RAM crisis highlights PC vs data center needs

Sam Altman, leader of OpenAI, which created ChatGPT, has said that personal computers are using too little of the limited RAM available globally. He pointed out that DRAM, or dynamic random-access memory, in PCs is only used about 5% of the time, while data centers could make better use of it.

AI drives demand for data center RAM

Altman argued that RAM in data centers is more valuable than in personal computers because it serves thousands of users at once. AI systems, which require large amounts of memory, are increasing the need for data center RAM, which operates continuously compared to the intermittent use in PCs.

Shift in RAM manufacturing priorities

RAM manufacturers are focusing more on server chips due to higher profitability. Micron, a major RAM producer, has moved away from consumer RAM under its Crucial brand, which was once a key source for PC users. This shift reflects the growing demand from AI-driven data centers, which offer more lucrative contracts.

RAM shortage expected to last until 2030

Executives from Micron and Samsung, another top RAM maker, predict the shortage will persist until 2030, with prices likely to rise sharply by 2027. Some users prefer running AI tasks on their own PCs to avoid cloud services, which might involve data privacy risks or subscription costs.

OpenAI’s needs challenge local AI preferences

Altman’s position is shaped by OpenAI’s growing demand for RAM, especially for projects like Stargate, which require vast amounts of memory. This stance contrasts with users who prefer running AI locally on their own devices instead of relying on centralized services.