Altman paired the warning with a strong long-term call. He said AI demand would grow “enormously” as models write custom software in real time and are used in robots, spacecraft and science, and that the top token user in early 2020 consumed about 100,000 tokens a month while this year’s heaviest user consumed 100 billion a month [1]. Asked about the biggest bottleneck, he listed supply chains, financing, AI safety and social acceptance, and enterprise adoption, then named chip supply chains and data center infrastructure as the biggest limits for the next several years and said more investment in the supply chain is needed [1].
On profitability pressure from OpenAI’s spending, he told the executives that a vertically integrated infrastructure cheaper than any other company’s would let OpenAI sell tokens at low margins and at large scale [1].
The remarks matter for memory makers because Samsung, SK hynix and Micron have been planning capacity around a multi-year shortage. Altman’s visit itself was read in Korea as a push for more DRAM supply, and no new supply contract was announced (our earlier coverage). Winbond’s chairman made a similar point from the supplier side this week, saying cloud buildouts are funded through 2028 but memory prices have hit their limit (MLQ.ai report).