Margin of Safety #22: The Reverse Acquihire

A new development in the AI landscape brings change to the way M&A has been done in tech

Summary of Events

Last week’s failed acquisition of Windsurf by OpenAI furthers a trend in how strategic M&A plays out in the AI space. Windsurf’s original $3 billion exit collapsed after Microsoft declined to restrict its access to Windsurf’s IP. For Windsurf (and potential owner OpenAI), which directly competes with Microsoft’s Copilot, the lack of exclusivity made the deal untenable.

Days later, Google stepped in—agreeing to license Windsurf’s IP for $2.4 billion and hiring CEO Varun Mohan along with 40 or so key staff members. The industry is now calling this a “reverse acquihire”: the acquiring party gets talent and rights to IP, but without taking on the company itself.

Then came the twist. Cognition, maker of the AI coding agent Devin, announced its own acquisition of Windsurf. The terms remain undisclosed, but Windsurf reportedly had $82M in ARR, 350+ customers, and over 200 employees. Unlike Google’s carve-out, Cognition is taking the whole entity and promising “no employee left behind”. From a strategic perspective, the combination makes sense. Windsurf built an agentic IDE; Devin is an agentic coder. Together, they start to resemble a full-stack AI development platform.

Our Opinions & Takes

• But this was not a traditional success story. Windsurf, once considered a breakout company, ended up in a multi-party carve-up more reminiscent of a distressed asset sale. If you really believe that windsurf was going to be the best IDE in five years, it’s easy to believe in valuations greater than $10B, even $50B. Investors and founders chose a different path. The company may have hit a ceiling on its own—limited by compute access, distribution, or data—and sought scale under a larger platform. Worse, the company may have concluded structural factors prevent another breakout star in the space.
• Cognition likely benefits meaningfully. It acquires a fully formed engineering org and positions itself as a white knight. Windsurf employees, previously uncertain about their future, now get continuity, equity upside, and a clear mission. For Cognition, this is a shortcut past the expensive, slow process of building world-class infra talent from scratch. Cognition also buys $82M ARR worth of customers for a low price.
• On the other hand, you have to wonder: the market has seen relatively few commercially successful IDE/developer environments over the years. Cognition benefits, but do they benefit enough to achieve breakout when Windsurf did not? The space in general is clearly in the crosshairs of the major labs.
• Still, this episode exposes an uncomfortable truth: the safety net for startup employees is fraying. Once, failed startups would find soft landings—acquired by Big Tech, with founders and employees rewarded, and VCs recovering their 1.0x liquidation preferences. That model is breaking. Today, acquiers can cherry-pick key employees and license IP, bypassing the cap table and leaving many behind.
• For employees evaluating fast-growing startups, founder charisma and fundraising headlines are no longer enough. Real IP, defensibility, and ecosystem leverage matter more than ever. This is generally difficult to do in the early stages (pre Series A), but especially at later stages I’d recommend a broader assessment.
• In this environment, Cognition may have pulled off a savvy deal as they are getting a fully built out turnkey engineering & sales team that built a legitimate product with ~$82M in ARR. But it’s not clear there will be another Cognition at the next reverse acquihire

Reach out to us if you are building in this space. We have some thoughts!
Kathryn Shih – kshih@forgepointcap.com
Jimmy Park – jpark@forgepointcap.com

This blog is also published on Margin of Safety, Jimmy and Kathryn’s Substack, as they research the practical sides of security + AI so you don’t have to.