AI’s biggest week: compute, agents, and new models
AI just got a lot more expensive and a lot more capable. From giant GPU deals to bigger context windows and enterprise agent rollouts, here are the must-know moves shaping the latest AI wave.[1][2][4]
🧵 1/6
Anthropic’s compute sprint: it signed for SpaceX’s Colossus 1 supercomputer, with 220,000+ NVIDIA GPUs and 300MW of power, and doubled Claude Code rate limits for paid plans overnight.[1] Compute is now the moat.
🧵 2/6
Google’s Gemini 3.1 Ultra is built for 2M-token context and native text, image, audio, and video input, plus a sandboxed code execution tool.[1] That is a big step toward one model doing the whole workflow.
🧵 3/6
Enterprise AI is getting more operational: IBM just launched Forward Deployed Units, small senior teams paired with AI agents to handle coding, testing, docs, and governance in production.[4] The race is shifting from demos to delivery.
🧵 4/6
OpenAI is pushing deeper into companies with DeployCo, a new deployment unit backed by more than $4B from 19 partners; at the same time, Anthropic is expanding with Claude for Small Business and new partnerships.[2] AI sales is becoming field ops.
🧵 5/6
Which of these AI moves feels biggest to you: mega-compute, giant context windows, or AI delivery teams? Reply and repost with your take.[1][2][4]
🧵 6/6
Sign Up To Try Advanced Features
Get more accurate answers with Super Pandi, upload files, personalized discovery feed, save searches and contribute to the PandiPedia.