The acqui-hire era has arrived
Since mid-2025, three deals tell the story: Meta invested $14B in Scale AI and hired CEO Alexandr Wang. Google spent $2.4B licensing Windsurf's tech and folding its team into DeepMind. Nvidia wagered $20B on Groq's inference technology and brought the CEO onboard.
These aren't traditional acquisitions. They're talent grabs dressed as licensing deals, a workaround for antitrust scrutiny and a shortcut in the AI talent wars.
Most recently, Meta acquired Manus for $2.5B with a $500M retention pool, navigating the autonomous agents startup's pivot from Chinese roots to Singapore. The pattern is clear: Big Tech pays premium prices to hire teams it can't recruit directly.
Why loyalty died
The four-year vest used to be sacred. Now researchers leave doctoral programs mid-PhD, founders sell after two years, and OpenAI rehires people who left 18 months earlier for Mira Murati's Thinking Machines.
Money drives some of this. Meta reportedly offered top AI researchers packages worth hundreds of millions. But Sayash Kapoor at Princeton points to something deeper: people understand institutional limitations now. The Windsurf founders likely calculated they'd have more impact at Google with its compute resources.
The 2000s belief in company missions has given way to pragmatism. When AI innovation moves this fast, staying put has opportunity costs.
Investors adapt
Max Gazor at Striker Venture Partners says his team now vets founding teams "for chemistry and cohesion more than ever." Deal terms increasingly include protective provisions requiring board consent for IP licensing.
The irony: Scale AI was founded in 2016, when a $14B acqui-hire would have seemed absurd. Now investors model for it in early term sheets.
What this means
AI-related deals represented 50% of tech M&A in 2025, up from 25% in 2024. Hyperscalers are projected to spend $100B each on AI infrastructure by 2027. The talent war isn't slowing down.
Some see Nvidia's Groq deal as a fire-sale signal if AI confidence wanes. Others see it as smart antitrust navigation. Either way, the old startup playbook where founders stay until exit or failure is gone.
The question for enterprise buyers: when your AI vendor's team can be poached for $2B, how do you plan for continuity?