OpenAI CEO Sam Altman Says You Having DRAM Instead Of Data Centers Is "Terrible"

Published Oct 30, 2026, 10:49 AM EDT James Lucas is an Executive Editor at TheGamer who has worked in professional journalism since 2018. A Newcastle University graduate with a degree in Journalism, Media, and Culture, they have written hundreds of guides, news stories, and original reports, with bylines in IGN, NME, GaymingMag, and VG247, sourcing interviews with studio heads, legendary industry figures, prominent actors, and more. An expert on FPS games, Soulslikes, RPGs, and survival horror, James has intimate knowledge of a wide variety of titles, providing in-depth coverage and exclusive profiles. Some of the games they have extensively covered include: Half-Life, Counter-Strike, Elden Ring, and Fallout. They pride themselves on their consistently thorough research, leveraging nearly a decade of industry experience. You can reach them at James.t@thegamer.com PCs, consoles, phones, and all other consumer-facing tech are in crisis, as AI data centers continue to swallow global memory demand in what has been dubbed the 'RAMpocalypse.' There has been plenty of humming and hawing over when prices will finally settle and computing will become affordable again, especially as we look down the barrel of $1,000+ home consoles. Even OpenAI CEO Sam Altman is concerned... just not about consumers. Altman argued in an interview on the Cheeky Pint podcast (via FRVR) that the DRAM shortage isn't the fault of data centers, but people daring to have memory in their own rigs for locally-run models. In so many words, the data centers eating up supply aren't the issue, you are. "From a global allocation perspective, actually, it's quite terrible to have everyone have a bunch of the world's limited supply of DRAM sitting in their own computer. And be used 5% of the time," Altman said. Altman Seems Confused That You Don't Want Data Centers To Have All Of The World's Memory Altman isn't strictly talking about gamers or everyday consumers, but it's not a leap to suggest that him referring to everyone having "a bunch of the world's limited supply of DRAM sitting in their own computer" would apply to those groups, not just users with locally-run open-source AI models. By his logic, having 32GB of RAM sitting in your rig for gaming is a complete waste of global supply. Why let consumers use up all that precious memory when it could be packed into a data center so that people can cheat on their homework assignments and generate deepfakes to harass women online? “A lot of people want local inference,” Altman said. “And somehow, they assume that the problem is solved now. Because we are going to have open-source models. We’re going to be able to run them on our own computers. And so we’re not going to need all the stuff [DRAM] in the data center. “You want maximum utilization in the cloud if you’re trying to deal with this very constrained environment,” he continued. “And somehow everyone’s, many people’s intuition is the opposite of that.” Altman argues that "we all have a shared interest" due to the world's limited silicon capacity, and that AI will discover efficiency gains and become a vital, everyday part of computing. Except, in this hypothetical future, nobody will be able to afford computers. It's a future that many are concerned about, where, instead of having local computers, everything is run through the cloud. Don't Worry, The AI Bubble Is Finally Beginning To Burst YouTube If there's a sliver of hope, though, it's that the AI bubble is finally starting to pop. Investors are growing cautious, US Democrats have proposed a bill to withhold bailouts for bankrupt AI firms, Google's value plummeted by $500 billion as its own AI summaries eat into traditional search ad revenue, and China's AI bubble has already burst. With giants like OpenAI and Anthropic burning through billions a year, Altman's dream of a world where data centers hoard every stick of memory and consumers get the scraps feels like a pipe dream.