Amazon Adds 2 Million More Nvidia GPUs to AWS as Compute Demand Surges

Skynet Chance (+0.06%): A tripling of frontier compute capacity plus adoption of Nvidia's robotics and world-model stack for physical systems expands the scale and embodiment of AI deployment faster than corresponding alignment or oversight work. Larger concentrated compute pools raise the ceiling on capabilities that could outpace control mechanisms.

Skynet Date (-1 days): Demand exceeding expectations within five months and $279 billion in locked-in supply commitments signal compute scaling accelerating ahead of prior projections, pulling forward the timeline for the most capable systems. Extending the stack into warehouse robots also shortens the path from digital models to physical actuation.

AGI Progress (+0.04%): Two million next-generation GPUs plus integrated CPUs and networking materially raise the training and inference compute available to AI labs, the most reliable historical driver of capability gains. Huang's claim that AI is generating 'profitable tokens' indicates an economic flywheel funding further scaling.

AGI Date (-1 days): The speed of the escalation — from 1 million to 3 million GPUs in five months — plus secured memory and manufacturing capacity through fiscal 2028 removes a key bottleneck and accelerates the compute curve toward AGI-scale training runs. Broad Vera CPU deployment across hyperscalers and neoclouds compounds the effect industry-wide.