As AI agents grow capable, more compute shifts from GPUs to CPUs
AccBalanced · x · 2026-09-11
A retweeted thread argues that as AI agents become more capable, more work happens around the model itself. A basic chatbot just takes a prompt, runs the model, and returns an answer, while an agent decomposes one instruction into dozens of subtasks — searching, running Python, calling APIs, querying databases, self-checking, managing sandboxes, and hitting the model repeatedly.
This means a growing share of agent computing is ordinary computing: GPUs still handle the parallel math inside the model, but CPUs and surrounding systems must coordinate everything before, between, and after model calls. Nvidia is designing hardware around exactly this — Vera is explicitly positioned as a CPU for agentic AI.
More from coding & agent
- 'Model doesn't matter': is opinionated model routing in coding tools obsolete? — hugobowne · 2026-09-11
- GitHub's agentic workflows engine now runs your custom Pydantic AI agents to maintain repos — samuelcolvin · 2026-09-11
- Why chat assistants fail at delegation, and a security model to fix it — uriwa · 2026-09-11
- Claude Code vulnerability lets untrusted git repo escape macOS sandbox for RCE — ziv_ravid · 2026-09-11
- Practitioner trio on agents: evals as customer outcomes, fewer guardrails, vertical fine-tunes — illscience · 2026-09-11
- Handing off AI support agents to humans without making customers repeat themselves — Wonderful_Ebb5860 · 2026-09-11