Anthropic revenue surpasses OpenAI

Multiple expert sources from early 2026 indicate that Anthropic's revenue has surpassed that of OpenAI, marking a significant shift in the AI market landscape. By April 2026, Anthropic had reportedly reached a $30 billion revenue run rate, a figure explicitly stated to be ahead of OpenAI [2, 6, 22, 26]. This development appears to be the culmination of a trend first noted in August 2025, when SemiAnalysis CEO Dylan Patel asserted that Anthropic had already overtaken OpenAI specifically in API revenue [1, 7, 12]. By March 2025, other experts noted that Anthropic's enterprise revenue was close to catching up with OpenAI's , suggesting a rapid and successful enterprise sales motion that has since accelerated dramatically. This contrasts with OpenAI's reported internal turmoil and strategic shifts, such as shutting down its Sora video model due to a poor cost-to-revenue ratio [8, 17].

Anthropic's revenue growth has been described as exponential, with several sources corroborating a trajectory of 10x year-over-year increases. The company reportedly grew from zero to $100 million in 2023, surpassed $1 billion in 2024, and reached approximately $10 billion in 2025 [24, 30]. Another account details ARR growth from $1 billion at the start of 2025 to over $19 billion within 14 months . The leap to a $30 billion run rate in early 2026 was fueled by staggering monthly performance, with one source claiming Anthropic generated $6 billion in revenue in February 2026 alone , a figure that surpassed the annual revenue of both Databricks and Snowflake . This rapid escalation supports claims that Anthropic's growth rate is faster than OpenAI's . It is important to note the methodology behind these figures; one source specifies that both companies calculate Annual Recurring Revenue (ARR) by annualizing the average revenue of the last four weeks, a method that can amplify recent growth spikes .

The competitive dynamic between Anthropic and OpenAI is occurring within a market expanding at an unprecedented rate. The two companies are collectively adding revenue on a scale comparable to entire established industries. In the last year, they added as much revenue as half of the public cloud market, excluding the Magnificent Seven [4, 10, 14, 18, 20]. Projections for 2025 suggested the two firms would add nearly half as much new run-rate revenue ($23 billion) as the entire public software industry combined ($46 billion) [9, 16, 19]. This duopolistic expansion appears to be primarily constrained by infrastructure rather than demand. Groq Founder Jonathan Ross posits that if either company were given double its current inference compute capacity, its revenues would almost double within a single month , highlighting that access to compute is a critical determinant of market leadership and revenue generation in the current AI landscape.