A New Leader in Corporate AI Spending
Corporate AI spending patterns just delivered a clear signal: Anthropic has edged ahead of OpenAI in business adoption, at least when companies vote with their credit cards. Expense platform Ramp reports that Anthropic now accounts for 34.4% of tracked corporate AI purchases, surpassing OpenAI’s 32.3% share for the first time. The dataset covers more than 50,000 companies buying AI as standalone SaaS-style tools, not bundled cloud services or giant master contracts. This nuance matters for interpreting enterprise AI adoption. The numbers highlight where businesses are making explicit purchasing decisions around AI vendors, rather than passively accepting whatever is embedded in existing platforms. In that arena, Anthropic’s rise from roughly 9% to the mid-30s within a year signals a rapid rebalancing in the business AI market and raises fresh questions about Anthropic vs OpenAI competitive positioning.
How Anthropic Won Technical Teams—and Then the Boardroom
Anthropic’s surge is not an accident; it reflects a deliberate go-to-market strategy built around technical power users. Ramp’s analysis notes that Anthropic focused first on highly technical customers in sectors such as finance, technology, and professional services, tailoring Claude to demanding workflows before broadening out with more accessible tools like Cowork. That beachhead is paying off: Anthropic reportedly wins about 70% of first-time, head-to-head evaluations against OpenAI among new AI buyers. For enterprise AI adoption, these early technical champions wield outsized influence. Engineering, data science, and quantitative teams often run initial pilots and then recommend preferred vendors to executives. As those teams standardise on Claude, procurement patterns follow, helping explain Anthropic’s steep climb and OpenAI’s recent share decline in Ramp’s index—including the platform’s largest single-month vendor drop, when OpenAI lost 1.5 percentage points in February.
Revenue Reality Check: OpenAI Leads, Anthropic Accelerates
Despite losing ground in discrete corporate AI spending, OpenAI remains ahead in overall revenue. According to figures reported by The Information, OpenAI generated USD 5.7 billion (approx. RM26.2 billion) last quarter, compared with Anthropic’s USD 4.8 billion (approx. RM22.1 billion). OpenAI’s income is diversified across its Codex coding assistant, growing enterprise sales, and paid ChatGPT subscriptions for 55 million paying users. Anthropic, however, is growing faster. Investor materials cited in reporting indicate Anthropic expects USD 10.9 billion (approx. RM50.1 billion) in revenue for the next quarter, more than doubling its previous result and potentially marking its first profitable quarter, with a projected operating profit of USD 559 million (approx. RM2.6 billion). While OpenAI is still operating at a loss, Anthropic’s path toward profitability, powered largely by enterprise demand for Claude, underscores shifting economics in the business AI market.
What the Shift Reveals About Enterprise AI Adoption
The Anthropic vs OpenAI battle now reflects broader dynamics in enterprise AI deployment rather than a simple winner-takes-all race. Ramp’s card-spend data excludes bundled cloud AI and mega deals, instead spotlighting situations where companies deliberately compare vendors for specific use cases—from coding and analytics to knowledge work and customer support. In those scenarios, enterprises increasingly treat AI providers like any other critical SaaS partner, scrutinising reliability, safety, support, model strengths, and integration paths. Anthropic’s rapid share gains suggest many organisations are willing to experiment beyond the most well-known brand, especially when technical teams perceive advantages in model behaviour or tooling. At the same time, analysts caution that the current lead could be fluid. Competitive responses, large enterprise contracts, and platform bundling may still reshape the landscape. For now, though, the spending data shows a more pluralistic business AI market than many expected.