The Twenty Minute VCOpenAI Restructuring: Who Wins and Who Loses & Mercor Raises $350M at a $10BN Valuation
CHAPTERS
OpenAI restructures: deal cut with Microsoft and state AGs, path to raising and IPO
Rory breaks down the headline: OpenAI has resolved negotiations with Microsoft and the attorneys general in Delaware and California, enabling a long-awaited restructuring. The new structure removes prior investment “straitjacket” constraints and makes an IPO materially more feasible.
What the new structure actually is: nonprofit foundation + Public Benefit Corporation
The hosts clarify what OpenAI becomes after the conversion: a well-capitalized charitable foundation above a for-profit Public Benefit Corporation (PBC). This removes earlier “treat as a donation” warnings and places OpenAI in a recognized corporate form that can access public markets.
Winners and “non-losers”: Microsoft’s leverage, foundation funding, employees, investors, and the chairman
Rory frames the outcome as a settlement aligning with leverage: Microsoft gets a great deal; the foundation ends up extraordinarily funded; employees gain liquidity potential; investors can breathe. He also credits Brett Taylor for navigating governance and negotiations.
The Sam Altman anomaly: CEO with no equity, and why that’s unprecedented
Jason highlights a standout detail: Sam Altman reportedly still holds no shares. They contrast it with the era’s mega-compensation expectations and discuss how owning nothing may reduce criticism while altering power dynamics.
Capital intensity and the “next $200B”: why restructuring matters for suppliers and commitments
They argue the restructure isn’t just governance—it's a financing unlock for enormous future commitments (cloud, chips, and supplier agreements). If OpenAI can access public markets, it can credibly fund massive forward contracts that previously looked speculative.
Valuation mechanics and euphoria: secondaries, markups, and the trillion-dollar question
They discuss conflicting price signals (primary vs secondary), rapid markups, and the possibility of OpenAI reaching $1T+ soon. Rory notes high multiples can hold while growth stays hot, but slowdowns can be brutal at 40x revenue-type pricing.
Andreessen Horowitz raises $10B: is this the new normal for mega-platform venture?
The hosts debate whether $10B is enormous or surprisingly “small” once broken into sub-funds. Rory argues mega-platforms are the dominant model now, with potential venture-industry bifurcation between mega-firms and boutiques—pending long-cycle returns.
Media, “wall of news,” and option-buying: how big funds win deals
A sharp disagreement emerges: does a “wall of news” and brand/media presence materially win deals? Rory argues AUM plus media strategy tilts the table; Jason counters that distribution and coverage can be engineered without mega-check sizes, while acknowledging option dynamics as the bigger edge.
Mercor’s $350M raise at $10B: RLHF labor market, real vs GMV revenue, and concentration risk
They unpack Mercor’s hypergrowth and what it sells: high-skill human feedback to train frontier models (RLHF). While revenue is considered “real” for accounting, they flag structural risks: heavy payout to experts, extreme customer concentration, and sensitivity to AI CapEx cycles.
Late-stage private rounds, dilution, and why headline valuation step-ups can be misleading
Ramp’s rumored $30B raise triggers a discussion about small step-ups, dilution, and the “public stock hiding in private” phenomenon. Rory argues mature private growth should deliver public-like returns; Jason notes that headline rounds often don’t move earlier investors’ outcomes as much as expected.
Spray-and-pray vs picking vs optioning: Carta’s Series B outcome distribution and what it implies
Using Carta data on 2018 Series B outcomes (547 deals), they debate whether diversified “spray and pray” works. Rory argues the data supports disciplined picking and that true “spray” only works if paired with option economics; Harry highlights multi-stage firms placing many seed “options.”
M&A and antitrust friction: Roomba’s collapse, acquisition timing risk, and why secondaries matter
Roomba’s blocked Amazon acquisition becomes a cautionary tale: regulatory delays can destroy companies and distort deal economics over long review periods. They tie this to founders’ sell/hold decisions and argue early managers should use secondaries more proactively given duration risk.
Amazon’s “bad fortnight”: layoffs, AWS share pressure, and missing the AI compute wave
They assess Amazon’s challenges: retail over-hiring during COVID and the need to automate, plus a cloud narrative of losing relevance in AI-era compute compared to Microsoft/Google. Jason argues Bezos stepping down before the AI inflection was poorly timed; Rory emphasizes the need to regain AI relevance without uneconomic deals.
Agree/Disagree lightning round: Ramp vs Brex, Andreessen’s momentum, and Anduril as the hottest private ‘public-ready’ company
The episode closes with rapid takes: valuation vs growth tradeoffs (Ramp vs Brex), whether Andreessen is the best mega-platform performer, and Anduril’s desirability as a late-stage private holding. Jason declines Anduril on personal interest despite acknowledging its status-symbol allure; Rory praises Andreessen’s operational machine and scalability.
