Anthropic overtakes OpenAI in business AI spending for the first time

Anthropic overtook OpenAI in business AI spending for the first time, with Ramp's AI Index showing Anthropic at 43.5% of U.S. businesses on its platform paying for subscriptions or tokens in July 2026, compared to OpenAI's 39.7%. Anthropic's business adoption quadrupled over the past year, driven by Claude Code, which captured 54% of the enterprise AI coding market, while OpenAI's Codex held 21%. However, Anthropic's higher pricing is a hurdle: its Fable 5 model, priced at about $10 per million tokens, accounted for only 11.4% of Anthropic model spend in its first month, versus GPT-5.6 Sol's 23% of OpenAI model spend.

Anthropic has moved past OpenAI in business AI spending, but the win comes with a harder question: whether companies will keep paying for the most expensive models. For most of the last three years, the assumption in Silicon Valley ran one way: OpenAI owned the enterprise wallet, and Anthropic was the challenger with a stronger reputation among developers. That changed fast. According to Ramp's AI Index, which tracks corporate card and invoice payments across its customer base, Anthropic passed OpenAI in business adoption share in April 2026, at 34.4% to 32.3%. By July, Ramp said 43.5% of U.S. businesses on its platform paid for Anthropic subscriptions or tokens, compared with 39.7% for OpenAI. The trend line matters more than the snapshot. Anthropic's business adoption has roughly quadrupled over the past year, while OpenAI barely moved on Ramp's measure. That's the point. One company is still huge, but the other is taking the business accounts that decide which models get built into daily work. Claude Code is the center of the story Menlo Ventures' enterprise research tells the same story from another angle. Its survey put Anthropic at roughly 40% of enterprise LLM API spend by late 2025, ahead of OpenAI at 27% and Google at 21%. Go back to 2023 and OpenAI controlled more than half the market, with Anthropic near 12%. Few companies lose that kind of lead by accident. Claude Code is the obvious reason. Menlo's mid-2026 AI coding work put Anthropic's share of the enterprise AI coding market at 54%, up from 42% six months earlier, while OpenAI's Codex sat near 21%. If you're buying AI inside a company, coding is where the bill becomes real first. Engineers use the tools all day, the usage compounds, and finance sees the result on the invoice. Microsoft's AI security overhaul sets a benchmark every enterprise security startup now has to beat https://startupfortune.com/microsofts-ai-security-overhaul-sets-a-benchmark-every-enterprise-security-startup-now-has-to-beat/ Microsoft's AI security overhaul sets a benchmark every enterprise security startup now has to beat - microsoft ai security vulnerabilities patch tuesday https://startupfortune.com/microsofts-ai-security-overhaul-sets-a-benchmark-every-enterprise-security-startup-now-has-to-beat/ - enterprise security startups competitive benchmark https://startupfortune.com/microsofts-ai-security-overhaul-sets-a-benchmark-every-enterprise-security-startup-now-has-to-beat/ Big contracts followed the usage. Anthropic said in May 2026 that its run-rate revenue had crossed $47 billion earlier that month, after announcing a $65 billion Series H round at a $965 billion post-money valuation. In February, the company had raised $30 billion at a $380 billion valuation. Those are extraordinary numbers. They also explain why this adoption fight is no longer a side plot in the AI race. None of this came free. Frankly, Anthropic's pricing is starting to work against it. The Information reported that Uber blew through its full 2026 AI budget in a few months after pushing engineers to use AI tools heavily, especially Claude Code. Fortune later noted that Uber CTO Praveen Neppalli Naga said the company had to treat AI efficiency as an engineering problem, not simply a budget problem. That's a warning for any buyer watching its developers fall in love with agentic coding. The best model is not always the easiest sell Ramp's August update adds the wrinkle investors should not ignore. Anthropic's Fable 5 was described by Ramp as the most capable model on the market, yet it made up only 6% of tokens businesses purchased from Anthropic in its first month and 11.4% of dollars spent on Anthropic models. OpenAI's GPT-5.6 Sol took 25% of OpenAI tokens and 23% of OpenAI model spend over the same period. Price matters. Fable 5 costs roughly $10 per million tokens, about twice the rate Ramp cited for GPT-5.6 Sol. Businesses like Anthropic. They are still hesitant to pay up for the highest-end version when a cheaper rival is close enough for many jobs. That is not a knock on Claude. It is how procurement works once the first wave of excitement turns into a monthly line item. The timing lines up with Anthropic's push toward public markets. The Financial Times reported that the company is planning a potentially record-breaking IPO in October 2026. Axios recently reported something bigger too: an annualized revenue run rate past $65 billion. The Ramp and Menlo numbers back that story up: Anthropic is winning business adoption, not just developer praise. Investors seem convinced. But if you're reading this as a buyer, not an investor, the lesson cuts differently. The company that wins your engineering team can also become the one that blows through your budget. That's the trade. Anthropic took the enterprise lead because people are using its tools for real work, day in, day out. Now it has to prove the spend keeps paying for itself. Monad's 1.2 Million-Follower X Account Was Suspended Overnight and the Crypto Community Had Feelings https://startupfortune.com/monads-12-million-follower-x-account-was-suspended-overnight-and-the-crypto-community-had-feelings/ Monad's 1.2 Million-Follower X Account Was Suspended Overnight and the Crypto Community Had Feelings - crypto project x account suspension overnight https://startupfortune.com/monads-12-million-follower-x-account-was-suspended-overnight-and-the-crypto-community-had-feelings/ - why did monad's twitter account get suspended https://startupfortune.com/monads-12-million-follower-x-account-was-suspended-overnight-and-the-crypto-community-had-feelings/ Also read: Nevada clears Tesla, Uber and Waymo to run up to 8,000 robotaxis in Las Vegas https://startupfortune.com/nevada-clears-tesla-uber-and-waymo-to-run-up-to-8000-robotaxis-in-las-vegas/ • A Mystery Model Called Ox Alpha Just Topped Coding Benchmarks For Free https://startupfortune.com/a-mystery-model-called-ox-alpha-just-topped-coding-benchmarks-for-free/ • How Vendor-Locked AI Coding Agents Are Quietly Raising Your Engineering Costs https://startupfortune.com/how-vendor-locked-ai-coding-agents-are-quietly-raising-your-engineering-costs/