On October 28, 2025, OpenAI finished converting its capped-profit LLC into OpenAI Group PBC — a public benefit corporation valued near $500B, controlled by a nonprofit that holds ~26% (~$130B), with Microsoft at ~27% (~$135B) and the old 100x profit cap gone for good. That's the short answer. The longer answer is where the real money question lives.
The restructuring wasn't cosmetic. It rewired who gets paid, removed the ceiling on investor returns, and turned a structure no bank could underwrite into one you can actually take public. For anyone holding OpenAI exposure — directly, through a fund, or via Microsoft stock — the conversion changed the math.
What the OpenAI For-Profit Conversion Actually Did
The OpenAI for-profit conversion, completed October 28, 2025, replaced the company's "capped-profit" LLC with OpenAI Group PBC, a Delaware public benefit corporation. The nonprofit, renamed the OpenAI Foundation, kept board control and took roughly 26% of the equity — about $130B at the ~$500B valuation. The change eliminated the 100x cap on investor returns and put every shareholder on conventional equity.
Before this, OpenAI ran one of the strangest corporate structures in tech: a nonprofit (OpenAI Inc.) controlling a capped-profit subsidiary (OpenAI LP) where the earliest investors could earn at most 100x their money, after which excess returns flowed back to the nonprofit. That worked when OpenAI was raising hundreds of millions. It broke completely once the company needed to raise tens of billions for compute. The conversion fixed the structure's core flaw — you cannot sell uncapped upside out of a capped-profit entity.
Who Owns What After the OpenAI Restructuring
The cleanest way to understand the recap is the cap table. The nonprofit's economic stake fell from majority owner to roughly a quarter of the company, but it kept control. Microsoft converted its prior investment into a conventional ~27% equity position. Employees and outside investors hold the balance. At a ~$500B post-money valuation, these percentages translate into very large dollar marks.
| Stakeholder | Approx. Stake | Implied Value (~$500B) | Control / Notes |
|---|---|---|---|
| OpenAI Foundation (nonprofit) | ~26% | ~$130B | Retains board control of the PBC |
| Microsoft | ~27% | ~$135B | IP rights extended through 2032 |
| Employees (current & former) | ~25% | ~$125B | Vested equity, sold via tenders |
| SoftBank | ~5–7% | ~$25–35B | Led the $40B 2025 round |
| Thrive Capital | ~2–3% | ~$10–15B | Multi-round insider |
| Other investors (Khosla, a16z, MGX, etc.) | ~10–14% | ~$50–70B | Includes prior funds and secondary buyers |
Figures are 2025–2026 estimates blended from OpenAI's own October 2025 announcement, Microsoft disclosures, The Information, Bloomberg, and FT reporting. Stakes are approximate and fully diluted; individual investor percentages are reported ranges, not audited figures.
Why the OpenAI For-Profit Conversion Killed the 100x Cap
The 100x return cap was the single most important thing the conversion removed. Under the old capped-profit model, OpenAI's first investors — Microsoft's initial $1B in 2019 among them — were contractually limited to 100x their money. That sounds generous until you do the arithmetic at scale: a fund that wants to write a $10B check cannot accept a structure where the upside is mathematically bounded while the downside is not.
By 2025, OpenAI needed capital on a scale no capped structure could support. The 2025 round alone was about $40B led by SoftBank — the largest private financing in history — and the Stargate infrastructure plan implied $500B of compute buildout over five years. SoftBank's commitment was explicitly conditioned on OpenAI completing the for-profit conversion by year-end. Miss the deadline and a chunk of the round reportedly fell away. The cap had gone from a mission safeguard to a fundraising liability.
Converting to a PBC solved three problems at once: it let OpenAI issue conventional uncapped equity, it gave employees liquid, ordinary shares instead of capped profit-participation units, and it created the clean structure a public listing requires. The trade-off the nonprofit accepted — dropping from majority economics to ~26% while keeping control — was the price of unlocking that capital.
OpenAI Restructuring Timeline: From Nonprofit to PBC
The conversion took nearly two years of legal and governance maneuvering, including a very public fight with Elon Musk, who sued to block it. Here is how the structure evolved.
2019
Capped LP
100x profit cap created
Dec 2024
PBC plan
Conversion proposed
Mar 2025
$40B round
SoftBank leads, cap-tied
Oct 2025
$500B PBC
Recap completed
Two side agreements mattered as much as the equity split. The revised Microsoft deal extended Microsoft's rights to OpenAI's models and IP through 2032 and changed how the "AGI" clause works — AGI declarations now require verification by an independent expert panel rather than OpenAI's board alone. And the nonprofit committed an initial $25B to health and AI-resilience programs, partly to satisfy the California and Delaware attorneys general, who had to bless the conversion because charitable assets were involved.
What the For-Profit Conversion Means for Investors
For investors, the conversion is overwhelmingly positive on economics and slightly more complicated on governance. On economics: removing the 100x cap means the people who backed OpenAI early now own ordinary equity with unbounded upside. At a ~$500B valuation rising toward a potential $1T IPO, that uncapped exposure is worth far more than capped units ever could be. Microsoft's ~27% alone is marked around $135B — one of the most valuable single positions any company holds in a private business.
On governance: the nonprofit still controls the board, so investors own economics without control. That is unusual but not unheard of — it resembles dual-class founder control at companies like Meta or Alphabet, except the controlling party is a charity bound to a mission rather than a founder. The risk for investors is that the foundation can, in principle, make decisions that favor the mission over returns. The PBC charter formalizes that the board must balance shareholder and public interest. For most LPs and crossover funds, the upside removed by losing the cap outweighs that governance ambiguity.
For LPs in venture funds with OpenAI exposure — Sequoia, Thrive, Founders Fund, a16z, Khosla — the conversion turned capped paper into conventional equity that can be marked and eventually distributed. That matters for fund performance metrics like TVPI and DPI: the position is now a clean, sellable security rather than a structurally encumbered one. Track the broader frontier-AI valuation race and capex commitments on our AI Spending dashboard, and see how the revenue side is scaling in our breakdown of OpenAI's $25B revenue run rate.
The OpenAI for-profit conversion wasn't about abandoning the mission.
It was about making the structure fundable — and the nonprofit traded economics for control to get there.
A nonprofit that owns ~26% but controls the board is a deliberate split: the mission governs, the capital flows. Whether that balance holds once OpenAI is a $1T public company is the real question the recap left open.
Explore Related Dashboards
Interactive tools with live data on this topic
Track frontier AI valuations, capex, and the IPO pipeline across OpenAI, Anthropic, Google, and xAI on the AI Spending Dashboard and VC Performance Dashboard at Value Add VC. Originally published in the Trace Cohen newsletter.