naveen rao

The brain computes at ~20 watts using nonlinear dynamics and stochastic oscillatory coupling rather than matrix math; by building synthetic circuits that exploit the time axis of physics for computation — eliminating the memory-compute separation — we can approach Landauer limits and achieve three orders of magnitude better energy efficiency than current 2D lithography.

Current AI trajectory consumes gigawatts for inference and training; with humanity's total brain power at only 160 GW vs 9,000 GW global electrical capacity, the von Neumann compute paradigm's inefficiency will exhaust available energy within years, making fundamental substrate innovation an existential necessity for AI scaling.

Traditional semiconductor scaling via 2D lithography has reached a plateau where manufacturing and packaging improvements lower cost but no longer reduce energy per flop with memory access, necessitating a shift to non-von Neumann architectures like neuromorphic dynamical systems to continue efficiency gains.