Sourcery

$4.5B Brain-Inspired AI Chip Company

61 snips
Feb 7, 2026
Konstantine Buhler, Sequoia partner who backs deep-tech pioneers, and Naveen Rao, serial AI founder building biology-inspired, energy-efficient chips. They discuss why AI’s real bottleneck is power, recreating neural dynamics in silicon, the return of analog-inspired compute, and what it takes to found and fund capital-intensive deep tech startups.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

Energy, Not Capital, Is The Real Bottleneck

  • AI compute is colliding with global energy supply, making energy the true bottleneck, not capital.
  • New computation paradigms beyond von Neumann and digital abstractions are required to achieve biology-scale efficiency.
INSIGHT

Think Of Neural Nets As Dynamical Physics

  • Neural networks act like dynamical systems rather than precise arithmetic machines, suggesting different hardware could be far more efficient.
  • Reclaiming physical device dynamics (time and nonlinearity) can yield step-function efficiency gains.
ADVICE

What Investors Seek In Deep Tech

  • Investors look for a unique, compelling insight plus an outlier founder when funding formation-stage deep tech.
  • If you can aggregate world-class talent and a clear physics-driven thesis, large early rounds are justifiable.
Get the Snipd Podcast app to discover more snips from this episode
Get the app