Oct 16, 2025, 10:23 AM ETFireside Chat, Interview, Podcast

Thinking Machines Co-Founder Joins Meta for $3.5BN, Industry Venture's $665M Acquisition

Goldman Sachs acquired Industry Ventures for a base price of $665 million with a potential earn-out of up to $300 million, contingent on performance through 2030.
Industry Ventures manages $7 billion in assets (AUM), representing a 10% multiple of AUM, which aligns with secondary business valuations (Carlyle/KKR trade at 20% of AUM; public funds trade at 1-2%).
Goldman Sachs' strategic rationale includes transferring Industry Ventures' high-net-worth clients into the "Apex" private market platform and expanding access to private capital for its public asset management division.
The deal underscores a structural difference in venture ecosystems: asset managers with productizable fee streams (like Industry Ventures) are acquirable, whereas pure venture firms relying on individual founder IQ are generally not.
Founders of asset managers like Industry Ventures often face different exit dynamics than venture general partners, as fee-based businesses can be sold independently of specific investment performance.

Andrew Tullock, co-founder and lead of Thinking Machines (raised $2B, $10B post-money valuation), departed to join Meta for a reported $3.5 billion cash/liquid stock package.
Panelists characterize this as a shift from "multi-period games" to "one-and-done" scenarios due to the sheer size of exit offers, which erodes traditional founder loyalty and long-term commitment.
Rory recommends that investors and founders implement extended vesting schedules (e.g., 6-year cliffs) and aggressive repurchase rights to mitigate the risk of co-founders leaving for competitors.
The transaction is described not as an acquisition of a company, but as an "asset purchase" of human capital, prioritizing specific talent over the enterprise's liabilities or remaining equity.
Jason argues that the $3.5B offer represents a "transactional" moment where individual financial optimization overrides the collective mission, potentially indicating a broken evolution in human behavior at the highest wealth levels.

SoftBank reportedly secured a $5 billion margin loan secured by ARM shares to invest in OpenAI, reinforcing a "high-risk, all-in" strategy by Masayoshi Son.
The panel notes that while margin calls are theoretically possible (e.g., during a NASDAQ-style 90% crash), the probability is low given ARM's current valuation of ~$90-100 billion.
Jason projects an insatiable demand for AI compute, citing "vibe coding" capabilities that allow him to build 100x more software than possible traditionally; he estimates token consumption could increase 1,000x if supply constraints were removed.
The consensus is that the primary constraint on AI scaling is no longer technical feasibility (scaling laws have held for 6-7 years) but economic viability (diminishing marginal returns on capital).
Data center construction is outpacing office building, reflecting a fundamental shift in infrastructure demand driven by the need to house compute, not by remote work trends.

Polymarket raised $2B at a $9B valuation, while Kalshi (direct competitor) raised from Andreessen and Excel at a $5B valuation shortly thereafter.
Panelists debate whether this represents "kingmaking" (capital picking a winner); they conclude these are "capital-intensive" races where money fuels marketing and distribution rather than creating an inherent winner through "anointment."
Harry and Rory identify the sector as a regulatory arbitrage play, where unregulated or lightly regulated entities (Polymarket, Kalshi) capture value from regulated sports betting markets.
The current U.S. administration's stance is viewed as a key factor enabling rapid growth for these entities, contrasting with the historical regulatory hurdles faced by legacy sports betting operators.
The discussion highlights a "structural budget deficit" argument: state tax revenue from regulated betting drops when customers migrate to offshore or unregulated platforms due to higher tax rates.

Peter Thiel (Founders Fund) is shifting to a concentrated strategy (10-12 deals in growth fund), while others (e.g., DST/GC) advocate for broad indexing across AI companies (Mistral, Anthropic, OpenAI) to hedge against uncertainty.
Roger's strategy involves a "funnel" approach: starting with 20-25 initial diversified bets to create a farm team, then concentrating 75% of capital into 3-5 winners during follow-on rounds where conviction is highest.
Harry notes that as exit bars rise from $200M ARR to $400M+ ARR, the timeline to exit has extended, necessitating a longer holding period and potentially more diversified initial portfolios to survive the "long tail."
The panel agrees that once a company generates 2 years of revenue, the probability of a 5x return jumps from 30% to the mid-70s, justifying aggressive concentration at that stage.
Jason highlights the "optionality" problem: if early checks are too large (8% of fund per deal), the fund may exhaust capital reserves before later, higher-conviction rounds occur, though this is mitigated by writing smaller initial checks.
Rory emphasizes that every check should be evaluated as an "option" with risk-adjusted value, where a follow-on round at a higher price can be a signal to increase allocation if the external validation is strong.

The "sweet spot" for identifying differentiation in seed-stage AI companies has compressed from months to weeks or days, making early-stage differentiation difficult to assess.
Investors face a trade-off between investing in "acute uncertainty" (seed) versus "acute uncertainty at higher valuations" (Series A+), with the latter often offering clearer product-market fit signals.
The panel suggests that "parallel LPs" and "cross-fund investing" structures are critical for maintaining the ability to fund follow-ons across a company's lifecycle without capital constraints.
There is a growing consensus that "stewardship" requires VC presence for the duration of the investment (up to 10 years), rather than exiting early, especially as founders have more leverage to depart.