Anthropic Gains on OpenAI in Enterprise Adoption as AI Competition Intensifies

Enterprise adoption of generative artificial intelligence is broadening, but revenue remains heavily weighted toward a small group of high-volume users. Unlike conventional software sold through fixed per-seat subscriptions, application programming interfaces from OpenAI and Anthropic are generally priced by usage. That model can rapidly turn AI-native businesses into major accounts, while leaving model providers exposed if a few customers reduce spending, switch suppliers or bring workloads in-house.

Ramp data published on Sept. 4, 2026, showed the top 1% of customers generated about 80% of enterprise revenue at both OpenAI and Anthropic. Cursor and GitHub Copilot together contributed roughly $1.2 billion toward Anthropic’s $5 billion revenue milestone in 2025, or nearly one quarter of the total. Financial services may produce the next wave of large users: 95% of firms reported broad AI adoption, while 80% expected spending to increase over the following 12 months.

Anthropic has positioned its most capable artificial-intelligence models for enterprises willing to pay for stronger reasoning and complex-task performance. Corporate buyers, however, are increasingly weighing inference costs and measurable returns alongside benchmark results. As lower-priced tools improve and attract more users, Anthropic’s flagship offering faces pressure to prove that its performance advantage is large enough to justify a premium.

Recent reports say Anthropic’s top model is struggling to gain traction among business users, with some customers switching to cheaper alternatives as the company prepares for an initial public offering. The available report headlines do not specify an IPO date, valuation, model prices or the number of customers moving away. Still, the shift highlights softer enterprise demand and intensifying commercial pressure from lower-cost AI competitors.

Ramp’s AI Index uses spending data from customers of the U.S. expense-management platform to track whether businesses are actually paying for artificial intelligence products, offering a different signal from consumer traffic or app downloads. Anthropic’s lead matters because enterprise adoption can generate recurring, higher-value revenue, while purchasing decisions increasingly weigh model performance against security, integration requirements and the cost of deploying AI at scale.

The latest Ramp data showed Anthropic’s paid enterprise adoption rate remained ahead of OpenAI’s for a fourth consecutive month, though newer readings indicated OpenAI was gaining ground. Ramp did not disclose an average dollar amount spent by the companies in the survey. The report also found limited appetite for premium-priced flagship models, with some businesses shifting toward lower-cost open-source alternatives as they seek to curb deployment and inference expenses.

Anthropic and OpenAI are competing simultaneously for consumer access and a place in enterprise workflows. Anthropic estimates that model APIs accounted for 86% of its roughly $4.5 billion in revenue in 2025, while the Claude chatbot contributed about $600 million. The figures show that individual usage habits are becoming an important foundation for enterprise adoption and revenue.

An Indagari analysis released on March 28, covering about 28 million U.S. consumers, showed that the number of paying Claude users grew to roughly 2.5 times its previous level between September 2025 and February 2026. Anthropic also confirmed that its subscriber base has more than doubled since the start of 2026. OpenAI remains ahead with 910 million weekly active ChatGPT users and projects that it will have 220 million paying subscribers by 2030.

Competition between OpenAI and Anthropic has expanded beyond model performance into enterprise adoption and business models. Anthropic signed a multiyear, $200 million partnership with Snowflake in December 2025, while OpenAI is leveraging ChatGPT's vast user base to expand in the enterprise market. Control of corporate workflows and advertising gateways could shape generative AI revenue and determine which platforms dominate.

At Super Bowl LX on February 8, 2026, Anthropic ran two advertisements mocking OpenAI's introduction of ads to ChatGPT, escalating their battle of brands. On March 9, a federal court in San Francisco granted Amazon a preliminary injunction barring Perplexity's Comet agent from logging into Amazon to shop on users' behalf. The court found that Amazon had presented compelling evidence of unauthorized access.