OpenAI chief Sam Altman warns of an unsustainable boom in building compute capacity
Sam Altman, CEO of OpenAI, described the boom in building AI compute infrastructure as “unsustainable silliness” and criticized neocloud providers building capacity without corresponding demand. He described the expansion of OpenAI itself as profitable.
Sam Altman, CEO of OpenAI, described the current boom in building AI compute infrastructure in an interview as “unsustainable silliness” (unsustainable silliness). His criticism targets neocloud providers announcing massive increases in capacity without corresponding demand or revenue to justify such investments. According to Altman, many of them operate on the principle that “price is no object” and will build enormous amounts of compute capacity even at higher prices. By contrast, he described the expansion of OpenAI itself as profitable and backed by real demand.
Altman also acknowledged that progress at OpenAI in reducing costs and improving efficiency could make the situation worse – if this happens quickly enough, today's expensive infrastructure investments could turn into bad bets. He also mentioned that a broader downturn in the industry could make it harder for OpenAI to pay for capacity it has already committed to, but he considers this risk manageable. OpenAI does not currently sell compute capacity to third parties, but Altman did not categorically rule out this possibility.
According to the source, the current tone from Altman is markedly more cautious than before, when Dario Amodei, CEO of Anthropic, accused Altman of investing money in building compute capacity recklessly (“YOLO”). However, it is unclear whether this represents a fundamental change in the stance of Altman on building compute capacity or merely a change in tone.
Why it matters
The warning points to growing concerns about excessive investment in AI infrastructure by providers without assured demand – if compute becomes cheaper and more efficient, today's expensive capacity commitments could prove unprofitable. This is particularly relevant to investors and companies entering into long-term contracts for compute capacity or considering entering the AI infrastructure business.
Relevant practical impact
What this means
For a business
The warning signals a risk of excessive investment in compute capacity by providers without corresponding demand – if compute becomes cheaper and efficiency improves, today's expensive commitments could become unfavorable, which is relevant to companies entering into long-term capacity contracts.
Risks and complianceCheck the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.