The big picture: The AI boom has siphoned most of the tech industry's RAM supply, driving up prices and making consumer electronics less affordable. Recent comments from Sam Altman, arguably the man most responsible for launching the boom, frame the situation as a conflict between cloud and local compute, and he clearly favors the cloud.
In an interview on the YouTube channel Cheeky Pint, OpenAI CEO Sam Altman argued that it's more efficient to direct the world's tightening supply of memory to AI data centers than to local devices. The comments drew immediate backlash as hardware prices continue to climb.
About 29 minutes in, Altman says it's "quite terrible" for DRAM to be locked inside consumer hardware, where it may be used only 5% of the time. Because data centers run their hardware almost constantly, he argues, they can push the global supply closer to full utilization.
Altman specifically mentioned PCs with local AI capabilities and open-source LLMs. Devices like Microsoft's Surface Laptop Ultra can ship with up to 128GB of unified RAM to support local AI workloads, including open-source models like DeepSeek, which OpenAI sees as competition.
However, given the AI boom's broader impact on consumer devices, Altman's claim could be seen as an attack on the very idea of owning hardware. Many fear that the companies driving the AI boom ultimately want owners to become renters of cloud services running on AI data centers.
The data center buildout has quintupled DRAM costs since last year, pushing consumer hardware prices to unaffordable highs. Game consoles are approaching the $1,000 mark, dragging sales downward. PC shipments have also dropped significantly this year for similar reasons, contributing to Intel's worst stock price collapse in decades.
– Jaydeep (@_jaydeepkarale) October 31, 2026
Meanwhile, high-end graphics cards like Nvidia's RTX 5080 and 5090 seem to have disappeared from store shelves after reaching astronomical prices. Even cameras could become more expensive because of RAM shortages, according to Canon.
Memory makers disagree on how long shortages will last. SK Hynix estimates that current conditions will continue past 2030, while Micron recently claimed that humanoid robots could keep demand high if the AI bubble pops (which might already be happening).
Acer's CEO, however, accuses the memory makers of wanting to prolong the situation to maximize profits, and thinks prices could start sliding as soon as next year.