Sam Altman Says Your Gaming PC's RAM Would Be Better Off in His Data Centers

Sam Altman has a suggestion for anyone frustrated by the soaring price of RAM this year: maybe you shouldn't have any in your computer at all.

In an interview with YouTuber Cheeky Pint, the OpenAI CEO was asked about the global shortage of compute and memory squeezing the AI industry. The conversation turned to DRAM, the memory chips used in everything from gaming PCs to game consoles, and Altman didn't hold back on where he thinks the market has gone wrong.

"From a global allocation perspective, actually, it's quite terrible to have everybody have a bunch of the world's limited supply of DRAM sitting in their own computer, and being used maybe 5% of the time," Altman said. "You want maximum utilization in the cloud if you're trying to deal with this very supply constrained environment. And somehow many people's intuition is the opposite of that."

His argument is essentially about efficiency. A chip sitting in a home PC is idle most of the day. The same chip pooled into a data center, Altman argues, could run constantly, serving AI workloads around the clock instead of waiting for someone to open a game or a browser tab. He went as far as to say the market's "invisible hand" is failing here, apparently because DRAM wafers keep ending up in consumer products instead of flowing directly into data centers built to train and run AI models.

It's a framing that makes a lot of sense if you're running a company whose entire business depends on having enough memory to make AI models faster and cheaper to operate. Altman has every incentive to want hyperscalers' data centers stocked with as much DRAM as possible, as fast as possible. But the logic runs into a problem: RAM isn't some niche AI component. It's a basic building block of nearly every consumer electronics product, and the current shortage is already hitting people who have nothing to do with running AI models locally.

The price evidence is hard to ignore. Used PS5 Pro consoles have reportedly climbed as high as $1,399 this month, following an official price increase back in March that Sony attributed to the RAM shortage. Nvidia's RTX 5090 graphics card has effectively been discontinued, and the RTX 5080 has now vanished from Nvidia's own marketplace as well. None of that is happening because gamers are secretly training large language models on their rigs. It's happening because memory supply is tight everywhere, and AI companies are absorbing a growing share of it.

That's the tension in Altman's comment. He frames the problem as people hoarding DRAM that would be better utilized in the cloud, but the squeeze consumers are feeling is a downstream effect of AI companies competing for the same limited manufacturing capacity. Samsung's latest earnings call reinforced that dynamic, indicating the memory crunch isn't expected to ease up next year, according to the company.

Altman's underlying belief, that cloud-based AI is where computing is headed and local hardware is increasingly beside the point, isn't new for him. But saying it out loud in the middle of a RAM shortage that's already made consoles and GPUs more expensive is a reminder of how differently the people building AI infrastructure and the people buying gaming hardware are experiencing this moment. One side sees underused memory as a global inefficiency to fix. The other just sees prices that keep climbing with no end in sight.