Top AI researchers are departing from technology giants like Meta and Alphabet at an unprecedented rate to launch startups that are quickly securing massive funding rounds, signaling a new phase of accelerated talent movement in Silicon Valley's AI sector.
On April 28, CNBC reported that former Google DeepMind researcher David Silver announced a record-breaking $1.1 billion seed funding round for his startup, Ineffable Intelligence, which was established just months ago. Separately, former DeepMind employee Tim Rocktäschel is reportedly seeking up to $1 billion in funding for his new company, Recursive Superintelligence. Concurrently, AMI Labs, founded by former Meta AI lead Yann LeCun after his departure, completed a $1 billion funding round in March of this year.
Analysts indicate that investor enthusiasm is providing strong impetus for this wave of departures. According to data from Dealroom, venture capital has injected $18.8 billion into AI startups founded since early 2025, putting 2026 on track to potentially surpass the total $27.9 billion invested throughout the previous year.
Experts note that departing talent is not only taking technical expertise but also deep insights into industry blind spots, which forms the core logic behind investor bets.
The funding amounts secured by these new ventures are astonishing, far exceeding the conventional boundaries of early-stage investment. Ricursive Intelligence, co-founded by former Anthropic and Google DeepMind researchers Anna Goldie and Azalia Mirhoseini and focused on AI tools for chip design, was established last September and subsequently closed two funding rounds totaling $335 million in December and January. Periodic Labs, founded by former OpenAI and DeepMind staff to develop autonomous labs, secured $300 million last September, merely months after its inception. San Francisco-based Humans&, founded by former Anthropic and xAI employees in October, completed a $480 million funding round in January.
Elise Stern, a Managing Director at Eurazeo, which invested in AMI Labs, attributes the competitive advantage of these founders to their unique internal perspective: "They know what truly works at scale and they clearly see which opportunities are being deprioritized internally. The opportunity is right there."
Reports indicate that these emerging companies are not simply replicating the strategies of large firms but are instead pursuing clearly differentiated technological directions. Joël-Carbonell of HV Capital pointed out that a growing number of AI researchers are questioning whether continued scaling of current large language models is sufficient to overcome the next bottleneck in AI capabilities.
AMI Labs is focused on developing AI systems capable of learning from continuous real-world data. A company spokesperson stated, "AI has made significant progress in content generation, but clear shortcomings remain in fundamental cognition, causal reasoning, and reliable behavior in real-world environments. As AI moves from screens into physical settings like industry, robotics, and healthcare, these limitations become increasingly critical."
Ineffable Intelligence is concentrating on reinforcement learning—training AI models through experience rather than relying on human-labeled data—contrasting with the mainstream approach of training on internet text. A source familiar with the matter told CNBC that this is also the technical path adopted by Humans&.
Anna Goldie of Ricursive Intelligence emphasized the strategic value of an independent identity: "For chip manufacturers to trust us with their most core intellectual property, we must remain neutral, which was impossible to achieve within Google."
Notably, after securing ample funding, these startups are extending their reach back to the major tech companies, creating a secondary flow of talent. Goldie revealed that Ricursive Intelligence has reassembled the core AlphaChip team, "which involved recruiting some of our former colleagues." The current team's backgrounds span Alphabet, Anthropic, NVIDIA, Apple, and xAI. This pattern is common among several new companies—founders, leveraging their personal reputations and investor-backed capital, are attracting top researchers from former employers and other AI giants, intensifying the talent competition between startups and large firms.
Behind this exodus lies a "internal involution" within major companies, creating a window for entrepreneurship. The arms race among large AI labs is inadvertently generating opportunities for smaller, more agile firms. As HV Capital partner Alexander Joël-Carbonell noted, as major AI labs face pressure to justify sky-high valuations, commercial objectives are increasingly prioritized, significantly compressing the exploration space for top researchers. "Within large foundational model labs, the pressure to deliver benchmark performance and maintain rapid release cycles leaves almost no room for truly exploratory research, especially in directions outside the mainstream large language model paradigm." This structural contradiction is making top talent, who wish to pursue cutting-edge but non-mainstream research directions, increasingly inclined to strike out on their own.
Comments