OpenAI’s annual recurring revenue is approaching $70 billion, powered by a sharp acceleration in enterprise sales that is narrowing the company’s gap with Anthropic.
According to Axios, OpenAI’s revenue run rate has increased more than 70% since the start of the third quarter, while enterprise sales have more than doubled since July. The jump gives the ChatGPT maker fresh momentum in a market where Anthropic had recently pulled ahead on reported revenue growth.
The shift is especially important because enterprise customers are becoming one of the central battlegrounds in the AI race.
OpenAI’s Enterprise Business Is Accelerating
OpenAI has increasingly pushed beyond consumer ChatGPT subscriptions into corporate AI spending through enterprise products, APIs and developer tools.
That strategy appears to be working.
Only weeks ago, Anthropic’s revenue run rate had reached roughly $65 billion, putting it ahead of OpenAI and reinforcing Claude’s strength among business customers.
Now OpenAI is closing that gap quickly.
The change also comes as price competition across the AI industry intensifies. Falling token costs have already pushed OpenAI and Anthropic into a broader AI pricing war, forcing both companies to grow usage while keeping infrastructure costs under control.
| Quarter | OpenAI Revenue | Anthropic Revenue |
|---|---|---|
| Q3 2024 | ~$1.6B* | ~$0.1B* |
| Q4 2024 | ~$1.8B* | ~$0.2B* |
| Q1 2025 | ~$2.7B* | ~$0.5B* |
| Q2 2025 | ~$3.5B* | ~$1.0B* |
| Q3 2025 | ~$4.0B* | ~$1.8B* |
| Q4 2025 | ~$2.8B* | ~$2.3B* |
| Q1 2026 | $5.7B | ~$4.7-4.8B |
| Q2 2026 | ~$6.7B | $11.5B+ |
IPO Competition Adds More Pressure
The revenue race is becoming more important as both companies prepare for potential public listings.
OpenAI has already begun moving toward an IPO filing, while Anthropic filed confidentially earlier this year and has been linked to valuations potentially approaching $2 trillion.
Public investors would get much deeper access to both companies’ finances, including a metric that remains less visible today: costs.
That matters because enormous revenue growth does not automatically translate into profits.
Training frontier models, running inference and securing enough computing capacity require billions of dollars in annual spending. Anthropic’s latest disclosures have already shown how large those commitments can become, while OpenAI has provided far less detail about its current expense base.