Synapse

RAM for your computer costs about 7 times what it did a year and a half ago, and the CEO of OpenAI just said it's "quite terrible" that so much of it is sitting in your PC. His company signed preliminary deals for up to 40% of the world's supply last year. A basic 32GB kit was $59 in April 2025. The cheapest one in the US is over $400 today. Sam said this on Stripe's Cheeky Pint podcast. He was pushing back on people who think open-source AI running at home means we won't need all these data centers. His answer was that from a "global allocation perspective," everyone keeping memory at home and using it 5% of the time is a bad deal for the world. Which is a funny thing for him in particular to worry about. When you run an AI model on your own machine, OpenAI never sees a dime. You don't pay a subscription or a usage fee, and every time a free open model gets better, the computer you already own gets more useful. That's the one competitor Sam can't undercut on price, and it runs on the exact memory he'd rather see in a server rack. And the reason that memory got so expensive traces right back to the data centers. Back in October 2025, OpenAI signed letters of intent with Samsung and SK Hynix for up to 900,000 memory wafers a month for its Stargate project, which TrendForce figured was around 40% of everything the industry makes. Meanwhile the chipmakers have been shifting their factories toward the special high-bandwidth memory that AI chips use, since it pays way better. Samsung's average memory price jumped about 20% just last quarter, and TrendForce sees contract prices climbing another 10 to 20% this quarter. Your phone is feeling it too. IDC says smartphone shipments fell 22% last quarter compared to a year ago, the biggest drop they've ever recorded, and they blamed memory costs that are up something like 350%. I'll give Sam this much. A data center serving millions of people really does keep its memory busy around the clock, while your gaming PC sits there all night doing nothing. That's basically why the cloud exists. The people building these rigs at home know all that and still pay $400 for 32GB, because it lets them run any model they like, offline, without anybody else setting the price. Even if Sam's right that your PC only touches its RAM 5% of the time, a lot of people would rather own that 5% than rent it back from OpenAI by the token.