Meta, Google staff leave for startups

Meta, Google and OpenAI are watching senior artificial intelligence researchers leave to build their own labs, with investors funding the breakaways almost immediately. (cnbc.com) CNBC reported Tuesday that former Google DeepMind researcher David Silver disclosed a $1.1 billion seed round for Ineffable Intelligence on April 27, one day before the broader talent-migration story. The new company emerged at a $5.1 billion valuation. (cnbc.com 1) (cnbc.com 2) CNBC also said former DeepMind researcher Tim Rocktäschel is raising as much as $1 billion for Recursive Superintelligence. In March, AMI Labs announced a roughly $1 billion raise after founder Yann LeCun said he was leaving Meta’s top artificial intelligence role. (cnbc.com) (techcrunch.com) These startups are pitching alternatives to today’s dominant large language models, including new model architectures, agents, interpretability tools and “world models” trained on continuous real-world data. Eurazeo managing director Elise Stern told CNBC big labs are narrowing their research focus during the race for artificial intelligence dominance. (cnbc.com) (techcrunch.com) The money behind that shift is rising fast. Dealroom data cited by CNBC shows venture capital firms have put $18.8 billion into artificial intelligence startups founded since the start of 2025, putting 2026 on pace to top last year’s $27.9 billion for companies launched since the start of 2024. (cnbc.com) At the same time, Meta is cutting 10% of its workforce, about 8,000 employees, beginning May 20, and scrapping plans to hire for 6,000 open roles, according to a company memo reported by CNBC. Microsoft said last week it would offer voluntary buyouts to some U.S. employees, with about 7% of its U.S. workforce eligible. (cnbc.com) CNBC reported April 24 that the Meta layoffs and Microsoft buyouts together raised concern about more than 20,000 jobs as companies redirect spending toward artificial intelligence infrastructure and automation. Meta has also shifted content moderation work away from third-party vendors and contractors toward artificial intelligence systems. (cnbc.com 1) (cnbc.com 2) Investors backing the new labs are also giving them money to recruit from the same companies their founders left. That leaves Big Tech concentrating spending on a smaller set of artificial intelligence priorities while former insiders try to build the next ones outside. (cnbc.com)