In February 2026, OpenAI's annualized run rate hit $25 billion. That sounds spectacular — until you learn the company's internal target for full-year 2026 sat around $30 billion, and the run rate barely moved from February through April, according to a Wall Street Journal report in late April. At the same time, Anthropic, which had just raised its own internal target in January to "as much as $18 billion" for the year per The Information, was already posting a $47 billion run rate by May, then $69 billion by July according to research firm Yipit's estimates. Two companies both described in the press as the fastest-growing software businesses in history, telling two very different stories about how close they've come to the plans they set for themselves.
OpenAI: Chasing a Target That Keeps Moving Away
OpenAI jumped from $3.7 billion in revenue in 2024 to a $21.4 billion run rate by the end of 2025 — growth that, by most measures, no company in tech history has matched at that pace. But momentum cooled once 2026 began. First-quarter 2026 revenue came in at $5.7 billion, which annualizes to roughly $22.8 billion — short of the full-year $30 billion target. Reuters reported OpenAI missed monthly revenue targets through early 2026, and net new ARR growth slowed from a peak of $2 billion a month to around $1.6 billion a month.
The mid-term target is even more ambitious: $62 billion in ARR by mid-2027. Independent research firm FutureSearch projects that figure is more likely to land near $44 billion — a third short of the internal goal. Margins remain deeply negative in the meantime: a -122% operating margin, with projected 2026 losses around $14 billion, though some estimates put that figure as high as $33 billion depending on the accounting method used. ChatGPT itself crossed 1 billion monthly active users in June 2026, but that scale of user base hasn't translated into revenue growth keeping pace with the company's own targets.
Anthropic: Clearing the Bar Before the Year Even Ends
The story runs in the opposite direction for Anthropic. The company closed 2025 with a run rate around $9 billion, then in late January 2026 raised its own internal forecast for the year to "as much as $18 billion" — already an aggressive number at the time. In reality, by mid-May 2026 its run rate had already hit $47 billion, according to Sacra estimates cited by CNBC, and growth kept accelerating: the average daily ARR addition climbed from roughly $400 million to $550 million a day between May and July. By July, the run rate had reached $69 billion — nearly four times the target the company had set for itself just six months earlier — and FutureSearch has since revised its full-year projection to a range of $75-90 billion.
The main driver is Claude Code, the coding product launched in May 2025. Its ARR crossed $500 million within its first three months, then grew fivefold to $2.5 billion by February 2026. Anthropic is also reportedly closing in on profitability — CNBC reported a projected $559 million operating profit for the second quarter of 2026 if the company hits its $10.9 billion quarterly revenue target, something no frontier AI lab has achieved yet. Its valuation followed suit, reaching $965 billion in May 2026, overtaking OpenAI's $852 billion.
Why the Gap Is So Wide
The difference isn't about which company builds smarter models — both are considered competitive on that front. It comes down to revenue composition. Roughly 70% of OpenAI's revenue comes from ChatGPT subscriptions — closer to a gym membership model: you need a constant stream of new members or price increases on existing ones who are sensitive to cost hikes, and both levers are slow and carry cancellation risk. Anthropic's revenue, by contrast, is 70-75% API consumption by businesses and developers — closer to a metered utility bill. As more enterprises embed Claude into coding workflows or internal tools they've already adopted, revenue climbs largely on its own, without needing to win a new customer every month. Menlo Ventures data backs this pattern: Anthropic holds roughly 40% of enterprise LLM spend, versus OpenAI's 27%.
For readers weighing exposure to AI-related stocks — whether directly, through the public listings both companies are preparing (OpenAI filed a confidential S-1 on June 8, 2026; Anthropic filed a week earlier, on June 1), or indirectly through Microsoft, Amazon, or Nvidia — consumption-based revenue tends to be far easier to project forward than volatile, subscription-driven consumer revenue.
Big Targets, Slower Adoption
Zoom out to the market level and the numbers get even more tempting. Gartner projects agentic AI spending will reach $201.9 billion in 2026, and Bloomberg Intelligence expects the generative AI market to hit $1.3 trillion over the next eight years. McKinsey puts the potential economic impact of generative AI as high as $4.4 trillion annually worldwide.
But there's a note of caution that rarely makes the headlines. McKinsey itself found that only 23% of organizations have actually scaled their AI agent deployments past the pilot stage, and Gartner projects 40% of agentic AI projects will be cancelled before 2027 due to ballooning costs or unclear business value. The money flowing into the sector is real and large, but the gap between "how big the market is projected to be" and "how many companies are actually capturing value from it" remains wide — the exact same pattern playing out between these two AI giants, where the target on paper and the outcome on the ground can move in entirely different directions.
If you're reading a revenue projection from an AI company — whether for an investment decision, a career decision, or just to sanity-check a headline — the more useful question isn't "how big is the number." It's where that revenue is actually coming from, and how much of its future growth depends on the company working harder to chase it versus simply happening on its own.