Top researchers are jumping ship from Big Tech firms meta And Google Startups are launching and raising massive amounts of funding in the process, as investors place big bets on the commercial potential of early-stage AI labs.
Amid huge spending on AI, many These new startups are raising hundreds of crores Within a few months of installation.
Yesterday, former Google DeepMind researcher David Silver announced that he had raised a record $1.1 billion seed round for his months-old startup Ineffable Intelligence. Another former DeepMind employee is Tim Rocktashell Allegedly Raising up to $1 billion for its new startup Recursive Superintelligence. Rocktashell did not respond to CNBC’s request for comment.
AMI Labs announced it was raising $1 billion in March, just months after its founder Yann LeCun said he was leaving his role as Meta’s AI chief. It is developing AI systems that can continuously learn from real-world data.
In the past year, former employees of OpenAI, DeepMind, Anthropic and xAI also raised hundreds of millions of rupees from investors for months-old ventures, including AI labs Periodic Labs, Recursive Intelligence and Humans End.
Many of these companies have hired extensively from the former employers of the founders themselves and from other AI veterans, as investors have supplied them with the money needed to lure top researchers to Big Tech.
Alice Stern, managing director of Eurazeo, a French VC that backs AMI Labs, told CNBC that the race for AI dominance among the largest AI labs has created a path for smaller, more nimble companies.
“When you’re in a race, you’re focused,” he said. “This creates a void. Entire areas of research like new architectures, agents, interpretability and vertical models are not being prioritized, not because they don’t matter, but because they don’t win the immediate race.”
striking out
Investors are rushing to pour money into AI labs founded by top researchers already working for leading tech companies.
According to Dealroom, in 2026, VCs have invested $18.8 billion in AI startups founded through the beginning of 2025. That’s on track to surpass the $27.9 billion raised last year by companies launched since the start of 2024.
Eurazeo’s Stern said founders who have worked in frontier labs have “unique” insight.
“They know what works at scale, and they know exactly what is being left on the table internally,” he said. “That’s where the opportunity lies.”
Alexander Joel-Carbonell, partner at HV Capital, which has also invested in AMI Labs, told CNBC that the focus increasingly on commercial goals — as major AI labs try to justify astronomical valuations — limits the freedom of top researchers.
“Inside large basic laboratories, the pressure to deliver benchmark performance and maintain rapid release cycles leaves limited space for truly exploratory research, especially outside the dominant LLM paradigm,” he said.
finding intervals
Recursive Intelligence, which raised $335 million in two rounds in December and January after being founded in September, is building AI tools to help with chip design. Founders Anna Goldy and Azalea Mirhosseini both previously worked for Anthropic as well as Google DeepMind, where they contributed to the Alphachip project, which focused on automating chip design.
Goldie told CNBC that potential customers are more likely to regard a new company as a neutral partner rather than a competitor.
“For chip makers to trust us with their most valuable IP, we have to be Switzerland, and that wouldn’t be possible if we were Google,” he said.
The company has also turned to former colleagues. “We got the core Alphachip team back and that included hiring some of our old colleagues,” Goldie said. Other team members previously worked at Google, Anthropic, NVIDIA, Apple And xAI, he added.
Periodic Labs, founded by former OpenAI and DeepMind staff, raised $300 million in September, just months after launching. It is considering developing autonomous laboratories.
HV Capital’s Joel-Carbonell told CNBC that a growing number of AI researchers are questioning whether leveraging existing large language model (LLM) approaches will be enough to reach the next level of AI capability.
AMI Labs, founded by former Meta AI head LaCun, said it is “the right time as AI has made huge strides in content creation, but still struggles with grounding, causality, and reliable behavior in real-world settings.”
Yann LeCun, chief AI scientist at Meta, speaks at the Viva Tech conference in Paris on June 13, 2023.
Chesnote | Getty Images News | getty images
“As AI moves beyond screens and into industry, robotics, health care and other physical environments, those limits become increasingly important,” a company spokesperson told CNBC.
Ineffable Intelligence will focus on reinforcement learning, which occurs when AI models learn from experience as opposed to human data — compared to many leading AI models that are trained on Internet text — a source familiar with the company told CNBC.
It’s an approach that is also being used by San Francisco-based Humans End, which was launched in October by former Anthropic and XAI employees and raised $480 million in January.
Google, Meta, Anthropic and OpenAI did not respond to requests for comment.
