- Top AI researchers from Meta, Google, DeepMind, and OpenAI are departing to launch their own startups—raising hundreds of millions within months of incorporating.
- VCs poured $18.8 billion into AI startups founded since January 2025, on pace to eclipse last year’s $27.9 billion, per Dealroom.
- Founders say Big Tech’s fixation on scaling LLMs has left entire fields of research—new architectures, interpretability, vertical models—deprioritized and ripe for outsiders.
David Silver, the former DeepMind researcher behind AlphaGo and AlphaZero, announced a $1.1 billion seed round for Ineffable Intelligence on Monday—the largest AI seed deal in European history. He’s not alone. Tim Rocktäschel, another DeepMind alumnus, is reportedly raising up to $1 billion for Recursive Superintelligence. Yann LeCun departed his role as Meta’s AI chief and raised $1 billion for AMI Labs in March. The pattern is unmistakable: the people who built the foundations of modern AI are walking out the door and taking the checkbook with them.
At least half a dozen AI labs founded by former staff at OpenAI, DeepMind, Anthropic, and xAI have each raised nine-figure sums in the past year. Periodic Labs, started by ex-OpenAI and DeepMind researchers, collected $300 million in September. Ricursive Intelligence, founded by former Anthropic and DeepMind staff Anna Goldie and Azalia Mirhoseini, pulled in $335 million across two rounds. Humans&, launched by former Anthropic and xAI employees, raised $480 million in January. These are months-old companies with no revenue and no product—just pedigree.
“The race for AI dominance among the biggest AI labs has created an opening for smaller, more nimble companies,” Elise Stern, managing director at French VC Eurazeo, told CNBC. Eurazeo backed AMI Labs.
Why AI Researchers Are Walking Away From Big Tech
The explanation is partly structural. As the largest AI labs chase astronomical valuations—OpenAI at $300 billion, Anthropic at $350 billion after Google’s $40 billion commitment—they’ve narrowed their focus to commercial benchmarks and rapid release cycles. The research that doesn’t directly serve those goals gets shelved. “Inside the large foundational labs, the pressure to deliver benchmark performance and maintain rapid release cycles leaves limited room for genuinely exploratory research, particularly outside the dominant LLM paradigm,” Alexander Joël-Carbonell, partner at HV Capital, told CNBC. HV Capital also invested in AMI Labs.
Stern put it more bluntly: “When you’re in a race, you narrow focus. That creates a vacuum. Entire areas of research, like new architectures, agents, interpretability and vertical models, are being deprioritised, not because they don’t matter, but because they don’t win the immediate race.” The companies emerging from this vacuum aren’t trying to build the next GPT. They’re targeting the problems Big Tech has decided can wait—reinforcement learning, chip design automation, autonomous laboratories, real-world grounding for AI systems.
Ricursive Intelligence exemplifies the strategy. Goldie and Mirhoseini led Google DeepMind’s AlphaChip project, which automated chip design using AI. Their new company applies the same approach—but as an independent vendor rather than a Google subsidiary. “For chipmakers to trust us with their most valuable IP, we have to be Switzerland, and that wouldn’t be possible if we were at Google,” Goldie told CNBC. The company reassembled the core AlphaChip team, hiring from Google, Anthropic, Nvidia, Apple, and xAI.
The Numbers Behind the AI Talent Drain
The capital flowing into these founder-labs is staggering by historical standards. In 2026 alone, VCs have funneled $18.8 billion into AI startups founded since the start of 2025, according to Dealroom data cited by CNBC. That’s on track to surpass the $27.9 billion raised last year by companies launched since the start of 2024. Ineffable Intelligence’s $1.1 billion seed at a $5.1 billion valuation makes it the largest seed round in European history—more than many Series C rounds. And that was announced the same day as this broader trend story.
What makes these bets unusual is that the investors aren’t funding technology—most of these companies have nothing to ship yet. They’re funding people. Founders who worked at frontier labs have “unique” insight, Stern told CNBC. “They know what works at scale, and they know exactly what is being left on the table internally. That’s where the opportunity lies.” It’s a bet that institutional knowledge about what Big Tech is choosing not to pursue is worth more than any prototype.
The question nobody in the funding chain seems eager to answer is what happens when these startups need to show revenue. Axios reported that investors remain “bullish” on AI even after OpenAI reportedly missed its own revenue targets—suggesting the appetite for risk in this sector has yet to cool. But OpenAI at least has millions of paying users. These new labs have whiteboards and ex-employer NDA violations waiting to happen.
Google, Meta, Anthropic, and OpenAI did not respond to CNBC’s request for comment on the departures.