OpenAI Is Behind Anthropic on Revenue, Missed Its 1B User Target, and Its CFO Is Worried About the Compute Bill
The Wall Street Journal reported on April 28 that OpenAI missed multiple internal monthly revenue targets in the opening months of 2026 and failed to reach its self-imposed goal of 1 billion ChatGPT weekly active users by the end of 2025. The company currently sits at roughly $24B in annualised revenue — trailing Anthropic’s $30B for the first time in the company’s history.
The CFO is not quiet about the implications. Sarah Friar has told senior colleagues she is concerned OpenAI may not be able to honour future computing contracts if revenue growth does not accelerate. That is not a minor concern: OpenAI has signed commitments totalling more than $600 billion in data centre and chip spend across Oracle ($300B five-year deal), AWS ($100B addendum), Microsoft ($250B through 2032), CoreWeave, Google Cloud, and the broader Stargate consortium. The Stargate ramp alone is described internally as a 7-gigawatt buildout that is already ahead of schedule.
The Competitive Shift That Triggered It
The revenue miss is not a random stumble. It tracks a well-documented market share reversal in the two segments that drive AI revenue: coding tools and enterprise API.
Twelve months ago, OpenAI held roughly 42% of the coding assistant API market. Anthropic held 12%. By April 2026 those numbers have inverted: OpenAI at 31%, Anthropic at 42%. Claude Code’s ARR crossed $2.5B in February and has more than doubled since January. Enterprise customers — Fortune 500 companies spending $1M or more annually — doubled at Anthropic in under 60 days.
Google Gemini has applied parallel pressure on the consumer side. SimilarWeb puts Gemini at over 25% of AI chatbot traffic against ChatGPT’s 50%, down from a 90% OpenAI dominance in 2024.
What the Numbers Actually Mean
OpenAI closed a $122B funding round at an $852B valuation in late March. Four weeks later the WSJ disclosed it has been missing its own monthly sales targets. That gap between fundraise narrative and internal dashboard is what sent OpenAI-adjacent equities lower on April 28: Oracle fell more than 6%, NVIDIA and AMD dropped 2-5%, CoreWeave slid 3.5%.
Sam Altman and Friar issued a joint statement calling the WSJ report “ridiculous” and pointing to the $122B raise as evidence markets remain confident. But the structural problem the CFO has privately identified does not go away with a press statement: a $20B 2026 revenue target against more than $400B of three-year infrastructure spend leaves essentially no room for the growth deceleration that the numbers currently show.
Friar has also reportedly flagged that OpenAI is not yet ready to meet public company financial reporting standards, which complicates the planned late-2026 IPO.
The Coding Market Is Now the Fulcrum
The highest-margin, stickiest AI revenue in 2026 comes from enterprise coding contracts — multi-year, volume-based agreements that compound once engineers build workflows around a specific API. OpenAI’s consumer-first distribution strategy, which made sense in 2023, optimised for the wrong metric. Anthropic’s API-first approach captured the contracts.
The Microsoft-OpenAI pact restructuring announced the same week — ending exclusivity and allowing OpenAI to sell via AWS and Google Cloud — is partly a response to this dynamic. It widens OpenAI’s distribution without fixing the underlying share problem in the enterprise coding segment.
Key Numbers
| Metric | OpenAI | Anthropic |
|---|---|---|
| ARR (April 2026) | ~$24B | ~$30B |
| WAU (ChatGPT) | ~800M | ~134M MAU |
| Coding assistant market share | 31% | 42% |
| Claude Code ARR | — | $2.5B+ |
| Enterprise $1M+ customers | undisclosed | 1,000+ |
| Planned compute commitments | $600B+ | $40B (Google) |
The company that made AI mainstream is now the follower in the market segment that determines enterprise cash flow. That is a structural problem, not a quarterly blip, and it is now officially the story ahead of the IPO.