Google has stepped into a deal that changes the shape of the AI coding race. After OpenAI spent months trying to acquire Windsurf, Google is bringing in Windsurf CEO Varun Mohan and much of the company's research and development team for work inside Google Deepmind.
The arrangement leaves Windsurf as an independent company. It also shows how major AI companies are competing for people, products, and intellectual property without always buying start-ups outright.
What Google is getting from Windsurf
Windsurf said former Head of Business Jeff Wang will become interim CEO as Varun Mohan and key team members move to Google Deepmind. Their new focus will be so-called "agentic coding" projects, a phrase that points to AI systems built to take more active roles in software development workflows.
According to The Information, Google is paying around $2.4 billion for the deal. That amount covers multi-year compensation for the employees joining Google and a non-exclusive license to Windsurf's intellectual property.
The structure matters. Google is not taking ownership of Windsurf. Instead, it gains talent and access to technology while Windsurf continues to operate on its own.
Windsurf, previously known as Codeium, builds AI tools that let users generate code through natural language. The company also plans to increase its emphasis on the enterprise market, meaning its independent path is still tied to customers that need AI coding tools for business use.
Why this is not a traditional acquisition
The Google-Windsurf arrangement fits a wider pattern in the AI industry. Large technology companies are increasingly using licensing agreements and talent-focused deals instead of full takeovers.
That approach can offer several advantages. A company can bring in experienced teams, gain access to important technology, and avoid some of the friction that can come with buying an entire start-up.
The source article also notes that these arrangements are seen as a way to sidestep regulatory scrutiny often attached to outright acquisitions. Major tech firms are already receiving closer attention from antitrust authorities, which makes deal structure more important than it used to be.
Google has used a related playbook before. Last year, it hired Character AI founders Noam Shazeer and Daniel De Freitas and paid $2.7 billion to the start-up as part of that deal.
Other large companies have made similar moves:
- Microsoft invested about $650 million to bring in Deepmind co-founder Mustafa Suleyman as "CEO of AI" through Inflection.
- Meta bought a 49% stake in Scale AI to bring on CEO Alexandr Wang and key staff.
- Meta has also recruited more than ten leading researchers from OpenAI recently.
- Amazon acquired a non-exclusive license for Covariant's robotics foundation models in summer 2024 and hired roughly a quarter of Covariant's workforce.
- Amazon also launched an "AGI Lab" in San Francisco, initially staffed by employees from Adept.
The Amazon transactions reportedly cost about $1 billion each. Together, these examples show that the market for AI talent and technology is being shaped by deals that do not always look like classic mergers.
How OpenAI lost the Windsurf deal
OpenAI had been discussing a full acquisition of Windsurf for months. The reported offer was around $3 billion, and the deal would have been OpenAI's largest to date.
The strategic reason was clear from the source article: OpenAI wanted to regain ground in the fast-growing market for AI-powered coding tools. Microsoft's GitHub Copilot and Anthropic's Claude Code are currently leading the space.
But according to The Information, the talks broke down because of concerns tied to OpenAI's existing agreement with Microsoft. That agreement requires OpenAI to share its technology with Microsoft. It also gives Microsoft exclusive rights to host OpenAI's models on Azure until at least 2030.
OpenAI was unable to secure an exception for Windsurf. That made the acquisition harder to complete, because Windsurf's technology could have become entangled with OpenAI's obligations to Microsoft.
The result is a sharp reversal. OpenAI was positioned as the likely buyer, but Google ended up hiring the CEO and major technical talent while also securing access to Windsurf intellectual property through a non-exclusive license.
What it means for the AI coding market
AI coding tools are becoming a major competitive front for the largest AI companies. The Windsurf episode shows that the race is not only about model performance. It is also about distribution, enterprise focus, engineering teams, and who controls access to key systems.
Windsurf's independence keeps the company in the market even as important leaders move to Google. Jeff Wang's role as interim CEO gives the start-up a continuity plan, while its enterprise focus suggests it will still try to compete for business customers.
The deal also leaves questions around model access. After details of the possible OpenAI-Windsurf transaction became public, Anthropic limited Windsurf's access to its Claude models to prevent its technology from ending up with OpenAI.
It is unclear how that situation will change now that Google has brought in parts of the Windsurf team. Google is an investor in Anthropic, and the two companies collaborate in some areas. At the same time, they directly compete in the AI coding market.
That tension captures the broader state of AI: companies can be partners, investors, customers, and competitors at once. In AI coding, those overlapping relationships now shape who gets talent, who gets technology, and which tools developers and enterprises may use next.