Altman’s RAM Claim Sparks Controversy
OpenAI CEO Sam Altman attracted attention this weekend after suggesting that the world’s limited DRAM supply is being wasted by personal computers. In an interview he described it as “pretty terrible” that gamers and other PC users have DRAM that is used only about 5% of the time, and argued that the supply should be redirected to data‑center training workloads.
Altman’s remarks come amid a global shortage of RAM and high‑capacity memory chips such as HBM, which AI companies need in massive quantities to build and train large models. Samsung, one of the world’s biggest memory manufacturers, signed a deal with OpenAI last year to provide roughly 900,000 DRAM/HBM wafers each month — about 40% of its total output. Recent reports say Altman has secured a new agreement with Samsung as the shortage persists.
The criticism is not aimed at active gamers who regularly use their memory, but at dormant machines that sit idle while still holding RAM that could be repurposed for cloud workloads. Still, the suggestion that consumers are “stealing” supply has drawn sharp pushback from observers who note that modern PCs make extensive use of RAM in everyday tasks.
Industry data adds context. PC hardware shipments fell 20.1% in the third quarter of 2026, according to IDC, reflecting a slowdown in consumer demand. At the same time, RAM prices have surged, with Micron locking in pricing for its customers through 2030 and VRAM costs climbing so high that Nvidia and AMD may delay new graphics‑card generations until 2027.
Critics argue that Altman’s view overlooks how essential RAM is for gaming, content creation, and general computing. A PC without memory would not boot, and gamers typically allocate a large portion of their budget to fast, plentiful memory to achieve performance goals. The CEO’s comment, made by a leading AI figure, underscores a growing disconnect between AI firms and the broader market that funds their development.
Beyond memory, the rapid expansion of AI data centers raises additional concerns, including higher energy consumption, water usage for cooling, and the risk of unregulated AI arms races. While Altman did not address these broader impacts, the ongoing RAM constraints highlight how AI ambitions can strain resources that ordinary users rely on daily.
