Companies to watch (AI's second-order winners): None of these are AI labs. All six stem from the AI buildout’s ripple effects: power, materials, security, and law scrambling to keep up. All were funded in a two-week window in early-to-mid July. 1. CuspAI - $450M Series B, $2.6B valuation (up from $520M last September) - Led by Kleiner Perkins and NEA, with Jeff Bezos's Bezos Expeditions joining - Builds AI that designs new materials for chips, clean energy, and manufacturing - launched the ai materials foundry coalition with 45+ orgs, including nvidia, meta, samsung, hyundai, and lam research - Watch it: the chip race needs new materials faster than labs can deliver, and the biggest players are already paying for early access 2. Neo - $100M seed and Series A - Backed by Andreessen Horowitz, Bessemer, Craft Ventures, and Merlin Ventures - Founded by ex-SentinelOne executives - Builds the control layer security teams need to police autonomous AI agents inside a company - watch it: "ai agents acting autonomously" is now a real, funded security category, not a hypothetical risk 3. Quantum Systems - $1.2B Series D (July 2), $8B valuation, Europe's largest defense-tech round - Munich-based drone maker; its Vector drone flew over 19,000 missions for Ukraine in 2025 - Funding goes toward expanding armed drone manufacturing across seven countries - watch it: defense capital is now moving at AI-startup speed and size 4. Proxima Fusion - €411M (~$468M), €2.4B (~$2.7B) valuation - Led by XTX Ventures and East X Ventures, with Google and RWE as strategic backers - Europe's best-funded fusion company, spun out of the Max Planck Institute - Targeting a net-energy demonstrator in the early 2030s - watch it: AI's power appetite just pulled Google into funding a fusion reactor directly 5. Quaise Energy - $134M Series B first close - Led by Prelude Ventures with JERA and Idemitsu, two of Japan's largest energy companies - Uses MIT-derived millimeter-wave drilling to reach superhot rock (300-500°C) for geothermal power - watch it: it's chasing the exact same problem as Proxima Fusion, AI's electricity demand, from the opposite technical direction 6. Norm AI - $120M Series C, $1.2B valuation - Led by Khosla Ventures - Builds "supervisory AI" that checks whether a company's other AI tools are following the rules - Clients cover $30 trillion in assets under management - Delaware's Secretary of State hired Norm AI to help define a new legal entity type specifically for AI agents - watch it: a US state government just outsourced part of "what is an AI agent, legally" to a startup Some interesting observations: • Combined, these six companies raised over $2.4 billion in roughly two weeks, and not one of them builds a foundation model. • Two of them (Proxima Fusion, Quaise Energy) are solving the same bottleneck, AI's power demand, from opposite ends: one fusion, one geothermal. • Norm AI getting hired directly by a state government is a bigger signal than its funding number. This is where the actual bottlenecks of the AI buildout, power, materials, security, and law, are quietly getting capitalized while ai labds make headlines.