Meta, Google and OpenAI lose top AI talent as staff quit to launch startups

Leading researchers are departing from prominent tech corporations such as Meta and Google to create new startups, securing significant funding during this transition, as investors are strongly drawn to the promising business opportunities of emerging AI labs.

With tremendous investment flowing into AI, several of these nascent startups are securing substantial funding merely months after being established.

Recently, ex-Google DeepMind scientist David Silver revealed he had acquired a historic $1.1 billion seed round for his young venture, Ineffable Intelligence.

Tim Rocktäschel, another alumnus of DeepMind, is allegedly raising as much as $1 billion for his new firm, Recursive Superintelligence.

Earlier this year, AMI Labs announced securing $1 billion following its founder, Yann LeCun, stepping down from his position as AI chief at Meta.

Their mission is to develop AI systems capable of learning from ongoing real-world data.

In recent months, former employees from OpenAI, DeepMind, Anthropic, and xAI have also attracted hundreds of millions in venture capital for freshly launched businesses, including AI labs such as Periodic Labs, Ricursive Intelligence, and Humans&.

These ventures often recruit extensively from their founders' past workplaces and other AI industry leaders, as investors provide the financial resources necessary to lure top experts from major tech firms.

The fierce competition for AI leadership among large AI labs has created opportunities for smaller, more agile companies, shared Elise Stern, managing director at French VC Eurazeo, a supporter of AMI Labs, in an interview with CNBC.

"Being in a race means narrowing your concentration," she mentioned. "This opens up areas previously not focused on, such as new architectures, agents, interpretability, and specialized models, not because they lack importance, but because they aren't immediately essential for winning the race."

Investors are eagerly funneling funds into AI labs started by top researchers who previously worked at prominent tech companies.

In 2026, venture capitalists have invested $18.8 billion in AI startups initiated since the beginning of 2025, reports Dealroom.

This is set to exceed the $27.9 billion collected last year by startups launched since early 2024.

Founders experienced in leading labs have "a unique perspective," according to Eurazeo's Stern.

"They are aware of what succeeds on a large scale and understand what innovative opportunities aren't being pursued internally," she explained. "This is where they see potential."

As major AI labs focus more sharply on commercial outcomes to validate their massive valuations, it limits the creative freedom of top experts, explained Alexander Joël-Carbonell, a partner at HV Capital, also an investor in AMI Labs, to CNBC.

Ricursive Intelligence, which secured $335 million through two funding rounds in December and January after being established in September, focuses on creating AI tools to aid in designing chips.

Founders Anna Goldie and Azalia Mirhoseini both previously worked for Anthropic and Google DeepMind, contributing to the AlphaChip project which aimed at automating chip design.

The venture has also attracted former colleagues. "We've reunited the core AlphaChip team, which meant hiring some of our previous collaborators," Goldie noted. Additional team members have backgrounds with Google, Anthropic, Nvidia, Apple, and xAI, she said.