For the first time, more businesses are paying for Anthropic’s AI than OpenAI’s. That is the headline finding from Ramp’s May 2026 AI Index, which tracks actual spending data across thousands of companies on the Ramp platform.

The numbers are striking. Anthropic’s business adoption rose 3.8 percentage points in April to reach 34.4% of companies. OpenAI’s fell 2.9 points to 32.3%. One year ago, OpenAI sat at around 32% and Anthropic was below 8%.

Anthropic has effectively quadrupled its business footprint in twelve months while OpenAI’s growth barely moved.

What Is Driving the Shift

The engine behind almost all of Anthropic’s growth is one product: Claude Code. Anthropic’s agentic coding tool has become the fastest-growing product in the company’s history, and it is pulling in a specific type of business customer: development teams, technical operators, and companies building internal tools.

This is not casual ChatGPT usage migrating over. These are teams making deliberate infrastructure decisions to build workflows on top of Claude. Claude Code lets developers write, test, and deploy code through an AI-powered command-line tool that operates as a collaborator, not just an autocomplete engine. Companies building internal tools nobody hates are among the biggest adopters.

The growth is also showing up in Anthropic’s financials. The company’s annualised revenue has now passed $30 billion, up from around $9 billion at the end of 2025. That is a more than three-fold increase in roughly five months, and it is attracting attention at scale: Anthropic is reportedly in talks to raise between $30 billion and $50 billion in a new funding round that would value it at close to $1 trillion.

The Business Implications

This is a meaningful signal for any organisation evaluating which AI platform to build on.

OpenAI was the obvious default for most of 2024 and 2025. It had the brand recognition, the API ecosystem, and the enterprise partnerships. But business adoption is now moving on the basis of actual utility, not familiarity. And in this particular cycle, Claude Code’s ability to handle complex, multi-step agentic tasks is resonating with the teams building real things.

There is also a quality story here. Claude’s extended context window and its behaviour in complex reasoning tasks have consistently ranked well in head-to-head evaluations. Businesses that started with GPT-4 for simple use cases are now evaluating what platform they want to run their serious workloads on, and a meaningful number of them are choosing Claude.

That said, the shift is not guaranteed to hold. Ramp’s analysis notes that some of the fastest-growing vendors on its platform are AI inference platforms offering access to cheap open-source models. If cost becomes the dominant purchase criterion, both Anthropic and OpenAI face pressure from the commodity layer underneath them.

What This Means for Business

If you are evaluating AI tools for your organisation right now, the Ramp data tells you something practical: the developer community, which is often the first to move and tends to validate tools before broader rollout, has voted for Claude in a meaningful way.

For teams that are building agentic workflows, internal tools, or AI-assisted operations, Claude Code and the broader Claude API ecosystem deserve serious evaluation alongside GPT-4 and GPT-5. The defaults have shifted.

For business leaders who are not in the weeds of tooling decisions, the more important takeaway is structural: AI platform choice is now a real strategic decision with real cost and capability implications. The market is competitive enough that the right answer today might look different from what was obvious eighteen months ago.

At Enterprise DNA, we have been teaching data and AI skills to more than 220,000 professionals across 50 countries, and we have watched this shift happen in real time inside our own community. The teams making the most progress right now are the ones who treat AI platform evaluation as an ongoing process rather than a one-time decision.

The Ramp data is a useful prompt to revisit that evaluation if you have not done so recently.

---

Source: Ramp AI Index, May 2026. Additional reporting by TechCrunch and Axios.