OpenAI CEO Altman’s house was burned down. In the early hours of Friday morning, a 20-year-old threw a Molotov cocktail at Altman's San Francisco residence. Fortunately, no one was injured. An hour later, the reckless young man appeared outside the OpenAI office building and threatened to burn down the entire building. The suspect's social account quoted the plot of the "Dune" novel in which humans launch a jihad to eliminate AI.

The shocked Ultraman then posted a rare family photo and wrote a long article. He said that AGI has become the Lord of the Rings. The real danger is not the technology, but the idea of ​​"becoming the one who controls AGI", which can make people do extremely crazy things. He also publicly apologized for the first time for the boardroom turmoil in 2023, admitting that his "fear of conflict" character caused a huge mess to the company.


This company has just completed the largest single round of private equity financing in human history: US$122 billion, with a post-money valuation of 852 billion. On one side, there are people who want to burn it down, and on the other side, there are people who are pouring hundreds of billions of dollars into it. Fear and greed have never been so nakedly present in the same company as they are today.

The institutions that pay out the money have their own calculations behind it. Amazon promised to invest 50 billion US dollars, but it is not a one-time payment. 15 billion will be received immediately, and the remaining 35 billion comes with conditions, requiring OpenAI to complete an IPO or realize general artificial intelligence (AGI) before the end of 2028. Nvidia and SoftBank each invested US$30 billion, which will be received in batches in July and October 2026. Microsoft continued to invest but did not disclose the specific amount. More than 20 top institutions including a16z, Sequoia, BlackRock, Blackstone, and Thrive Capital have all entered the market.

Amazon and Nvidia are not single bets. They are also investors in Anthropic, OpenAI’s biggest competitor.

Amazon spent billions of dollars on Anthropic to deeply bind the Claude model to AWS; then it turned around and wrote a check for 50 billion to OpenAI. The logic of this hedging bet is very clear. It is not to bet on who will win, but to ensure that no matter who wins, you will sit at the card table.

Nvidia, which is not short of money, has invested US$30 billion in OpenAI, which will eventually flow back into its own pocket in the form of a GPU purchase order. It is more like an advance payment for a chip sales contract than an investment.

OpenAI also did something unprecedented: it opened subscriptions to individual investors through bank channels for the first time, and received more than $3 billion from retail investors. ARK Invest also announced that it has purchased US$240 million in OpenAI shares through multiple ETFs. A company that has not yet been listed is already making comprehensive preparations for its IPO.

But after US$122 billion was poured into it, whose pocket did the money end up in? At the same time as the financing news was announced, an equity table was leaked.

A leaked shareholding sheet

This equity structure table, said to be from within OpenAI, puts the company’s ownership distribution under the sun for the first time.

The top five shareholders are: Microsoft 26.79%, OpenAI Foundation 25.8%, SoftBank 11.66%, Amazon 4.66%, and Nvidia 3.47%. Current and former employees collectively hold about 20% of the shares.

Microsoft is the biggest winner on this table. Starting from the first US$1 billion in 2019, to 10 billion in January 2023, and to an additional 2 billion in 2024, Microsoft has invested a total of approximately US$13 billion in OpenAI. Based on the current valuation of 852 billion and a shareholding ratio of 26.79%, the book value of this 13 billion US dollars is 228.3 billion US dollars, with a return multiple of more than 17 times. This may be one of the most rewarding single bets in the history of technology investing. Microsoft's advantage is that it entered the market early. When it made its first investment in 2019, ChatGPT did not exist yet, and almost no one outside the academic circle had heard of the term "large language model."

Look at the other players. SoftBank’s investment of US$64 billion has a current book value of approximately US$99.3 billion, and a paper return of about 1.5 times. This number does not look sexy. The reason is simple: SoftBank’s valuation was already very high when it entered the market. By the same logic, Amazon and Nvidia are more expensive this time around. They are not betting on short-term returns, but on strategic positions at the AI ​​infrastructure level.

But the most eye-catching number on this table is not Microsoft's 228.3 billion, but another number - zero. Altman, CEO of OpenAI, holds zero shares. This leads to the most dramatic line in the entire OpenAI story.


CEO shares zero

In a company valued at $852 billion, the CEO doesn’t own a penny of equity. His annual salary is $76,001, which, according to Altman, is "the minimum wage required to get health insurance."

Ultraman's explanation of this matter is always honest. At the New York Times DealBook Summit last December, he said it was his "dream job since I was a kid" and that "being able to sit in a room with the smartest researchers in the world and be a part of this crazy adventure is worth more than any extra money to me." But then he said something even more intriguing: "If I could go back in time, I would take some equity, even just a little bit, so I don't have to answer this question forever."

Not taking equity was initially an active choice. OpenAI was a non-profit organization when it was established in 2015, and managers holding equity will affect tax-exempt status. Altman himself has said that he needs to prove to the outside world that he can separate personal interests from the company's mission. But as the company's valuation has risen from zero to 852 billion, the price of this "active choice" has become higher and higher.

Zero shareholding implies governance risks. In 2023, the OpenAI board of directors suddenly dismissed Altman. In a normal company, the CEO is usually one of the largest shareholders, and it is difficult for the board of directors to expel the CEO without his consent. But Altman has no equity voting rights on the board of directors, and his CEO position relies entirely on the trust of the directors. Later, under pressure from Microsoft and its employees, Altman was reinstated, but this incident also reminded that a CEO without equity is extremely fragile in terms of governance structure.

Altman later reflected on the incident and said he "learned a lot about communication in times of crisis," but also felt for the first time "the pride that a team can function without me."

The outside world has been speculating that Ultraman will get the equity sooner or later. In September 2024, Bloomberg reported that the OpenAI board of directors was discussing giving Ultraman a 7% stake. Later, Altman denied it at the general meeting, saying that "there is no plan to obtain a huge equity stake" and that the numbers reported by the media were "too outrageous."

But both he and CFO Fryer acknowledged the fact that investors are worried about him not holding equity because investors want the CEO to be bound to their own interests. Bret Taylor, chairman of the board of directors, confirmed to the media at the same time: The board of directors has indeed discussed this matter, but has not yet discussed specific numbers.

In October 2025, after the company completed its transformation, Altman talked about this topic again on social media, and his tone was much more frank than before: "I wish I had taken the equity long ago, so there would be fewer conspiracy theories." He also said that the decision not to take the equity was "very stupid, trying to express 'I already have enough money.'" From "voluntarily giving up" to "a very stupid move", the change of wording itself says a lot. The “transformation” Altman calls is the biggest change OpenAI has experienced in the past year.

From non-profit to nearly one trillion valuation

On October 28, 2025, OpenAI completed its transformation from a non-profit to a for-profit company. This restructuring, which has lasted nearly a year, has to solve a fundamental contradiction: a company that needs to burn hundreds of billions of dollars can no longer continue to use a non-profit structure to raise funds.

The new structure is like this: the original non-profit organization became the "OpenAI Foundation", which holds 26% of the equity of the for-profit subsidiary "OpenAI Group PBC", with a valuation of approximately US$130 billion. Microsoft obtained 27% of the shares, worth 135 billion. The foundation's board of directors has "special voting rights" and can appoint and remove all directors of PBC. California Attorney General Rob Bonta said of the plan that it "ensures that charitable assets are used for their intended purposes, safety will be prioritized, and OpenAI will remain in California." The subtext of this sentence is that without these commitments, the Attorney General may take the plan to court.

But the outside world is not without doubts. Johnston, director of the nonprofit watchdog group Midas Project, pointedly pointed out: The board members of the foundation and PBC almost completely overlap, and the supervisory power of the nonprofit organization is essentially reduced to "the power to fire itself." This question touches on the core contradiction of the entire transformation. When mission and profit conflict, will a two-tier structure managed by the same group of people really abandon private interests and choose public welfare?

The transformation is complete, and the next step is the IPO. The IPO timetable is becoming an increasingly public point of disagreement within OpenAI.

According to The Information, Altman hopes to go public as early as the fourth quarter of this year, but CFO Fryer has privately expressed concerns. She believes that it is not yet qualified for listing in 2026, process and organizational preparations have not been completed, and huge expenditure commitments are also risky.

Amid the disagreement, Altman allegedly began excluding Fryer from some financial planning discussions. This kind of rift between the CEO and CFO is not a good thing for a company that is about to go public.

External competitive pressure is also accelerating. Anthropic is also considering going public by the end of 2026, with an expected financing size of more than US$60 billion. If Anthropic is listed first, it will suck away the enthusiasm for AI stocks in the market. Therefore, for OpenAI, IPO is not only a financing need, but also a race against time. The end point of this race points to a question that everyone is asking.

Is $852 billion worth it?

Is OpenAI worth $852 billion?

Let’s look at income first. According to Reuters citing Fryer's statement at the beginning of the year, OpenAI's current monthly revenue is US$2 billion, which is approximately US$25 billion annually. ChatGPT has more than 900 million weekly active users, more than 50 million paid subscribers, and enterprise revenue accounts for more than 40% of total revenue. Within six weeks of the advertising pilot, it generated more than $100 million in annualized recurring revenue. Looking at the growth rate alone, OpenAI is indeed running faster than Google and Meta in the same period.

A valuation of 852 billion corresponds to an annualized revenue of 25 billion, with a price-to-sales ratio of approximately 34 times. If calculated based on the trillion-dollar valuation expected by the market at the time of the IPO at the end of the year, the price-to-sales ratio will be even higher. For comparison, NVIDIA's market-to-sales ratio did not exceed 40 times at its craziest time, and NVIDIA is actually making money. OpenAI is expected to lose US$14 billion this year.

The pressure on the cost side is equally staggering. According to The Information, OpenAI’s computing power investment plan in the next five years is as high as US$600 billion. It is expected to achieve positive cash flow in 2030 and will lose more than US$200 billion before then. At the same time, its API pricing is still several times that of its competitors. Once a price war breaks out, revenue growth will be under pressure, and computing power expenditures will be rigid.

Revenues are elastic and costs are rigid, which is a red flag.

Drawing our attention back to China, the contrast becomes even more intense. Since the beginning of 2026, China’s AI financing is also looking very hot. Dark Side of the Moon raised US$700 million, and Step Star completed a RMB 5 billion B+ round, setting a new record for single financing for large models in China. But adding together the historical financing amount of all large model companies in China, it is probably a fraction of OpenAI’s current round. This is not a problem of technology gap, but a completely different capital structure.

AI financing in the United States is essentially an arms race between technology giants. Amazon invested 50 billion to sell cloud, Nvidia invested 30 billion to sell chips, and SoftBank invested 30 billion to promote its Stargate project. The money went around in circles and returned to the hands of investors.

China's AI financing is still at the stage dominated by VCs and local industrial funds, and has not formed such a closed loop of "investment as procurement." When one party is fighting with the balance sheet of a giant, and the other party is still looking for financial investors round after round, the capital base of this competition is different from the starting point.

So what exactly did 852 billion buy? Ultraman himself gave the answer. "We are now very confident that we know how to build AGI," he wrote in his blog. He also said that by 2035, everyone on the planet should be able to mobilize the equivalent of the combined intelligence of all humans in 2025.

Investors are betting on this vision. If AGI really comes, all today's numbers will be rewritten, and 852 billion will look cheap. If it doesn’t come, or if it comes slower than expected, today’s valuation will be a huge bubble. This $122 billion is not a bet on a company’s financial statements, but on the end of AI.

in conclusion

From a non-profit laboratory when it was founded in 2015 to sprinting for an IPO with a valuation of US$852 billion in 2026, it took OpenAI 11 years to complete a path that most companies cannot take in a lifetime. In the past 11 years, it has transformed from an idealistic research institution into the most funded commercial company in history, experienced the farce of the CEO being fired and reinstated, and completed the legal transformation from non-profit to for-profit. However, its leader has not received a penny of equity.

There are no ready answers to any of the questions facing this company. Differences between the CEO and CFO on the pace of listing have surfaced. It will burn US$600 billion in the next five years, and will fall into a passive position once revenue growth slows down. A “non-profit controlled for-profit” structure run by the same people has not been tested for true conflicts of interest.

Ultraman once predicted, "Close to the singularity; not sure which side it is on." A $122 billion bet has been made, and it will take several years for the answer to be revealed.