Menlo Ventures partner Deedy Das just updated his "Neolab" roster — a list of special AI startups that are worth over a billion dollars but have almost no revenue. The list grew from 50 to 63 companies, with a combined valuation of roughly $300 billion. Das defines Neolab as a class of AI startups that are valued at over $1 billion, have little to no revenue, are founded by top AI researchers, and are focused on long-term AI breakthroughs rather than short-term commercialization.
The founders on this list are heavily concentrated at top labs like OpenAI, DeepMind, and Meta AI. Notable names include SSI (Safe Superintelligence), founded by former OpenAI chief scientist Ilya Sutskever, valued at around $32 billion; Thinking Machines Lab, started by ex-OpenAI CTO Mira Murati, which is reportedly raising $4 billion to $5 billion at a valuation of up to $50 billion; and Reflection AI, co-founded by DeepMind veteran Ioannis Antonoglou and former DeepMind researcher Misha Laskin, which is raising $2.5 billion at around a $25 billion valuation with backing from Nvidia and a focus on open-source models.
https://twitter.com/deedydas/status/2052266037979320752
The list also shows that the AI startup hot spots are shifting from pure large language models to robotics, agents, world models, and AI biotech. Companies like Physical Intelligence, World Labs, and Nous Research are all on the list.
PitchBook data shows that 60 of the 95 new unicorns added globally in Q1 2026 were AI-related; AI companies now account for 47.6% of global unicorn valuations, or about $4.1 trillion. Meanwhile, Crunchbase reports that 26 of the 28 new unicorns added in April were AI companies.
