OpenAI President Brockman Says Company May Be Ready for IPO in 2027

Greg Brockman rarely admits failure in public. But at the WSJ Tech conference on Wednesday, he made a rare concession.

“Q1 was quite tough, and Q2 was quite tough too,” said the co-founder and president of OpenAI. “On the coding front, we're not even in the conversation. We have Codex, but nobody talks about it.” He was describing a period in which competitors had left them behind. Anthropic's Claude Code captured 54% of the enterprise coding market, with annualized recurring revenue soaring from about $9 billion at the end of 2025 to $44–47 billion. In terms of overall share of enterprise AI spending, Anthropic's share jumped from about 10% in early 2025 to more than 65% by February 2026, overtaking OpenAI. Brockman did not try to hide this.

Then he pivoted: by the third quarter, “the picture was completely different.” The releases of GPT-5.6 Sol and ChatGPT Work brought what he called a “revenue inflection point”: OpenAI's annualized revenue grew 70% quarter over quarter, enterprise business revenue doubled, and the share of revenue contributed by enterprise customers jumped from around 30% previously to a majority. He cited JPMorgan as an example of the kind of customer the company is winning. When asked whether he was ready for an IPO in 2027, he replied: “I think so. Yes.”

There is a true side to this narrative. An OpenRouter analysis of about 120,000 enterprises using tools from both companies showed that by September, OpenAI and Anthropic's spending shares were “roughly even”; at the start of the year, Anthropic still held three-quarters of it. The absolute scale of enterprise revenue is also expanding: by the end of September, OpenAI's annualized revenue was close to $50 billion, up 77% year over year. Brockman's statement that “the enterprise business has really taken off” is not empty talk.

But “ready for an IPO” is a far more complex judgment than “enterprise revenue doubled.”

Enterprise revenue growth masks the structural costs paid for that growth

OpenAI's enterprise business is indeed accelerating. The problem is that the cost and sustainability of this acceleration are folded into the phrase “the painful work behind the scenes” in Brockman's account.

One repeatedly discussed but never settled issue is the revenue metric. OpenAI uses the net method in cloud distribution arrangements, recognizing only its own share of the transaction; Anthropic uses the gross method, counting the full amount customers pay through cloud platforms as revenue. OpenAI argues that Anthropic's gross method inflates reported revenue by billions of dollars. This dispute itself does not constitute a verdict on either side, but it points to a deeper vulnerability: when an AI lab's revenue depends heavily on cloud giants that are simultaneously investors, compute suppliers, and channel partners, the boundary between “revenue growth” and “cost spending” becomes blurred. Every dollar of enterprise revenue growth simultaneously creates compute orders and channel fees for the same set of counterparties.

Brockman attributed the jump in enterprise revenue to “a series of earlier investments finally paying off,” but what OpenAI has not disclosed is how much of those investments are locked in as committed spending for the future. As of July 2026, OpenAI's cloud infrastructure spending commitments through 2030 were about $750 billion; Bernstein, after rebuilding the model under the assumption of on-time contractual delivery, estimated actual spending at about $614 billion, peaking in 2031 at about $184 billion. These figures are of the same order of magnitude as the $895.4 billion in compute commitments disclosed in Anthropic's prospectus. Enterprise revenue growth is real, but it is happening against a backdrop in which capital expenditure is also expanding in units of hundreds of billions of dollars. Doubling revenue and doubling the bill are two sides of the same coin.

The private market is already repricing, and the public market will only be harsher

The most noteworthy thing about Brockman's remarks is not what he said, but the context he did not mention.

Just days before he made these remarks, OpenAI's latest private funding round had its target cut from $30 billion to about $20 billion, and its valuation target lowered from $1.4 trillion to $1.25 trillion. As of the end of October, the company had secured about $11 billion in investment commitments, only 55% of the lowered target. Some investors reduced the size of their subscriptions. This is not a marginal adjustment but a directional signal: in the private market, pricing for OpenAI has already begun to swing back from a “narrative premium” toward “cash flow constraints.”

And the constraints of the public market will only be tighter than those of the private market. In the first quarter of 2026, OpenAI's operating loss was about $9.3 billion, its cash burn was about $3.7 billion, and its adjusted operating margin was about negative 122%; in the second quarter, revenue was $6.7 billion, but the operating loss widened to about $12.3 billion, equivalent to burning more than $1.8 for every $1 earned. Internal forecasts show that cumulative negative free cash flow from 2026 to 2030 will reach $278 billion, and spending on compute and infrastructure alone will reach about $856 billion by the end of 2030—exceeding the cumulative revenue forecast of $840 billion for the same period. A research report published by HSBC in November 2025 already warned that even by 2030, OpenAI would still find it very difficult to achieve profitability.

Brockman said he has “never been more confident,” and at the micro level of the enterprise business, that statement holds. But IPO pricing is not a price set on confidence in any single quarter. It is a discount of future free cash flow over the next decade. When the endpoint of that discount remains far out of reach, while committed spending has already been locked in by contract as irrevocable obligations, the public market's first question will not be “how much has enterprise revenue grown,” but “how far away is the point at which external financing is no longer needed.”

The organizational structure cleared the legal obstacles, but did not answer the governance question

The restructuring in October 2025 did, at a technical level, pave the way for an IPO. Investors' previous investments were uniformly converted into common equity, the cap on shareholder returns was removed, the for-profit entity was reorganized as OpenAI Group PBC, the nonprofit OpenAI Foundation holds about 26% of the equity and has ultimate control, and Microsoft holds about 27%. From the SEC's perspective, the legal structure is already in place. OpenAI even confidentially submitted a draft S-1 in June.

But the distance between “can file an application” and “can successfully price” is not determined by legal documents. The arrangement under which the foundation holds control can, in the private market, be regarded as an institutional safeguard for its mission commitment; in the public market, it will be re-examined as a governance variable: when the entity holding control prioritizes a nonprofit mission, to what extent can the interests of minority shareholders be protected? Altman himself does not hold equity in the restructured company, which is a plus in the mission narrative but an anomaly that needs to be explained within the principal-agent framework of the public market. These are not disqualifying questions, but they are questions that cannot be avoided on an IPO roadshow. Brockman's “ready” answers confidence at the operational level; the pricing logic at the governance level requires a different language.

A hole has just been punched in Anthropic's IPO window, and OpenAI's 2027 is not a blank calendar

Brockman said 2027, which comes right after Anthropic's IPO probability suffered a sharp plunge.

On November 5, the Nasdaq fell more than 5%, and the Philadelphia Semiconductor Index plunged more than 12%. On Kalshi, the probability that “Anthropic files an IPO application before November 30” fell from 84% to 60%, and the probability before the end of the year dropped from 90% to 74%. This selloff occurred after the midterm elections, when the Democrats won both the Senate and the House, turning data center construction from electoral rhetoric into real uncertainty at the approval level. Meta President Dina Powell McCormick said bluntly at the same WSJ Tech event: after the election results were announced, building data centers will “become more difficult.” OpenAI quickly put out a data center policy proposal on Thursday, promising to end confidentiality agreements with towns and to pay the full cost of the electricity and infrastructure it uses. This is an AI company's immediate adaptation to a change in the political environment, but a policy proposal can be written overnight; the approval cycle and construction cycle for data centers cannot.

Anthropic's IPO is widely seen as a demand test for the entire AI sector: if the public market is willing to accept, at a price of $1.5 trillion to $2 trillion, a lab with quarterly revenue of $17 billion and committed spending of nearly $900 billion, then OpenAI's listing path will have a reference anchor. Conversely, if Anthropic's pricing is cold, what OpenAI faces is not a question of “whether it will be ready in 2027,” but a question of “under what conditions and within what length of window it can complete pricing.” Brockman said it would go public at “the right time,” and the passive voice of that sentence is accurate: whether the timing is right is not entirely determined by OpenAI's degree of readiness, but also by how much patience the public market is willing to pay at that moment for an AI lab's losses and commitments.

Brockman's candor deserves to be taken seriously. He acknowledged falling behind in the first half of the year, acknowledged that Codex “no one talks about,” and acknowledged that the enterprise was still regarded as a consumer company a year ago. These admissions give a certain credibility to the “Q3 inflection point” narrative. But when the president of a company says “ready” after its private financing shrank by one-third, after a competitor's IPO window was punched through by the market, and when the company's cumulative funding gap is counted in hundreds of billions of dollars, the weight carried by that sentence far exceeds what “enterprise revenue doubled” can support.

2027 is not a date. It is a condition. Whether the condition is met is not up to Brockman; it is up to the public market.

This article comes from the WeChat public account “Tech Business,” by Zisen and Bowen, and is published by 36Kr with authorization.