AI data centers continue to devour global memory production capacity, triggering an industry crisis known within the sector as "Memory Doom." All consumer electronics, including PCs, consoles, and smartphones, are mired in hardware price hikes. There has been ongoing debate about when memory prices will fall and when consumer computing devices will return to reasonable price points—a topic of even greater concern now that the next generation of home consoles is likely to break the $1,000 barrier. Even OpenAI CEO Sam Altman has expressed "concern" about this, though he has never cared about the plight of ordinary consumers.
In an interview on the Cheeky Pint podcast, Altman stated that the root cause of the DRAM shortage is not the massive demand from data centers, but ordinary users insisting on configuring memory in their personal devices to run local AI models. In short: data centers consuming memory production capacity is not the problem; your personal computer is.
"From the perspective of global resource allocation, it's terrible to have everyone's computer holding onto the world's limited DRAM inventory while only being used 5% of the time," said Altman.
Altman's remarks were not specifically targeted at gamers or ordinary consumers, but it is clear that his description of "everyone's computer hoarding the world's limited DRAM inventory" applies equally to these groups, not just to those running local open-source AI models.
By his logic, installing 32GB of RAM in your computer for gaming is a complete waste of global resources. Why should consumers occupy this precious memory? It would be better to stuff it all into data centers so people can use it to cheat on homework and generate deepfake videos.
"Many people want local inference capabilities," Altman said. "And for some reason, they think the problem is solved—that with open-source models, they can run them on their own computers without needing the DRAM in data centers."
He continued, "To cope with such extreme resource scarcity, the cloud should achieve maximum resource utilization. But for some reason, many people's intuition is exactly the opposite."
Altman believes that with limited global semiconductor production capacity, "our interests are aligned," and AI will eventually achieve efficiency gains, becoming a routine and indispensable part of computing. However, in the future he envisions, no one will be able to afford a personal computer. Many are worried about this future, where personal computing devices will cease to exist and all computation will be done via the cloud.
However, if there is a glimmer of hope, it is that the AI bubble is finally beginning to burst.
Investors are becoming increasingly cautious. The U.S. Democratic Party has proposed a bill refusing to bail out bankrupt AI companies. Google's market value plummeted by $500 billion as its AI summary feature cannibalizes traditional search advertising revenue. Industry giants like OpenAI and Anthropic burn through billions of dollars annually. Altman's dream of a future where "data centers hoard all memory, leaving consumers with scraps" is ultimately nothing more than an unrealistic fantasy.