Ai As Filter, Moat Logic, And Economic Mechanism
Key takeaways
- Zeev disputes the claim that AI will broadly kill incumbents, arguing incumbents often have superior data and can benefit if they adapt quickly.
- Zeev states that if an investment thesis looks weird or wrong at entry, it often implies fewer competitors and a multi-year moat-building window, conditional on the thesis being correct.
- Zeev states macro variables like interest-rate velocity can break otherwise strong-looking businesses and that stress tests should consider faster and deeper adverse scenarios than prior worst-case assumptions.
- Oren Zeev states that he is the largest LP in each of his funds and pays himself zero management fee.
- Zeev disputes that growth expectations have fundamentally changed, arguing the compounding math is unchanged and the key question is sustainability and health of growth.
Sections
Ai As Filter, Moat Logic, And Economic Mechanism
The deltas emphasize AI as a gating criterion for new investments, but not as a uniform disruption force. A key mechanism is that in complex/regulatory/integration-heavy businesses, AI alone is less likely to be a full substitute, while in certain operations-heavy businesses AI can improve unit economics (example: support automation improving gross margin). The corpus also highlights a selection bottleneck: even if a category is attractive (e.g., AI support), extreme crowding can make early picking unattractive.
- Zeev disputes the claim that AI will broadly kill incumbents, arguing incumbents often have superior data and can benefit if they adapt quickly.
- Zeev states he is worried that AI-driven disenfranchisement could trigger political unrest that poses significant risk to society.
- Zeev states that operationally complex, distribution-heavy, integrated, data-rich, and regulated businesses are harder to disrupt with AI because software is only a small part of the full system.
- Zeev states he now requires potential investments to be likely beneficiaries of AI and views being AI-neutral as usually insufficient.
- Zeev states he avoids investing early in crowded AI customer-support startups because he does not trust his ability to pick the winner among thousands.
- Zeev states AI can materially improve Navan’s gross margins by automating customer support that historically represented a large cost component.
Decision Process And Communication Constraints (Lp-Gp And Investor-Founder)
The corpus highlights decision bottlenecks beyond pure analysis: round dynamics can force suboptimal structures (example: capped SAFE leading to fixed ownership), and profitable companies can block follow-on ownership-building. It also claims that communication style changes outcomes (LP feedback delivery affecting fund sizing; founder receptivity improving when support is unconditional). Separately, Series A pricing risk is framed as structural due to optics-driven repricing rather than true risk reduction.
- Zeev states that if an investment thesis looks weird or wrong at entry, it often implies fewer competitors and a multi-year moat-building window, conditional on the thesis being correct.
- Zeev states he invested in Descartes via a capped SAFE and ended up with about 5% ownership from a $1.5M check.
- Zeev states that great venture outcomes typically require being meaningfully contrarian and being right.
- Zeev states he generally avoids markets with many early competitors because he wants each investment to have a high probability of becoming the market leader.
- Zeev states Series A pricing risk is structural because valuation often jumps after seed based on optics rather than true risk reduction.
- Zeev states founders are more receptive to investor advice when they believe the investor will support their decision regardless of disagreement.
Portfolio Construction, Pacing, And Vintage/Price Risk
The deltas specify high concentration, fast deployment, and a retrospective claim of 2021 overpayment magnitude, leading to a concrete strategic change: reducing fund size. Macro shock speed (interest-rate velocity) is presented as a key failure mode that can break companies that otherwise look strong, implying stress tests should explicitly model rapid adverse moves. On liquidity management, proactive secondary selling is framed as structurally adverse selection.
- Zeev states macro variables like interest-rate velocity can break otherwise strong-looking businesses and that stress tests should consider faster and deeper adverse scenarios than prior worst-case assumptions.
- Zeev believes he invested too fast in 2021 because the market forced him to overpay by roughly 3–4x on many deals, reducing expected fund multiples materially.
- Zeev states he uses a 20% per-company concentration limit, above the industry-standard ~10%.
- Zeev states he is willing to deploy a fund very quickly (sometimes ~12 months) and will not slow down to accommodate LP pacing preferences.
- Zeev states his 2021–2022 fund sizes (over $500M) were too large for top returns and that he cut his 2024 fund size to roughly half (around $250M), aiming for even less for the next fund.
- Zeev states he targets roughly 15 companies per fund currently, and due to cross-fund overlap he is in fund 11 with about 40 unique companies total.
Vc Incentives, Fund Governance, And Mark Reliability
The corpus places incentives at the center of interpreting VC behavior and reported performance. Zeev presents an unusually LP-aligned fee/economics structure for himself, and separately claims that reported valuations and audits can be weakly constraining due to motivation and latitude in marking. Net implication: evaluate claims about NAV/TVPI through the lens of who benefits from optimistic marks.
- Oren Zeev states that he is the largest LP in each of his funds and pays himself zero management fee.
- Zeev states he reinvests 100% of management fees into the fund, takes no personal income from fees, and receives economics only after LPs have received 100% of their capital back.
- Zeev states he is typically about 13–14% of each fund as an LP, caps any other single LP below 10%, and takes 30% carry, making him roughly 40%+ of the fund economics.
- Zeev states reported VC valuations are heavily influenced by GP motivation, where brand-name funds have less incentive to inflate marks while marginal funds may rationalize higher paper values to facilitate fundraising.
- Zeev states auditors are weak constraints on VC valuations because they lack context to challenge economically relevant assumptions and there is wide latitude to justify marks.
Growth Evaluation: Compounding Vs. Sustainability And Revenue Quality
The deltas dispute the idea that growth expectations have fundamentally changed, instead pushing a conditional view where growth is evaluated through sustainability, unit economics, and credible forward indicators. A specific failure mode flagged is metric-gaming through circular revenue driven by growth-only incentives. Market structure conditions matter: winner-take-most dynamics can justify temporarily sacrificing margins.
- Zeev disputes that growth expectations have fundamentally changed, arguing the compounding math is unchanged and the key question is sustainability and health of growth.
- Zeev states a 1-to-5 revenue jump is sufficient when unit economics are strong and the forward trajectory implies continued rapid growth (e.g., 5 to 15–20 next year).
- Zeev argues that overweighting top-line growth incentives can push companies into unsustainable behavior such as circular revenue deals that inflate sales without creating real value.
- Zeev states that in winner-take-most competitive dynamics, companies may be forced to prioritize market share over margins until leadership is secured.
Watchlist
- Zeev states macro variables like interest-rate velocity can break otherwise strong-looking businesses and that stress tests should consider faster and deeper adverse scenarios than prior worst-case assumptions.
- Zeev states he is worried that AI-driven disenfranchisement could trigger political unrest that poses significant risk to society.
- Zeev expects a potential 2026–2027 wave of very large IPOs could create a liquidity surge that reshuffles venture fundraising dynamics.
Unknowns
- To what extent does the AI-beneficiary investment filter actually improve portfolio outcomes relative to prior criteria, and what measurable KPIs would confirm beneficiary status (e.g., margin expansion, retention, velocity)?
- Are operationally complex/regulatory/integration-heavy businesses empirically less disruptable by AI in the relevant markets discussed, and over what time horizon?
- Is Navan realizing the specific AI-driven gross margin improvement mechanism claimed (support automation), and what is the magnitude and timeline?
- How prevalent are circular or low-quality revenue practices in the cohorts being discussed, and what indicators reliably detect them early?
- What is the realized impact of 2021 overpayment (3–4x) on eventual DPI/TVPI outcomes for the affected funds, versus earlier/later vintages?