The money is moving into artificial intelligence (AI) built around specific problems, and the amounts keep growing. The biggest check this week went to a research lab that has not shipped anything yet.
Ineffable Intelligence, a London-based startup founded in late 2025 by David Silver, former head of reinforcement learning at Google DeepMind, raised $1.1 billion in a seed round at a $5.1 billion valuation. Sequoia and Lightspeed co-led the round, with Nvidia, Google, Index Ventures and DST Global among the participants. The U.K. government’s Sovereign AI Fund and the British Business Bank co-invested, marking the fund’s second direct equity deal since its April launch. The round is the largest seed round in European history.
Silver’s pitch is a departure from large language models trained on internet text. Ineffable is building what it calls a superlearner, an AI system that generates knowledge by interacting with environments and learning from experience, applying reinforcement learning at a scale Silver believes can reach superintelligence. The company has no product, no revenue and no published roadmap. The check size reflects Silver’s standing in the field.
Agentic Workflows Find Their Verticals
Below the frontier research tier, a separate class of companies raised capital this week by applying AI to industries where the workflows are well-defined and the inefficiency is expensive.
Avoca closed a Series B at a $1 billion valuation, bringing total raised to more than $125 million. The company builds AI agents for home services businesses like HVAC, plumbing and electrical to handle inbound calls, scheduling and lead follow-up for contractors. Backers include Kleiner Perkins, Meritech, General Catalyst and Y Combinator. The company said it is on track to book $1 billion in jobs through its platform.
Sereact, a Stuttgart, Germany-based robotics AI company, raised $110 million in a Series B led by Headline to scale its Cortex 2.0 model and open its first U.S. office in Boston. Sereact’s AI operates picking robots in warehouses for customers including BMW, Mercedes-Benz and PepsiCo. More than 200 systems are live across Europe. The company said its robots have completed more than a billion real production picks, with one in roughly 53,000 requiring human intervention.
Orkes secured $60 million in a Series B led by AVP. The company, built by the original architects of Netflix’s microservices orchestration platform, helps developers run AI agents in production environments. Its open-source foundation, Conductor, is in use at Netflix, JPMorgan Chase, Atlassian and Tesla. Orkes has tripled its customer base since its Series A in 2024 and now counts United Wholesale Mortgage and Quest Diagnostics among its enterprise accounts.
Infrastructure Bets Follow the Agent Wave
As agentic AI moves into production, the supporting infrastructure is attracting its own capital.
Cloudsmith, an artifact management platform, announced a $72 million Series C led by TCV and Insight Partners one year after its Series B. The company helps engineering teams manage, secure and govern the software packages and dependencies that AI coding agents produce. As AI tools generate code, the software they ship creates a growing security surface that many enterprise teams are not equipped to manage. Cloudsmith positions itself as the governance layer for that output.