gary tan

Gary Tan cites SemiAnalysis: 150GW load requests in Texas vs 60GW capacity, only 1.5GW approved; Elon using generators, Boom Supersonic pivoting to engines; Madrron's cooling saves 30% of grid permit ($500M/GW for chillers); power/cooling is the primary constraint for AI buildout.

AI makes mechanical/electrical engineering, supply chain management, and vendor discovery dramatically easier; Initialized Capital pivoted early; YC seeing more defense/hard tech companies this batch than all last year; investor sentiment shifted from 'hardware is hard' to 'hard tech is the way' post-Claude Code.

Gary Tan reports hard tech companies now comprise 25% of YC batch (40x increase), as AI tools like GPT-6 Astra enable small teams to solve complex engineering problems previously requiring massive capital and specialized hires.

Gary Tan predicts inference, data center, semiconductor, GPU and CPU demand will increase 1000x, creating massive opportunity in AI infrastructure that he believes is not currently priced in.

Gary Tan argues the AI safety discourse should shift from policy debates to building concrete cybersecurity defenses—provenance tracking, shutdown strategies, and prompt injection prevention—to stop agent swarms from seizing infrastructure.

Gary Tan predicts a massive turnover in software as AI-native startups with ~20 engineers using agentic systems will out-execute legacy SaaS companies employing thousands of engineers using old methods.

Gary Tan showcases nuclear reactor at YC demo day (no uranium); Har Tagar notes team building nuclear reactor inspired by Elon/Palmer/Anduril; investor appetite for hard tech now includes nuclear; structural shift from software-only funding.