dwarkesh

AI keeps demand for compute insatiable, so the share of the economy spent on compute rises and GPU rents stay high.

China's economy runs on sea lanes that close in any conflict, a permanent geographic discount on its assets.

A Second Cold War redirects capital from trade efficiency to rearmament, and defense contractors' order books compound.

If AI compute demand never satiates, GPU pricing power persists and NVIDIA's margins and revenue keep compounding.

Privatized AI value is now coming public; Anthropic's S-1 starts the re-rating of the biggest private AI name.

If intelligence falls out of raw scale, the compute owner compounds while efficiency tricks get arbitraged away.

RL's bits-per-flop problem pushes spend off the training update onto rollout/inference compute, where CoreWeave bills.

HBM bandwidth is the hard ceiling on AI scaling, and the dominant supplier owns the scarce asset and the pricing.

Rack-scale NVLink fabric (8 to 72 to 576 GPUs) is the real moat; rivals can't drop-in match it, so NVIDIA's premium holds.

Nvidia's moat is durable: electron-to-token transformation is too hard to commoditize even as software cheapens

Agents make tool-maker software companies skyrocket as tool instances explode, not get commoditized

Micron is a key memory winner as HBM and LPDDR go from specialty to mainstream computing

Silicon photonics suppliers Lumentum and Coherent benefit as Nvidia reshapes the optical supply chain

AlphaFold's value came from decades of wet-lab structural data, so the instrument makers that generate it capture the upside.