In short.

In short. The scheme is like this. We fund a startup — sell smart dudes to OpenAI. What could possibly go wrong.
1/
OpenAI buys Windsurf for 3 billion (https://www.reuters.com/business/openai-agrees-buy-windsurf-about-3-billion-bloomberg-news-reports-2025-05-06/)
(that’s when there are tons of prompts flying around and all that). We’re happy!
2/ But!
Three months later OpenAI decides they don’t need it anymore. (https://fortune.com/2025/07/11/the-exclusivity-on-openais-3-billion-acquisition-for-coding-startup-windsfurf-has-expired/)
3/ Investors write the money off, boys
go to Google. (https://www.theverge.com/openai/705999/google-windsurf-ceo-openai)

Ez.
👍
P.S. It’s kinda sweet how the fund got hooked up with API credits, since all these coding assistants are an omega-unprofitable business. Of course everyone will write that the investors didn’t get screwed. But usually the terms of deals like this are that the fund gets paid if the team hits KPI, and nobody ever hits them. Such is life.