$25 billion is where OpenAI's revenue run rate has been stuck since February 2026. Anthropic, meanwhile, blew past it in April and hit $47 billion by May. That's the real OpenAI revenue story right now โ not the growth, but the stall.
For three straight years, "OpenAI's revenue just did something remarkable" was a safe headline to write. It went from roughly $1 billion annualized in mid-2023 to $3.7 billion for full-year 2024, to $13.1 billion in 2025, to $20-25 billion in early 2026 โ a trajectory nobody in enterprise software history has matched. Then, sometime around February, the curve went flat. It's still flat in July. And in the same stretch, Anthropic did to OpenAI's growth story what OpenAI did to everyone else's.
What is OpenAI's revenue and ARR in 2026?
OpenAI's annualized revenue run rate is approximately $25 billion as of mid-2026, up from $21.4 billion at the end of 2025 and $13.1 billion in actual full-year 2025 revenue. Monthly revenue is running near $2 billion. The company crossed $1 billion ARR in mid-2023, $10 billion by June 2025, and $25 billion by February 2026 โ then growth essentially stopped, holding flat through April.
To put the flattening in perspective: OpenAI added roughly $15 billion in new ARR between June 2025 and February 2026, an eight-month stretch. It has added close to zero in the five months since. For a company that raised money in 2025 at a valuation implying decades of continued hypergrowth, a five-month plateau at $25 billion is the kind of data point that shows up in every subsequent term sheet negotiation, whether or not OpenAI's bankers want it to.
Why did OpenAI's revenue growth stall at $25 billion?
The plateau matters because of what came before it. OpenAI's run rate more than doubled every few months for most of 2024 and 2025 โ $3.7 billion to $13.1 billion in a single year, then to $21.4 billion by year-end, then a further jump to $25 billion by February. That kind of compounding doesn't slow down for no reason. ChatGPT has essentially saturated its addressable consumer base at 900 million+ weekly active users and 50 million+ paying subscribers, per OpenAI's own disclosures, and the 9 million business users generating enterprise contracts haven't been converting to net-new ARR at the same pace enterprise sales teams promised investors during the company's $40 billion-plus fundraising rounds in 2025. Consumer growth has a ceiling. Enterprise growth takes longer to close than a subscription toggle.
Figures are 2023โ2026 annualized run-rate estimates blended from OpenAI public statements, Sacra, and FutureSearch financial tracking. The FebโJul 2026 flat line reflects reported run-rate holding near $25B across multiple data points.
OpenAI revenue 2026 vs Anthropic: who's actually winning ARR?
Anthropic ended 2025 with roughly $9 billion in ARR โ barely a third of OpenAI's $21-25 billion at the time. Then it did something no AI lab has ever done at this scale: it went from about $1 billion to $30 billion annualized in fifteen months, officially overtaking OpenAI on April 7, 2026, and continuing on to $47 billion by May. Claude's coding products and API-first enterprise motion, not a consumer chatbot, drove nearly all of it. OpenAI grew about 2x over the same period; Anthropic grew closer to 30x. I've written before about how Anthropic's valuation was catching up to OpenAI's โ revenue catching up, and passing it, is a much bigger deal.
OpenAI vs Anthropic: 2026 Revenue Trajectory
Sacra, Epoch AI, The AI Corner, July 2026 estimates
OpenAI still has more than double Anthropic's weekly active users and a much bigger consumer brand โ but on the revenue line that VCs and LPs actually underwrite valuations against, Anthropic is no longer the smaller lab.
The $14 billion problem: OpenAI's losses in 2026
Revenue is only half the OpenAI story. The company's internal projections show roughly $14 billion in net losses for 2026 against total spending near $22 billion โ OpenAI burned through $3.7 billion in Q1 alone against $5.7 billion in revenue, meaning it spent more than 65 cents of every revenue dollar just in that quarter. Compute is the driver: over $13 billion in committed spend to Microsoft Azure, Stargate infrastructure buildout, and roughly $4 billion in talent costs, with inference costs alone reportedly quadrupling year over year. OpenAI's own plan reportedly targets cash-flow profitability around 2029 at close to $125 billion in revenue โ a number more than 4x today's run rate โ while outside analysts including HSBC have publicly doubted the company turns a profit by 2030 at all, citing a roughly $207 billion funding gap between committed infrastructure spend and disclosed capital raised.
None of this is unique to OpenAI in isolation โ every frontier lab is burning cash to buy compute right now. What's unusual is the size of the gap between the loss and the run rate: $14 billion in losses against $25 billion in revenue is a worse ratio than OpenAI posted a year ago, even with monthly revenue nearly doubling since mid-2025. Growth used to be the cover story for the burn. With growth flat, the burn is just the burn.
Can the new ChatGPT ads business change the math?
The most interesting new data point in OpenAI's 2026 revenue mix isn't enterprise โ it's ads. OpenAI's ChatGPT ads pilot, running only in the free and Go tiers, surpassed $100 million in annualized revenue in under six weeks, with more than 600 advertisers already on the platform, even though fewer than 20% of eligible free-tier users saw an ad daily during the test window. Self-service ad buying is now expanding access to a much broader advertiser base. It's a small number relative to $25 billion in ARR today, but it's the fastest-scaling new revenue line OpenAI has launched since ChatGPT Plus, and it points at a third leg of the business โ consumer subscriptions, enterprise contracts, and now advertising โ that could matter more once consumer subscriber growth flattens for good. Enterprise, for context, already makes up more than 40% of total revenue and is on track to reach parity with consumer revenue by the end of 2026, a shift worth tracking on the AI Valuations Dashboard.
The Bottom Line:
OpenAI's revenue run rate has been stuck at $25 billion since February 2026 while Anthropic passed it and reached $47 billion. OpenAI is still burning roughly $14 billion a year to hold that ground, and the new ChatGPT ads business, however fast it's scaling, is a rounding error next to the compute bill. The growth story isn't over โ but for the first time since ChatGPT launched, it isn't OpenAI's alone anymore.
Track how frontier AI lab valuations are moving on the AI Valuations Dashboard and see which AI companies just crossed $1B on the Unicorn Tracker at Value Add VC. Originally published in the Trace Cohen newsletter.
Get VC data most people never see
โ 100% free
Weekly benchmarks, valuations, and fund data. Join 5,000+ investors. No spam.