The Biggest Seed Round in History Just Went to a Reinforcement Learning Lab, and Your Agent's Negotiation Skills Depend on Which Model It Runs

David Silver, former head of reinforcement learning at Google DeepMind, closed a $1.1 billion seed round for Ineffable Intelligence on Monday, valuing the months-old London startup at $5.1 billion. The round, led by Sequoia and Lightspeed with backing from Nvidia, Google, and the British government, is the largest seed financing in European history. Silver's bet: a "superlearner" trained through reinforcement learning that discovers knowledge without human data, sidestepping the LLM paradigm entirely. (CNBC | Reuters | TechCrunch)

Silver is not an outlier. CNBC reports that VCs have funneled $18.8 billion into AI startups founded since the start of 2025, on pace to surpass last year's $27.9 billion. Cognition (the team behind Devin) is in funding talks at a $25 billion valuation, up from $10.2 billion. Sierra, Bret Taylor's agent startup, just made its third acquisition in weeks. The pattern is clear: the pressure inside frontier labs to ship benchmarks and justify valuations is pushing the most ambitious researchers toward their own ventures, where they can pursue architectures that the incumbents will not prioritize.

What this means if you are building: The talent and capital flowing into non-LLM approaches (reinforcement learning, self-play, neuro-symbolic) will produce alternative agent architectures within 12 to 18 months. Do not bet your stack on a single paradigm.