Memory shipped through 2027 could run 33–171 million concurrent frontier-model agents, or billions using efficient open models. Epoch AI estimates inference capacity from HBM supply, serving benchmarks, and agent-hour costs.
How many AI agents could run on the AI chips shipped through 2027?
About the authors
Jason LiJason Li is a researcher at Epoch AI, where he studies topics related to AI inference. Before Epoch, he worked at NVIDIA on LLM serving and inference optimization.

Newsletter
May 25, 2026
[](https://epoch.ai/gradient-updates/is-a-compute-crunch-coming)Is a compute crunch coming?
Related topics
ChipsEconomic impactFinances