For the past two years, the narrative around enterprise AI adoption has been remarkably consistent. OpenAI, with its ChatGPT Enterprise product and the formidable Microsoft distribution pipeline, was assumed to be the default choice for businesses looking to integrate large language models into their workflows. The company had first-mover advantage, brand recognition, and the most powerful models. What else did it need?
According to data from Ramp, the corporate expense management platform that processes billions of dollars in business spending, the answer is: more than OpenAI has. In a surprising disclosure that is already reshaping how investors and analysts think about the AI market, Ramp reported that Anthropic now has more business customers on its platform than OpenAI. Not slightly more. Meaningfully more. And the trend is accelerating.
This is not a user satisfaction survey or a theoretical market analysis. This is payment data — the kind of data that cannot be gamed, massaged, or spun. Companies using Ramp's corporate cards and expense software have voted with their wallets, and they are voting for Claude.
The Data That Changes Everything
Ramp's dataset is uniquely valuable for understanding enterprise AI adoption because it sits at the intersection of procurement and usage. When a company subscribes to an AI service, the transaction flows through Ramp's platform. When that same company renews, upgrades, or cancels, Ramp sees that too. The platform processes expenses for tens of thousands of businesses, from seed-stage startups to public companies, giving it a near-real-time view of how the corporate world is actually spending money on AI.
The headline finding is striking: Anthropic has overtaken OpenAI in total business customers on the Ramp platform. The crossover happened recently, and the gap appears to be widening rather than narrowing. For a company that was widely perceived as the underdog in the enterprise race — technically sophisticated but commercially naive compared to OpenAI's aggressive sales machine — this is a remarkable turnaround.
What makes the data particularly credible is that Ramp has no obvious incentive to favor one AI company over another. The platform is neutral infrastructure. It makes money from payment processing and expense management, not from AI partnerships. If anything, OpenAI's deeper pockets and more aggressive marketing might have been expected to produce an advantage in Ramp's data. The fact that Anthropic is winning anyway suggests something genuine is happening in the market.
Why Claude Is Winning the Enterprise
The obvious question is why. Why would businesses, many of which started their AI journeys with ChatGPT, switch to or prefer Claude? The answer, according to the patterns visible in Ramp's data and the qualitative feedback from customers, is a combination of technical differentiation, pricing strategy, and a fundamentally different approach to the enterprise relationship.
On the technical side, Anthropic has invested heavily in areas that matter disproportionately to business users. Claude's extended context window — capable of handling hundreds of thousands of tokens in a single conversation — is not just a benchmark bragging point. It is a practical advantage for enterprises dealing with long documents, complex contracts, extensive codebases, and detailed financial reports. A lawyer analyzing a merger agreement, a consultant reviewing a due diligence binder, or a developer debugging a sprawling codebase benefits enormously from being able to feed the entire document into the model rather than chopping it into fragments.
Anthropic has also been more aggressive about building features specifically for business workflows. Claude's artifact system, which allows the model to generate and display structured outputs — code, documents, spreadsheets, diagrams — alongside the conversational interface, transforms the tool from a chatbot into a genuine productivity platform. Businesses are not looking for a better search engine or a more eloquent email writer. They are looking for systems that can produce work products, and Claude's architecture is designed with that use case in mind.
The Pricing Advantage
Pricing has also played a significant role. OpenAI's enterprise offerings, while powerful, have been perceived by many customers as expensive relative to the value delivered. The company's strategy of capturing maximum value from its technological lead made sense when it was the only game in town, but the competitive landscape has changed. Anthropic has priced Claude aggressively for enterprise customers, often undercutting OpenAI by significant margins while delivering comparable or superior performance on the tasks that matter most to businesses.
For a finance team managing software spend across dozens or hundreds of seats, the difference between $50 per user per month and $30 per user per month adds up quickly. And when the cheaper option is also the one that handles longer documents better and produces more structured outputs, the business case becomes obvious. Ramp's data shows that companies are not just trying Claude as an experiment. They are committing to it at scale, with multi-year contracts and seat counts in the hundreds.
The Trust Factor
There is also a less tangible but equally important factor at play: trust. Anthropic has cultivated a reputation for transparency, safety research, and responsible AI development that resonates with enterprise risk managers and compliance officers. In an era where regulators are increasingly scrutinizing AI deployments, and where high-profile failures can generate headlines that damage brands, companies want to work with AI providers that take safety and ethics seriously.
OpenAI, by contrast, has struggled with trust issues. The company's internal turmoil in 2023, the public disputes between its leadership and safety-focused researchers, and its pivot toward a more aggressive commercial posture have created a perception — fair or not — that OpenAI prioritizes growth over responsibility. For a procurement team evaluating AI vendors, that perception matters. It shows up in due diligence questionnaires, in board presentations, and in the final vendor selection meetings where intangibles tip the balance.
Anthropic's constitutional AI approach, its public commitment to safety research, and its more measured public communications have created a brand identity that enterprise buyers find reassuring. The company is seen as the thoughtful alternative to OpenAI's brash disruptor — and in many corporate environments, thoughtfulness wins.
What OpenAI Is Doing About It
OpenAI is not blind to this competitive pressure, and the company is responding on multiple fronts. It has accelerated the development of enterprise-specific features, including enhanced data privacy controls, admin dashboards for managing team usage, and integrations with popular business software platforms. It has also adjusted its pricing strategy, introducing more granular tiers and volume discounts designed to compete with Anthropic's aggressive enterprise rates.
Most importantly, OpenAI is leveraging its relationship with Microsoft to maintain distribution advantages that Anthropic cannot easily replicate. Microsoft Copilot, which integrates OpenAI's models directly into Office 365, Teams, and the Windows operating system, represents a distribution channel that no standalone AI company can match. For businesses already embedded in the Microsoft ecosystem, the path of least resistance is to use the AI tools that come with the software they already pay for.
But Ramp's data suggests that this distribution advantage is not as decisive as Microsoft and OpenAI had hoped. Companies are willing to go outside the Microsoft ecosystem and adopt standalone AI tools if the value proposition is strong enough. Claude's technical advantages, particularly around context length and structured outputs, appear to be compelling enough to overcome the friction of adding another vendor to the stack.
The Broader Implications
The Anthropic-OpenAI enterprise rivalry is not just a two-horse race, and Ramp's data provides glimpses of how the broader competitive landscape is evolving. Google's enterprise AI offerings, delivered through Gemini for Workspace and Vertex AI, are gaining traction, particularly among companies that are already committed to Google Cloud. Microsoft's own AI strategy is increasingly decoupled from exclusive reliance on OpenAI, with the company developing its own models and deepening partnerships with other AI labs.
What the Ramp data reveals is that the enterprise AI market is more competitive and more fluid than the headlines suggest. The assumption that OpenAI's first-mover advantage and technical leadership would translate into permanent market dominance is being tested in real time. Companies are evaluating AI tools based on the specific value they deliver for specific workflows, not based on brand recognition or parameter counts. And in that evaluation, Anthropic is winning.
What Investors Should Watch
For investors trying to understand the AI market, Ramp's disclosure is a valuable signal — and a warning. Valuations in the AI sector have been driven largely by assumptions about market share and pricing power that may not hold up as the competitive landscape matures. OpenAI's reported $300 billion valuation, and the even more speculative valuations of other AI labs, depend on the assumption that these companies will capture and retain dominant positions in enterprise AI. If Anthropic can overtake OpenAI in business customers while operating with a fraction of the funding and staff, that assumption deserves scrutiny.
The more nuanced investment thesis is that the enterprise AI market will support multiple large winners, each with different strengths and different customer bases. OpenAI may dominate in creative applications and consumer products. Anthropic may lead in document-heavy enterprise workflows. Google may capture the cloud-native segment. And Microsoft may win by default in organizations that prioritize integration over best-of-breed capabilities.
What to Watch Next
Ramp has committed to updating its AI adoption data quarterly, which will provide an ongoing window into how the enterprise market is evolving. The next update, expected in three months, will be particularly telling. If Anthropic's lead continues to widen, it will validate the thesis that technical differentiation and enterprise-focused product development can overcome distribution disadvantages. If OpenAI reverses the trend, it will suggest that the company's recent enterprise investments are paying off and that the Copilot integration strategy is working.
Also watch for pricing moves from both companies. Enterprise software markets tend toward price competition as they mature, and AI is unlikely to be an exception. If Anthropic and OpenAI engage in a price war for enterprise customers, margins will compress and the path to profitability for both companies will lengthen. The companies that can maintain pricing power will be those that deliver genuinely differentiated value, not just comparable performance at a lower cost.
Finally, watch the regulatory environment. Enterprise AI adoption is happening against a backdrop of increasing government scrutiny, particularly around data privacy, algorithmic bias, and competitive practices. The companies that navigate this scrutiny most effectively — building compliance into their products rather than treating it as an afterthought — will have advantages that are not visible in current market share data but will matter enormously over the long term.
The AI market is still young, and today's leader may not be tomorrow's. What Ramp's data makes clear is that the enterprise race is far from over. And the company that everyone assumed would win? It just fell behind.