Anthropic gains ground as OpenAI’s enterprise share falls to ~27%

Enterprise AI is where the model wars stop being vibes and start being revenue. That is why this shift matters. Anthropic now appears to be ahead of OpenAI with paying enterprise model customers, even though OpenAI still looks much bigger overall. The gap is basically between broad consumer reach and the part of the market that tends to be stickier, higher value, and easier to forecast. (menlovc.com) ### What actually moved? The clearest new data point comes from Menlo Ventures’ 2025 enterprise AI report. It estimates Anthropic at 40% of enterprise LLM spend, up from 24% a year earlier and 12% in 2023. OpenAI, which held 50% in 2023, fell to 27%. Google also gained share, which makes this look less like one-off churn and more like a real broadening of the market. (menlovc.com) ### What does “enterprise share” mean here? It does not mean total AI revenue. It means share of enterprise spending on foundation models — basically the money companies spend to power products and internal tools with model APIs. That is a narrower slice than all of OpenAI’s business, because OpenAI also has big consumer revenue through ChatGPT subs(menlovc.com)It means Anthropic is winning a very important lane. (menlovc.com) ### Why would enterprises be shifting? Coding is a big reason. Anthropic has built a strong reputation with developers, and that tends to spread inside companies fast — one team adopts it, then another team standardizes on it. Menlo’s report also says enterprises are spending more on applications that deliver immediate productivity gains, not just o(menlovc.com)l workflows. (menlovc.com) ### Does this mean OpenAI is in trouble? Not exactly. OpenAI still said on March 31 that it closed a $122 billion funding round at an $852 billion post-money valuation. It also still appears to be operating at huge scale, with multiple reports pegging annualized revenue around $25 billion. But the catch is that a company valued like that gets judged(menlovc.com) asking whether the strongest margins and stickiest customers are moving elsewhere. (openai.com) ### Why are people suddenly more sensitive to that? Because growth expectations are extreme. Reuters, citing the Wall Street Journal, said OpenAI recently fell short of some internal user and revenue goals. That does not erase the company’s scale. But it changes the tone. Missing targets is one thing when you are a fast-growing startup. It is another when the market is implic(openai.com)economy. (msn.com) ### Is Menlo’s data the whole story? No — and that is important. Menlo is measuring enterprise spend through its survey and market model, not publishing audited revenue for every vendor. It is useful directional evidence, not the final ledger. Also, Menlo has exposure to the AI startup ecosystem, so t(msn.com)th broader reporting that Anthropic has become unusually strong with enterprise buyers. (menlovc.com) ### Why does enterprise share matter more than headline hype? Because enterprise customers are usually harder to win and slower to leave. If a model gets embedded into support tools, coding assistants, search, compliance, and internal workflows, that revenue can compound. Think of consumer AI as attention and enterprise AI as plumbing. Attention can(menlovc.com) is not just a ranking change — it hints at where the durable control points may end up. (menlovc.com) ### Bottom line? Anthropic taking the lead in enterprise model spend does not dethrone OpenAI everywhere. But it does puncture the easy story that scale alone will lock the market. Right now, the battle looks less like “who has the most users?” and more like “who becomes the default system inside companies?” That is the revenue stream everyone cares about. (menlovc.com)