing field; four unverified D&R leads are excluded. Coverage fraction: not computable is therefore the only defensible output. D&R reference entry, company-list paragraph · Hearing, printed pp. 133–135 · full reconciliation

Deal, team, and vehicle cross-checks

Single-source cases remain single-source: Catalytica rests on Cusumano's founder memoir; the Maxfield initial-plan pass rests on Maxfield; Applied Biosystems rests on Alafi; and Mayfield Fund I's result rests on Myers. They stay witness-labeled, with the missing independent record stated instead of being promoted by repetition.

Link, locator, quotation, and number audit

A 2026-08-21 direct GET sample covered twelve high-risk URLs: the hearing, Electronics, Rock, Myers, Swanson, the three SEC News Digests, the licensed Cusumano book route, the official Berkeley Kenney/Florida paper, Mayfield history, and the Los Angeles Times obituary. All twelve resolved; the SEC returned 403 to generic automation but 200 with a browser-style contact User-Agent. The author-hosted Cusumano chapter had developed a TLS/404 failure, so every citation now points to the resolving licensed Everand book route and records that the full author excerpt was inspected on 2026-08-20 but is no longer public there. The broken CiteSeerX interview route was replaced by the official Berkeley paper, which supports the derivative Doriot claim at printed p. 24 and lists the 16 December 1986 interview at printed pp. 41–42. Cusumano licensed route · Kenney/Florida official paper, printed pp. 24 and 41–42

All 49 source-map rows were reconciled against the checkbox protocol: 29 are checked and 20 are open overall. The priority subset contains 36 rows; all 28 accessible/processed priority rows are checked, while eight priority rows remain open because the underlying body is unavailable or unprocessed: the original 1966 speech, the 1986 interview body, The New Venturers, One Man's War, the paywalled/robots-blocked New York Times obituary, Done Deals licensed body, Mallaby's licensed body, and Princeton AC233 materials. The other twelve open rows are lower-tier ★ retrieval leads. The Forbes article was inspected during review and is checked; the Kenney/Florida derivative paper, Nicholas preview, Terman finding aid, Mayfield legal page, and Swanson duplicate route are also checked without implying that the missing primary materials were recovered. Coverage assessment

The quotation scan found no quoted span above 25 words. Every numeric economic statement remains reported, self-reported, contemporaneous, or explicitly unaudited; no IRR, TVPI, DPI, hit rate, personal multiple, or unsupported literal 80/20 split is computed. Direct locators were re-opened for the 1979 hearing (prepared pp. 133–138, oral pp. 138–143, Q&A pp. 157 and 159), the 1981 feature (Davis clusters pp. 94, 98, 104), Rock, Myers, Swanson, and the three SEC records.

Reproducible ordinary-claim sample

The sample was fixed before checking as a two-cell stratified grid for each core document: one ordinary biographical/source/access/market fact and one ordinary role/outcome/late-record fact, explicitly excluding claims already in the central-claim table. This prevents a convenient cluster in the strongest files and yields sixteen ordinary claims across all eight documents.

Transfer limits for modern VCs

[researcher inference · moderate] Five ideas transfer as questions, not as timeless prescriptions: test adaptable people with falsifiers; require direct founder-customer learning; match risk-bearing capital and duration to uncertainty; make employee ownership legible; and measure a network by resource conversion. Davis's sources do not establish a modern fund term, fee/carry norm, reserve ratio, ownership target, option design, portfolio concentration, governance cadence, or return benchmark. Each model is therefore dated 2026-08-21, identifies changed assumptions and likely misuse, and uses an analyst-created hypothetical when no grounded current case is supportable. Mental-model field guide

The most important non-transfer is epistemic. A sparse pioneer record cannot justify a present-day investment because its surface story resembles SDS, Genentech, or Catalytica. The decision labs force a fresh invest/pass choice, kill criteria, ownership/terms, and next-check milestones before revealing the historical outcome. Five decision labs

Unresolved evidence

The remaining gaps are material but not concealed. They are exhausted or precisely routed public-record limits, not permission to waive a phase gate.

1. Complete vehicle records. D&R and Mayfield agreements, capital calls, quarterly reports, fees, carry, partner accounts, reserves, cash flows, and key-person terms would decide the economic and incentive story. Princeton AC233 and D&R/Mayfield/LP archives are the highest-value routes; the remote Princeton finding-aid transfer timed out during this audit. Vehicle retrieval ledger
2. Deal/pass denominator and role logs. A complete company roster, rejected proposals, IC memos, board minutes, cap tables, follow-ons, write-offs, and distributions are needed to test selection, the record excluding SDS, and Davis-versus-partner attribution. The current result remains 0/12 end-to-end and coverage not computable. Denominator
3. Original Davis bodies. The 1966 WEMA speech needs an exact Crisp box/folder and lawful body; the 16 December 1986 Kenney/Florida interview needs recording, transcript, notes, accession, and permissions. The New Venturers and One Man's War need licensed bodies. Direct-voice routes
4. Independent case controls. Catalytica, the Maxfield initial proposal, Applied Biosystems, and Fund I each rely materially on one participant account. Contemporaneous company, Mayfield, LP, and filing records could confirm, narrow, or overturn them.
5. Identity and formative chronology. Harvard class years, Army unit/theater/duties, NLRB matters, birth day/place, and the exact path from those experiences to investment beliefs remain unproved. The biography does not manufacture causation.
6. Link preservation. The original Cusumano author PDF is now unavailable; the licensed route preserves a lawful access path but not the former public excerpt. The official Berkeley paper preserves the 1986 interview citation and derivative Doriot claim, not the interview body. These boundaries must remain in future refreshes. Cusumano note · Kenney/Florida note

No unresolved item supports a stronger claim than the package makes. If new public or legitimately accessible material appears, reopen the relevant claim, vehicle, case, or model; do not treat this review as freezing history.

Twenty-point acceptance evidence

Source-ranking protocol

Priority is retrieval order, not endorsement: ★★★ must process, ★★ should process, and ★ is a lower-priority lead. Reliability A denotes Davis's words, participant testimony, or an authoritative record; B denotes strong journalism or scholarship; C denotes an interested retrospective, metadata-only route, or discovery lead. Checkboxes track body processing across phases: checked means the accessible body has been processed; open means retrieval or later-phase processing remains outstanding. Mirrors and later retellings do not create independent evidence.

Davis direct voice and written record

• ★★★ | Thomas J. Davis Jr. statement to joint congressional hearings on innovation and small technology firms | https://ipmall.info/sites/default/files/BAYHDOLE/3_DIV_SCAN/3118_001_OCR_DIV.pdf | 1979-11-01 | prepared statement and oral testimony; reliability A; full public scan | Prepared statement pp. 133–138, oral testimony pp. 138–143, and later Q&A pp. 157 and 159 cover Davis's self-reported chronology, deployment, companies, failures, portfolio snapshot, fund structure, incentives, and regulation; company sales and jobs are not investor returns.
• ★★★ | “Time is ripe for entrepreneurs” | https://www.worldradiohistory.com/Archive-Electronics/80s/81/Electronics-1981-08-25.pdf | 1981-08-25 | edited magazine interview; reliability A for attributed Davis words, B for framing; full public body validated from printed pp. 93–104 | Davis passages appear on pp. 94, 98, and 104: people-first selection, customer-facing market orientation, and entrepreneurial mobility; referral remarks and other speakers’ figures remain separate.
• ★★★ | “How to Pick a Winner in the Electronics Industry” | https://www.library.hbs.edu/special-collections-and-archives/collections/highlights/crisp-papers | 1966-09 | Nicholas-mediated speech report; reliability B for the description, HBS landing page C as a general collection route | Tom Nicholas, printed pp. 198–200 reports the speech and manager screen; the HBS landing page does not index Davis or an item, so retrieve an exact box/folder before quoting or treating the rules as direct evidence.
• ★★★ | Martin Kenney and Richard Florida interview with Thomas Davis | https://brie.berkeley.edu/sites/default/files/wp163.pdf | interview 1986-12-16 | scholarly interview citation; reliability A if transcript is recovered, currently metadata/quoted fragments only | The official Berkeley paper cites the interview at printed pp. 24 and 41–42 and attributes a Doriot-influence claim to it; transcript/audio custody, exact wording, broader content, and lawful access remain unresolved.
• ★★ | Western Association of Venture Capitalists past-speaker roster | https://wavc.org/past-speakers/ | 1970s and 1980s, exact dates unavailable | association metadata; reliability B for current historical listing, no spoken body | Two distinct Davis roster entries appear under 1970–1979 and 1980–1989; no title, program, recording, transcript, or independent delivery confirmation surfaced.
• ★★ | First Atlanta high-technology venture-capital conference | https://repository.gatech.edu/bitstreams/6ca8c561-14c4-4abf-966f-f4c941e7e86c/download | 1980-05 event; later Georgia Tech retrospective | institutional event history; reliability B for invitation/event, no Davis spoken body | Printed p. 34 says Davis agreed to come and the conference occurred; it does not expressly confirm his attendance or preserve a session or remarks, so this route is excluded from the appearance denominator.
• ★★ | John W. Wilson, The New Venturers | https://search.worldcat.org/title/The-new-venturers-%3A-inside-the-high-stakes-world-of-venture-capital/oclc/14954671 | 1985 | licensed-book/catalog and derivative quotation route; reliability B if body recovered, C for later traces | Retrieval lead for Davis-attributed selection criteria; current evidence appears to relay the 1966 speech and is collapsed into that event family rather than counted independently.
• ★★ | One Man's War—A Boston Lawyer on the Jungle Trails of Burma | https://search.worldcat.org/search?q=%22One+Man%27s+War%22+%22Thomas+J.+Davis%22 | privately published 1989 | memoir/catalog route; reliability A for Davis's account if a lawful copy is inspected, current access metadata-only | Military-biography route for unit, theater, duties, and chronology; it is not evidence of investment philosophy without explicit linkage.
• ★ | Portraits of Success profile of Thomas Davis | https://books.google.com/books?q=%22Portraits+of+Success%22+%22Tommy+Davis%22 | 1986 | Carolyn Caddes book/profile; reliability B, current access metadata/snippet only | Near-contemporaneous biographical and portrait lead; retrieve a licensed copy and distinguish Davis quotations from author framing.

Participant oral histories and spoken countervoices

• ★★★ | Gibson S. Myers oral history | https://digicoll.lib.berkeley.edu/record/103607/files/myers_gib.pdf | interviews 2008–2009 | official Bancroft transcript; reliability A for Myers's participant account | Printed pp. 10–19, 22, 26, 36–38, and 45–47 distinguish Tommy, Wally, and Myers; describe Mayfield I underperformance, informal early decision process, fundraising, Fund II, Genentech role boundaries, Davis's illness, and succession.
• ★★★ | Robert A. Swanson oral history | https://digicoll.lib.berkeley.edu/record/54355/files/swansonrobiveco00roberich.pdf | interviews 1996–1997, published 2001 | official Bancroft transcript; reliability A for Swanson's account; repository requests paraphrase rather than quotation | Printed pp. 52–53 call Davis a mentor and describe his Corning/LP introduction; this supports network help, not sole Genentech sourcing or board credit.
• ★★★ | An Evening with Arthur Rock | https://archive.computerhistory.org/resources/access/text/2012/05/102658253-05-01-acc.pdf | 2007-05-01 | official CHM transcript; reliability A for Rock's retrospective account | Printed pp. 7–8 and 13 cover Kern, Watkins-Johnson, Davis & Rock formation, investors, and Teledyne; exact vehicle capital and partner credit require controls.
• ★★★ | Robert “Bob” Maxfield oral history | https://archive.computerhistory.org/resources/access/text/2012/07/102702237-05-01-acc.pdf | 2004 | official CHM transcript; reliability A for Maxfield's account | Printed pp. 16–17 document Mayfield/Davis declining the team's initial services/systems plan; do not relabel it a pass on the later ROLM product company without the missing proposal sequence.
• ★★ | Arthur Rock, HBS Entrepreneurs Oral History | https://www.library.hbs.edu/special-collections-and-archives/exhibits/entrepreneurs/arthur-rock | 2001-03 | official HBS interview/transcript; reliability A for Rock, B for HBS framing | Printed pp. 4–9 cover D&R, SDS, Teledyne, partnership duration, and aggregate reported economics; HBS's $3 million-to-$100 million framing is unaudited and not a Davis-personal return.
• ★★ | Arthur Rock oral history, Bancroft Library | https://digicoll.lib.berkeley.edu/record/218509/files/rock_arthur.pdf | interviews 2008–2009 | official university transcript; reliability A for Rock's account | Printed pp. 36–45 give the fullest partner narrative of Kern, Davis, formation, SDS, vehicle term, and split; Rock cannot independently prove his own versus Davis's work.
• ★★ | Herman Fialkov oral history | https://digital.sciencehistory.org/works/8910jv49c#oh-t-p866 | interview sessions 2009–2010 | official Science History Institute transcript/audio; reliability A for an LP participant | Paragraphs oh-t-p866–905 and 1027–1035 / printed pp. 88–91 and 103–104 give approximate commitment/distribution, reports, vehicle term, and actor boundaries; no audited call/distribution ledger survives publicly.
• ★★ | William H. Draper III oral history | https://archive.computerhistory.org/resources/access/text/2019/03/102740503-05-01-acc.pdf | interviews 2008–2009 | official CHM/Bancroft transcript; reliability A for Draper's account | Printed pp. 67–68 identify Mayfield as Davis plus Myers and describe reciprocal syndication among early firms; not deal-specific Davis attribution.
• ★★ | William K. Bowes Jr. oral history | https://digicoll.lib.berkeley.edu/record/218438/files/bowes_william.pdf | 2009 | official Bancroft transcript; reliability A for Bowes's account | Printed pp. 12–14 describe six months raising a reported $3.5 million from Stanford, Berkeley, and institutions and preserve fundraising doubts; not a governing agreement or audited close.
• ★★ | Moshe Alafi oral history | https://digicoll.lib.berkeley.edu/nanna/record/218715/files/alafi_moshe_2014.pdf?registerDownload=1&version=1&withMetadata=0&withWatermark=0 | 2014 | official Bancroft participant transcript; reliability A for Alafi's account | Printed pp. 65–66 give Applied Biosystems fundraising and Davis's late questions; later filings must establish security, entity, board, and economics.
• ★★ | James A. Cusumano, Balance, chapter 1 | https://www.everand.com/book/898675078/Balance-The-Business-Life-Connection | 2013 | founder memoir; reliability A for Cusumano’s experience, B for remembered Davis words; full author excerpt inspected 2026-08-20, author URL now broken, current licensed route partial | Printed pp. 25–29 describe Catalytica financing, founder support, a Davis maxim, later company outcomes, and an unquantified “significant” Mayfield return; cash flows and partner allocation remain unknown.
• ★★ | Steve Blank on the investor–entrepreneur relationship | https://www.fastcompany.com/1767059/steve-blank-broken-relationship-between-investors-and-vcs | 2011-07-25 | founder/employee recollection; reliability A for Blank's memory, B for edited framing | Under “The Investor/Entrepreneur Relationship,” paragraphs 1–5, a MIPS-era encounter illustrates longitudinal attention; it does not establish that Davis sourced, led, or governed MIPS.
• ★ | Reid W. Dennis oral history | https://digicoll.lib.berkeley.edu/record/218448/files/dennis_reid.pdf | 2009 | official Bancroft transcript; reliability A for Dennis's account | Printed pp. 27–29 preserve pre-D&R conversations and prudent-man context; anecdotal and not a partnership record.
• ★ | Mark Bertelsen oral history | https://archive.computerhistory.org/resources/access/text/2018/11/102740324-05-01-acc.pdf | 2018 | official CHM transcript; reliability A for Bertelsen's experience | Printed pp. 9–10 credit John Wilson's Mayfield formation help but explicitly speculate about SDS proceeds and misdate the Xerox transaction; use as a lead/countervoice.
• ★ | Thomas Weisel, HBS Entrepreneurs Oral History | https://www.library.hbs.edu/content/download/60632/file/Weisel_Thomas.pdf | 2002 | official HBS transcript; reliability A for Weisel's account | Early-career witness says Davis made an introduction that led to Weisel's first investment job; network evidence only.

Firm histories, journalism, and obituaries

• ★★★ | Mayfield Turns 50 | https://www.mayfield.com/mayfield-turns-50-insights-from-5-decades-of-people-first-investing/ | 2019 | official firm retrospective; reliability C for self-description, B for leads | Founding narrative, claimed early portfolio, team practice, and motto continuity; never convert Mayfield-company attribution into Davis-person credit or current-firm practice into historical proof.
• ★★ | Thomas Davis; High-Tech Investor | https://www.latimes.com/archives/la-xpm-1990-09-17-mn-518-story.html | 1990-09-17 | contemporaneous obituary; reliability B | Strong compact identity control for death at 77 in September 1990, Harvard, Army captain, NLRB, Kern, D&R, Mayfield, reported Teledyne figures, and survivors; obituary economics still need primary records.
• ★★ | Thomas J. Davis, 77, Investment Executive | https://www.nytimes.com/1990/09/13/obituaries/thomas-j-davis-77-investment-executive.html | 1990-09-13 | obituary; reliability B; canonical page paywalled/robots-blocked, preserved route https://perma.cc/N694-L7BG | Independent obituary route for identity and chronology; canonical and Perma copies are one source family.
• ★★ | Founding Funders | https://www.forbes.com/asap/2000/0529/145_print.html | 2000-05-29 | paired Rock/Kleiner interview and business-press history; reliability B | D&R formation/LP and venture-industry context; principally Rock/Kleiner, not Davis direct voice.
• ★★ | Done Deals | https://openlibrary.org/books/OL8089991M/Done_Deals | 2000 | edited oral-history book; reliability B; licensed borrowing/catalog route | Rock and Mayfield chapters preserve first-person firm histories and Davis practice leads; use a licensed body, not an unauthorized mirror.
• ★ | Silicon Catalyst fireside chat with Navin Chaddha | https://www.mayfield.com/silicon-catalyst-fireside-chat-with-navin-chaddha/ | 2023 | later firm-leader transcript; reliability C | Useful evidence of institutional memory and its drift: it misnames D&R and wrongly places Davis in Vietnam; treat “driving Terman” and book/philosophy claims as retrieval leads only.
• ★ | Santa Fe Capital portfolio history | https://sfcapital.com/portfolio/ | undated, accessed 2026-08-20 | adviser retrospective; reliability C | Claims David Silver helped Davis raise a Mayfield fund and names institutional context; verify against close documents and Bowes/Wilson evidence.

Books and scholarship

• ★★★ | Tom Nicholas, VC: An American History | https://books.google.com/books/about/VC.html?id=uSmWDwAAQBAJ | 2019 | scholarly book; reliability B; licensed preview/library route | Printed pp. 194–204 synthesize Davis, the 1966 speech, D&R, and early Mayfield; follow its notes to the primary record and do not rely on unauthorized mirrors.
• ★★ | Sebastian Mallaby, The Power Law | https://books.google.com/books?q=Sebastian+Mallaby+The+Power+Law | 2022 | reported history with extensive interviews/notes; reliability B; publisher URL returned 404 on 2026-08-21, current route is catalog/licensed-book search only | D&R capital, LP letter, returns, and split narrative require source-note tracing; overlapping Rock accounts are not independent confirmation.
• ★★ | Martin Kenney and Urs von Burg, “Technology, Entrepreneurship and Path Dependence” | https://kenney.faculty.ucdavis.edu/wp-content/uploads/sites/332/2018/03/Paths-and-Regions-The-Creation-and-Growth-of-Silicon-Valley.pdf | 1999/2000 | academic chapter; reliability B | Printed pp. 140–143 give D&R capitalization, actor network, and reported aggregate outcome; reconcile differing $3.0m/$3.4m/$3.5m/$5m measures.
• ★★ | Martin Kenney and Richard Florida on Silicon Valley venture capital | https://brie.berkeley.edu/sites/default/files/wp163.pdf | 2000s public copy of scholarship using 1986 interview | academic analysis; reliability B | Key route to Davis's own unavailable interview, Doriot influence, limited partnership, and carry history; cited interview must be recovered for exact wording.
• ★ | Solving the problem of abundance: venture capital and asset-driven inequalities | https://researchonline.lse.ac.uk/id/eprint/125890/1/Solving_the_problem_of_abundance_venture_capital_and_the_making_of_asset-driven_inequalities.pdf | 2026 | academic critical history; reliability B | Strong skeptical context on elite capital, LP infrastructure, and incentive distribution; not evidence of Davis-specific misconduct or performance.
• ★ | Steve Blank, Secret History of Silicon Valley installment | https://steveblank.com/2009/08/10/ | 2009-08-10 | practitioner history; reliability B/C | Links Kern/Stanford microwave ecosystem, D&R, limited-partnership form, and 20% carry; verify every institutional/mechanical claim against primary documents.

Deal, vehicle, filing, and company records

• ★★★ | SEC News Digest — Scientific Data Systems offering | https://www.sec.gov/news/digest/1964/dig051264.pdf | 1964-05-12 | contemporaneous SEC publication; reliability A | Printed p. 1 identifies Rock as SDS board chair and offering facts, constraining lone-Davis governance claims; retrieve the prospectus and D&R ownership schedule.
• ★★★ | SEC News Digest — Continental Telephone secondary registration | https://www.sec.gov/news/digest/1966/dig052566.pdf | 1966-05-25 | contemporaneous SEC publication; reliability A | Printed p. 2 lists Davis & Rock as holder of 11,457 Continental Telephone shares received in the Telephones, Inc. merger and proposed for secondary sale; it gives no cost, sale proceeds, partner allocation, or Davis-personal economics.
• ★★★ | SEC News Digest — Automatic Drilling Machines offering | https://www.sec.gov/news/digest/1969/dig070169.pdf | 1969-07-01 | contemporaneous SEC publication; reliability A | Printed p. 3 reports that the Davis & Rock Liquidating Trust held 23.97% of 1,462,950 outstanding shares; it does not disclose acquisition cost, dilution, sales, distributions, or partner economics.
• ★★★ | Princeton Office of Vice President for Finance and Treasurer records, AC233 | https://findingaids.library.upenn.edu/records/PRIN_MUDD_AC233 | 1968–1974 folder span | official archival finding aid; reliability A for metadata; remote fetch timed out | One box is indexed “Venture Capital — Mayfield Fund, 1968–1974”; highest-value LP/partnership route, requiring item request and exact box/folder/access result.
• ★★ | Frederick Emmons Terman papers, SC0160 | https://oac.cdlib.org/findaid/ark%3A%2F13030%2Ftf029000zm | 1901–1978 | official Stanford finding aid; reliability A for metadata | Correspondence/minutes route for Davis, Stanford, Watkins-Johnson, and technology transfer; finding aid confirms Terman's W-J board service, not Davis mentorship mechanics.
• ★★ | Mayfield terms of use / entity boundary | https://www.mayfield.com/mayfield-terms-of-use/ | current, accessed 2026-08-20 | official firm legal notice; reliability A for current entity disclaimer only | Confirms “Mayfield” is a collection of separate legal entities; current structure cannot be projected backward without formation documents.
• ★★ | Genentech founder oral history and company financing record | https://digicoll.lib.berkeley.edu/record/54355/files/swansonrobiveco00roberich.pdf | 1976–1980 events, later oral history | participant evidence; reliability A for Swanson | Reconstruct Kleiner/Perkins formation, later Mayfield/Corning links, and board boundary before assigning Davis credit; pair with contemporaneous offering/annual-report records.
• ★ | Watkins-Johnson and Kern/Tenneco records route | https://www.sec.gov/edgar/search/ | 1958 onward | SEC/company-record search route; reliability A when filings are recovered | Recover W-J prospectus, Kern/Tenneco annual reports, security, check, ownership, boards, and disposition; corporate capital is not Davis-person capital.
• ★ | Teledyne historical filings route | https://www.sec.gov/edgar/search/ | 1960 onward | SEC/company-record search route; reliability A when filings are recovered | Reconcile initial Hayden Stone placement, D&R follow-on/exchange, 26,600 shares, boards, and distributions; avoid deriving GP returns.
• ★ | Tandem Computers records route | https://www.sec.gov/edgar/search/ | 1974 onward | SEC/company-record search route; reliability A when filings are recovered | Test Mayfield entity, round, board, Jimmy Treybig/partner roles, and outcome against Davis's 1979 testimony and later firm history.
• ★ | Dysan records route | https://www.sec.gov/edgar/search/ | 1970s onward | SEC/company-record search route; reliability A when filings are recovered | Davis named Dysan in 1979; retrieve financing/ownership/board/outcome records and ordinary-case evidence.
• ★ | Applied Biosystems records route | https://www.sec.gov/edgar/search/ | early 1980s onward | SEC/company-record search route; reliability A when filings are recovered | Test Alafi's recollection against security, Mayfield entity, directors, follow-ons, dilution, and exit.

Coverage assessment

Search frame and identity controls

SCOUT searched Thomas J. Davis Jr., Thomas Jefferson Davis Jr., Tommy Davis, Tom Davis, Davis & Rock, Mayfield Fund, Kern County Land, and combinations with interview, speech, testimony, letter, memo, oral history, portfolio, lawsuit, SEC, complaint, failure, loss, and named companies. Results were rejected unless Harvard, Woodside, Kern, Davis & Rock, Mayfield, or a supported company/partner established identity. The 49ers kicker (1934–1987), baseball player, Scientology spokesman, politicians, and contemporary court defendants are namesakes.

Direct voice and primary archives

The strongest public direct voice is the 1979 congressional statement; the 1981 Electronics item is a second contemporaneous source. Tom Nicholas reports the 1966 speech, while the Peter O. Crisp papers remain only a general archive-request route; the 1986 Kenney/Florida interview is cited but not publicly located. One Man's War remains a physical/licensed-copy route. Archive leads attempted or logged include HBS/Baker Crisp papers, Princeton AC233, Stanford Frederick Terman papers, Kenney/Florida custody, Mayfield/firm records, Corning and company records, and NARA/Harvard/service-record routes for biography. No public D&R or Mayfield governing agreement, LP schedule, IC minutes, partner capital accounts, complete quarterly-report series, carry ledger, or audited cash-flow schedule surfaced.

Candidate denominators and actor boundaries

Separate universes are required: (1) Kern County Land corporate-development investments, beginning with Watkins-Johnson; (2) the finite 1961–1968 Davis & Rock vehicle; (3) Mayfield Fund I; (4) Mayfield II and later funds/partnerships; (5) Davis personal positions, if any; (6) board or advisory roles without investment; and (7) syndicates/co-investors. Davis self-reported in 1979 that his then-current partnership held 23 companies acquired for $8,343,607 over five years, but the company list is incomplete and the figures are not audited. D&R is reported as roughly fifteen companies, but the complete ledger, ordinary outcomes, and failures remain unavailable. A career coverage fraction is therefore not computable.

Historical context, skeptical routes, and completed peers

The favorable “people first” story is bounded by Myers's report that Mayfield I produced only a few-percent result, an initially informal/intuitive partner process, the Maxfield/ROLM-precursor pass, Davis's own admission that companies failed, missing denominator/cash flows, and Davis's illness/key-person succession. Exact lawsuit/SEC searches mostly returned namesakes; absence of a located case is not proof of no litigation. For comparison, the completed Georges Doriot and Arthur Rock packages are defensible on pioneer vehicles, people selection, technology transfer, and attribution; incomplete investors are not comparators.

Highest-value retrieval leads

1. Search the Peter O. Crisp papers for the 1966 speech; if located, record the box/folder and permission boundary.
2. Locate the 1986 Kenney/Florida interview transcript/audio and custody terms.
3. Request Princeton AC233's Mayfield 1968–1974 material and search Stanford investment-office/Terman correspondence.
4. Recover D&R and early Mayfield agreements, LP schedules, reports, capital calls, distributions, partner accounts, and complete portfolio ledgers.
5. Reconstruct contemporaneous company financings/boards for Watkins-Johnson, Teledyne, SDS, Tandem, Dysan, Genentech, Applied Biosystems, Catalytica, Atari, and MIPS.
6. Retrieve a lawful copy of One Man's War and primary Harvard/service/NLRB records to resolve biography without projecting military experience into investing causation.

Source notes