OpenAI president Greg Brockman has said the company now earns the majority of its revenue from enterprise sales, rather than from individual subscribers. Speaking at a Wall Street Journal technology conference in Napa, California, he was asked whether OpenAI would be ready to go public in 2027. He answered, in his words, that he thinks so, giving the clearest public signal yet that business customers are now OpenAI’s financial centre.
An enterprise majority, without a number
Brockman gave no revenue total, no growth rate, and no split between products, so his claim is a direction of travel rather than an auditable figure. A majority share can be reached by enterprise sales growing, by consumer revenue flattening, or by both at once, and his answer does not distinguish them.
What the answer does establish is OpenAI’s own framing, presenting itself first as a supplier to businesses rather than a consumer brand. Consumer products remain the entry point in that picture, while business contracts are treated as the destination for revenue and growth.
The timing is part of the message, since OpenAI is racing to raise fresh funds to capitalise on recent growth after delaying earlier listing plans. A company preparing for public markets needs a story investors can model, and recurring business contracts offer steadier revenue than monthly subscriptions.
Brockman announced no filing, timetable, or venue, so readiness in his answer means a business approaching a shape that could support a listing. It does not mean a listing itself is imminent, and readers should not treat a stage remark as a financial result or a filing date.
Codex is the wedge against Anthropic
Brockman tied the enterprise push to competition with Anthropic, which he noted is likely to go public before the end of this year. He also pointed to OpenAI’s Codex coding agent, the product at the centre of the contest for engineering teams and their budgets.
Coding tools sit at the sharp end of that contest, because they are bought by engineering teams, embedded in daily work, and hard to remove. Winning that ground brings recurring seats and usage, plus a foothold from which to sell broader products into the same company.
That contest sharpened this week when Claude Code’s product leader moved to OpenAI to lead product management for its core products. The hire underlines how closely the two companies track each other’s enterprise roadmaps, and how talent moves with the market they share.
OpenAI has also built enterprise delivery muscle through large services partners, with one consultancy committing 25,000 staff to its enterprise work. Those partnerships matter because big customers rarely buy software alone, instead buying implementation, training, and support as one package.
The IPO race is now a two-company story
Anthropic’s expected listing before year end sets the pace Brockman was implicitly answering, since the first major AI lab to list sets the benchmark. The second lab to list will then be judged against that valuation, whether it welcomes the comparison or not.
Anthropic has already opened its own listing to retail investors through a major bank, giving public investors an early route into its shares. Its amended filings also show the scale of its infrastructure commitments, offering detail that OpenAI has so far given only in outline.
OpenAI, by contrast, has offered only a stage answer so far, leaving the market to weigh its enterprise claim without audited prospectus detail. The contrast cuts both ways, because waiting can strengthen the enterprise story while delaying lets a rival define how investors judge AI labs.
Brockman’s answer threads that needle, signalling confidence without committing to a date and keeping the company’s options open. Customers deciding between the two vendors will hear a practical message too, that OpenAI intends to be a durable, business-first supplier for the long term.
Why It Matters
If enterprise sales truly form the majority of OpenAI’s revenue, the company’s incentives shift in ways that affect every user, including free users. Business buyers demand security controls, compliance, uptime guarantees, and predictable pricing that consumer products can often postpone without penalty.
That shift should, over time, produce more sober product decisions and stronger enterprise features, with clearer support and roadmaps. It can also push consumer offerings toward serving as funnels into paid business tiers, a trade-off users should watch in future releases.
Readers should watch for the next concrete evidence, such as a filing, audited figures, or named enterprise deals, before treating the claim as settled. A stage remark is a signal of intent, not a financial result, and the gap between the two is where careful coverage must stay.
The remark also reframes the rivalry as a race for the same enterprise budget, fought through coding agents, partners, and public-market credibility. Whoever lists first will face the hardest questions about losses and infrastructure spending, but will gain a currency for acquisitions and talent.
Brockman’s short answer therefore does double duty, reassuring customers while warning rivals that OpenAI will not watch from the sidelines. Whether that confidence survives contact with a prospectus is the question the next year of enterprise results will have to answer.
Background
OpenAI rose to prominence with ChatGPT, which put conversational AI in front of hundreds of millions of people before the enterprise push began. Anthropic followed a parallel path with Claude and its coding tools, so both companies now compete for engineering teams as much as for benchmarks.
Source: Reuters report
0 Comments