TechCrunch Disrupt 2026 Day 2: AI Hardware, Robotics, and Venture Insights

San Francisco’s Moscone West hosted the second day of TechCrunch Disrupt 2026, drawing more than 10,000 founders, investors, and technology leaders. The expo floor featured over 300 startups, while the agenda centered on artificial intelligence, robotics, public‑safety technology, and venture‑capital perspectives.

Key Sessions

• AI, Surveillance, and the Future of Public Safety – Garrett Langley (Flock Safety) discussed how edge AI cameras are reshaping city‑scale monitoring.
• When AI Starts Designing Its Own Hardware – Anna Goldie and Azalia Mirhoseini (Ricursive Intelligence) presented research on automated chip‑floor‑planning and its impact on next‑generation GPUs.
• What We Believe Now – Benchmark’s full partnership (Jack Altman, Peter Fenton, Chetan Puttagunta, Everett Randle, Eric Vishria) shared investment theses for AI infrastructure and deep‑tech.

Why it matters for GPU / AI infrastructure: The sessions underscored a rapid shift toward custom silicon and accelerator‑rich clusters. As startups move from model training to real‑time inference at the edge, demand for high‑density GPU farms, low‑latency interconnects, and automated hardware‑design tooling is accelerating. Cloud providers and colocation operators must plan capacity and power envelopes that can support both massive training runs and distributed inference workloads.

Attendees left with concrete leads on emerging chip architectures, partnership opportunities with robotics firms, and a clearer picture of where venture capital is directing the next wave of AI‑hardware investment.