Anthropic gains ground vs OpenAI
Private-market signals show buyer preferences fragmenting: one platform reports Anthropic has overtaken OpenAI in valuation, and another report estimates Anthropic’s annual revenue above $30bn versus OpenAI’s roughly $24bn annualised, suggesting institutional buyers are diversifying away from a single-model narrative. That shift matters for product vendors because agencies and buyers are increasingly judging models by task fit and reliability rather than brand alone. (officechai.com | socialsamosa.com)
A year ago, the easy story was that one company had already won generative artificial intelligence for business. In April 2026, that story looks shakier: Anthropic says it has crossed a $30 billion annualized revenue run rate, while OpenAI was recently reported around $2 billion a month, or roughly $24 billion annualized. (anthropic.com) (officechai.com) Private markets are reflecting that shift too. OpenAI announced a $300 billion post-money valuation on March 31, 2025, and Anthropic announced a $380 billion post-money valuation on February 12, 2026. (openai.com) (anthropic.com) These numbers are not the same kind of score. Valuation is what investors think a company could be worth later, while annualized revenue is a snapshot that takes a recent month or quarter and stretches it across a full year. (openai.com) (officechai.com) The buyer shift is showing up in spending data, not just fundraising decks. Ramp says its Artificial Intelligence Index tracks card and invoice spend across more than 50,000 United States businesses, and OfficeChai reports that the industries using artificial intelligence most heavily are leaning more toward Anthropic than OpenAI. (officechai.com) That pattern is a clue about where the money is coming from. OpenAI became the household name through ChatGPT, which reached mass consumer scale, while Anthropic has spent much of the last year pushing Claude deeper into coding, application programming interface usage, and enterprise workflows. (openai.com) (cnbc.com) Coding has become one of the clearest battlegrounds. CNBC reported on April 11, 2026 that Claude Code, launched to the general public in May 2025, was generating more than $2.5 billion in annualized revenue as of February. (cnbc.com) That helps explain why the market no longer looks like a one-brand race. A company buying an artificial intelligence model for customer support, legal review, or software development is starting to act less like a fan choosing a smartphone brand and more like an information technology team choosing separate tools for email, payroll, and cloud storage. (officechai.com) (cnbc.com) OpenAI is not disappearing from that picture. It still said in its March 2025 funding update that ChatGPT serves 500 million weekly users, and Bloomberg reported on April 10, 2026 that OpenAI has been telling investors its early push to secure more computing power is still a core advantage. (openai.com) (bloomberg.com) What changed is the shape of the contest. Instead of one company owning “artificial intelligence” in the abstract, buyers are splitting spend across models, and the winner in one task is not automatically the winner in the next. (officechai.com 1) (officechai.com 2) For software vendors and agencies, that means the sales pitch is getting narrower and more concrete. “Best model” is giving way to questions like which model writes production code with fewer errors, which one follows policy rules more reliably, and which one is cheaper to run at scale. (officechai.com) (siliconangle.com) Anthropic’s rise does not prove OpenAI is losing. It shows the market is getting more normal: one giant consumer brand, one fast-rising enterprise rival, and customers who are no longer treating model choice as a one-time bet. (openai.com) (anthropic.com)
