Windsurf — the three-way acquisition battle
The data
Windsurf (Cognition AI):
- $82M ARR reached within months of launch.
- OpenAI bid $3B to acquire the company.
- Google executed a $2.4B reverse-acquihire of the human capital (Varun Mohan, Douglas Chen, and core team).
- Cognition AI simultaneously acquired the brand, codebase, enterprise contracts, and developer behavior data.
A three-way bidding contest for the same asset, with each party valuing a different slice (talent / brand+data / strategic neutralization).
Source
Reporting synthesized in where-value-lands-2026, drawn primarily from Varun Mohan’s appearance on Lenny’s Podcast and trade-press coverage of the acquisition in 2025.
What it shows
The most legible market-price reading on what the distribution layer is worth. Three different acquirers, each with a different theory of the case, all converging on multi-billion-dollar valuations for a months-old company at the IDE / developer-distribution layer. The fact that none of the bidders were trying to acquire foundation-model capability (they have their own) is the load-bearing observation: the asset that was contested was the human-developer relationship and the behavioral data it produces.
The decomposition of the acquisition (talent → Google, brand+assets+data → Cognition) is itself an analytic artifact: it reveals which slices the market prices independently.
How to cite
Windsurf reached 2.4B reverse-acquihire of the founding team) and Cognition AI (which absorbed the brand, codebase, contracts, and developer data) after OpenAI’s $3B bid (2025).
Connections
- middle-layer-defensibility — keystone evidence alongside Cursor.
- distribution-moat — the asset class the bidders were paying for.
- H2_u-curve-of-value — same partial-validation pattern as Cursor.
- cursor-50b-valuation — adjacent data point.
- where-value-lands-2026 — primary source.