Sam Altman says it’s better for RAM to be in a data center than on your desk

Judgment

CIOs and developers planning local AI inference now face a sharper choice: keep buying RAM-heavy desktops for autonomy, or shift workloads to cloud, where Altman says DRAM is used five percent of the time—while OpenAI’s Stargate deals tighten supply.

Basis

On the Cheeky Pint podcast, OpenAI CEO Sam Altman called it “terrible” to have RAM in personal PCs instead of data centers. He said local-inference advocates assume open-source models will run on their computers and reduce data-center need, but from a global allocation view it is bad to have the world’s limited DRAM sitting in machines used five percent of the time; he wants maximum utilization in the cloud. The mechanism: Samsung and SK Hynix have long-term deals with cloud operators and projects like OpenAI’s Stargate that have reserved much of global RAM output for the next few years, and the billions spent on chips are why laptops, phones, and other tech are getting more expensive.