Google hires Windsurf's CEO in a $2.4 billion deal as OpenAI's acquisition falls apart

A startup growing quickly

Windsurf was an AI coding startup, one of the companies building tools that use AI models to help programmers write and edit software. Its annual recurring revenue, the yearly value of its subscription business, had reached about $100 million in April 2025, up from about $40 million a few months earlier ^[1]. That growth made it a target for acquisition, and OpenAI had a deal in place to buy the company for $3 billion ^[1].

The deal falls apart

On Friday 11 July 2025, the $3 billion acquisition fell apart, TechCrunch reported, citing The Verge ^[1]. The same day, Google DeepMind hired Windsurf's CEO Varun Mohan, co-founder Douglas Chen and some of the startup's top researchers ^[1]. Bloomberg reported that Google was paying $2.4 billion to license Windsurf's technology and to hire its top employees ^[1] — a sum in roughly the same range as the price OpenAI had agreed to pay for the whole company.

What Google did and did not get

The structure of the arrangement mattered as much as the money. Google took no stake in Windsurf and gained no control over the company ^[1]. What it received was a nonexclusive license to certain Windsurf technology, which left Windsurf free to license the same technology to others ^[1]. That is a different thing from owning a company: Google got access to the technology and the people it wanted, while Windsurf remained a separate business that could still sell or license its work to other buyers.

What was left at Windsurf

Windsurf did not shut down. Its head of business, Jeff Wang, became interim CEO, and most of the roughly 250-person team stayed on to keep serving enterprise customers ^[1]. So the company continued with a large majority of its staff, a new leader and its existing corporate customers, even as its founding CEO and co-founder left for Google DeepMind.

Why the OpenAI deal collapsed

TechCrunch reported that the collapse of the OpenAI deal had been a major tension point in OpenAI's contract renegotiations with Microsoft, since OpenAI did not want Microsoft to get access to Windsurf's technology ^[1]. The exclusivity period on OpenAI's offer had also expired that Friday ^[1]. Both details point to the same difficulty: an acquisition of a fast-growing AI coding company touches the agreements between the companies involved, not just the buyer and the seller.

A recognizable pattern

TechCrunch described the arrangement as a reverse-acquihire, a pattern it compared to Google's earlier Character.AI deal with Noam Shazeer and Microsoft's hiring of Mustafa Suleyman ^[1]. In a conventional acquisition the buyer takes the company, its contracts and its liabilities along with its staff. In this pattern the buyer takes the senior people and a license to the technology, and the original company stays independent, keeping a smaller team, its customers and its ability to license its technology elsewhere. For readers following the competition between AI labs, the episode shows how the most valuable assets of a young AI company — its researchers and its software — can change hands without the company itself being sold.

What the numbers show

The revenue figures explain why two large AI companies were interested at the same time. Going from about $40 million to about $100 million in annual recurring revenue within months is fast growth for any software business ^[1], and it happened in a product category — AI-assisted coding — where the large model developers were themselves competing. OpenAI's offer was for the company; Google's $2.4 billion was for a license and a hiring package ^[1]. The outcome left Windsurf operating under an interim CEO with most of its roughly 250 employees ^[1], and left OpenAI without the company it had agreed to buy.