Failed OpenAI Windsurf Acquisition
Codeium
The failed OpenAI deal showed that AI coding had become too strategic to stay a normal startup sale. By May 2025, OpenAI was willing to pay about $3B for Windsurf, a price that reflected more than current revenue. It reflected the value of owning a fast growing AI IDE, the developer usage flowing through it, and a direct distribution channel into coding, where Cursor and Windsurf were emerging as the leading independents.
-
Windsurf was not just another coding copilot. It had reached about $40M ARR by February 2025 and sat as the clear number two AI native IDE behind Cursor at about $200M ARR in March 2025. That made it one of the few scaled assets in a market where product, distribution, and model feedback loops were all compounding at once.
-
OpenAI’s interest was practical. Windsurf runs inside a VS Code style workflow that developers already use, so OpenAI could have dropped its models into a familiar editor, bundled it with ChatGPT plans, and gained both paying users and a large stream of real coding interactions to improve model behavior on software tasks.
-
What happened next makes clear why the acquisition mattered. After the OpenAI process ended, Google agreed on July 11, 2025 to pay about $2.4B to hire Windsurf’s CEO, co founder, and roughly 40 employees and license certain technology, while Cognition signed a separate deal on July 14, 2025 to buy the remaining business, including the product, brand, IP, and customer base.
Going forward, the winning AI coding companies are likely to be the ones that control both the model and the workflow where developers spend the day. The collapse of the OpenAI transaction and the later split outcome pushed the market toward tighter vertical stacks, where labs, IDEs, and coding agents are increasingly being combined rather than sold as separate layers.