OpenAI faces $47B revenue gap

OpenAI’s problem is not that demand disappeared. ChatGPT is still huge. Revenue is still huge. The problem is timing — and timing is brutal when your business depends on buying absurd amounts of computing power before the money fully shows up. That’s why the “$47 billion gap” framing matters. It’s not a claim that OpenAI suddenly lost $47 billion in actual revenue. It’s a way of describing the distance between the scale of infrastructure spending tied to OpenAI and the pace at which its commercial engine is proving it can support that spending. OpenAI has pushed back hard on the bleakest versions of that story, but the underlying tension is real. ### What actually went wrong? The clearest reported miss was on internal growth targets. OpenAI reportedly failed to hit an internal goal of 1 billion weekly active ChatGPT users by the end of 2025, and it also missed several monthly revenue targets earlier in 2026. The reported issue was not broad AI weakness so much as share loss in some valuable categories — especially coding and enterprise — where Anthropic and Google have been more aggressive. (cnbc.com) ### But isn’t OpenAI still enormous? Yes — and that’s what makes this tricky. OpenAI said on March 31 that it had closed a $122 billion funding round at an $852 billion post-money valuation, was generating $2 billion in revenue per month, and made $13.1 billion in 2025 revenue. ChatGPT also had more than 900 million weekly active users as of March. Those are monster numbers. But monster numbers do not automatically mean the business can comfortably pre-fund every data center, chip order, and cloud contract now being discussed around it. (forbes.com) ### Why does compute make this so dangerous? Because OpenAI’s commitments are not normal startup commitments. Oracle has been tied to a reported $300 billion, five-year computing deal for OpenAI starting in 2027. OpenAI also expanded a spending agreement by another $100 billion and has been racing to secure more capacity from Nvidia, AMD, Amazon, and others. Basically, the company is trying to lock in future AI supply before competitors do. (cnbc.com) That makes strategic sense. But it means the bills arrive on infrastructure timelines, not software timelines. ### Where does Broadcom fit in? Broadcom matters because custom chips are supposed to be one of the escape hatches. OpenAI and Broadcom announced plans in October 2025 to build and deploy 10 gigawatts of custom AI accelerators, with OpenAI arguing that custom silicon could lower compute costs over time. But a newer report says the first phase has hit an $18 billion financing snag, with Broadcom wanting Microsoft to commit to buying a big chunk of the chips before funding moves ahead. (cnbc.com) If that delay is real, OpenAI stays more dependent on Nvidia-style economics for longer. ### Why does Microsoft matter so much here? Because Microsoft used to be the tightest commercial wrapper around OpenAI. But the relationship changed on April 27. OpenAI can now serve customers across cloud providers, Microsoft’s license is no longer exclusive, and OpenAI’s revenue-share payments to Microsoft are capped, though the 20% share reportedly continues through 2030. That gives OpenAI more freedom to chase enterprise revenue wherever customers already are. (cnbc.com) The catch is that more freedom usually comes with more execution burden. ### So is this an IPO problem? Partly. Reports say Sarah Friar has worried internally about both funding future compute agreements and whether OpenAI is ready for public-company reporting standards on Sam Altman’s preferred timeline. That does not mean an IPO is dead. It means the old story — unbeatable product, infinite capital, inevitable listing — now has friction in it. (cnbc.com) ### What’s the real bottom line? OpenAI is not facing a demand collapse. It is facing a conversion problem. The company proved it can attract users and capital at historic scale. Now it has to prove something less glamorous — that it can turn that lead into predictable, fast-growing, enterprise-grade cash flow before its infrastructure appetite outruns its financing flexibility. (forbes.com)