OpenAI CEO Sam Altman Argues DRAM Should Prioritize Data Centers Over PCs
OpenAI CEO Sam Altman argued that during global memory shortages, DRAM resources should be prioritized for centralized data centers rather than consumer PCs used for local AI inference. He suggested that widespread low-utilization local AI deployment leads to a wasteful allocation of limited chip supplies. This perspective highlights the intense global competition for semiconductor and memory resources driven by the booming AI infrastructure sector. It underscores the ongoing tension between centralized cloud-based AI scaling and the growing trend of decentralized local AI hardware adoption. Critics point out a major flaw in Altman's logic: consumer PCs are multi-purpose devices used for gaming, professional software, and multitasking, meaning DRAM demand in PCs cannot simply be equated to or replaced by cloud AI workloads. Meanwhile, memory suppliers like Samsung expect AI-driven supply constraints to persist through 2028.
## BACKGROUND
DRAM (Dynamic Random-Access Memory) is a type of volatile semiconductor memory widely used as the main system memory in modern computers and graphics cards. In recent years, the massive expansion of artificial intelligence infrastructure has driven up demand, creating severe supply shortages and increasing hardware costs across the technology industry.