OpenAI CEO Sam Altman has triggered widespread online backlash after arguing that the memory inside personal computers is largely a waste of global resources and would be better deployed in artificial intelligence data centers.
Speaking on an Oct. 29 episode of the “Cheeky Pint” podcast hosted by Stripe co-founder John Collison, Altman criticized the distribution of high-end hardware. He described it as “quite terrible to have everybody have a bunch of the world’s limited supply” of dynamic random-access memory sitting inside personal computers, noting that active device memory in consumer machines is utilized just 5% of the time.
RAM lets computers, smartphones, and consoles juggle open applications, but soaring demand from AI data centers has created a severe global shortage that is driving up prices for everyday electronics. When Collison asked why sky-high hardware costs had not naturally attracted new suppliers into the market, Altman responded bluntly: “I think the invisible hand is failing right now.”
Altman also dismissed the growing popularity of local inference—running open-source AI models directly on personal computers—arguing that scarce memory belongs in cloud servers running continuously rather than sitting idle in home machines. While Altman was addressing local AI rather than gaming rigs, home enthusiasts often run local models to maintain data privacy and avoid monthly subscription fees. Critics quickly pointed out the irony, noting that OpenAI is one of the largest buyers of the exact memory in short supply and that aggressive data center expansion by AI firms is driving the crunch.
The remarks ignited swift and widespread backlash across social media, with some online posts escalating into violent threats against Altman. Among the more restrained reactions was a post on X from Forbes senior contributor Paul Tassi, who wrote, “You know what I don’t think I like this guy very much.”
Altman’s comments coincide with a period of severe pricing pressure for computer components. Conventional DRAM contract prices paid by major manufacturers jumped roughly 93% to 98% during the first quarter of 2026 alone, according to data from research firm TrendForce. TrendForce projected that price increases would moderate to 10% to 15% in the fourth quarter, while warning that PC memory allocations could shrink as production shifts heavily toward servers. Other industry analysts expect the supply squeeze to persist at least through 2027.
At retail checkout, a 32GB kit of DDR5 RAM now commands about $500, up from roughly $400 in early April, in what investigators at Tom’s Guide have labeled a “cloud tax.” OpenAI represents a major catalyst for that demand. In October 2025, Altman signed letters of intent with Samsung and SK Hynix to secure memory for Stargate, an ambitious data center buildout. SK Group indicated that the project would require up to 900,000 DRAM wafers per month, translating to as much as 40% of global output based on production estimates from TechInsights.
Addressing why memory manufacturers have been hesitant to build new fabrication plants, Altman told Collison that the industry remains scarred by past cycles, recalling historical periods when a DRAM producer “would invest $20 billion of CapEx to make $19 billion.” The most recent industry bust occurred in 2023, when Samsung scaled back memory production after reporting a roughly 96% drop in first-quarter operating profit, according to .
Even as Altman advocated for shifting global memory capacity into data centers, OpenAI faced hurdles in financing those infrastructure ambitions. Reports from indicated the company initially sought at least $30 billion at a pre-money valuation of about $1.4 trillion. By Oct. 22, the target valuation had slipped to $1.25 trillion for a funding round led by investors from the United Arab Emirates and the investment firm Coatue, as reported by CNBC.
An Oct. 30 deadline for investor commitments saw OpenAI reduce the funding round to roughly 20 billion, with the initial cash tranche falling short after several participants scaled back their commitments amid a broader tech-stock selloff tied to AI spending concerns and a race ahead of rival Anthropic’s expected mid-November IPO. The Wall Street Journal reported that OpenAI ultimately raised less than half of its original $30 billion target.
Despite the scaled-back funding round, a $1.25 trillion valuation remains comfortably above the $852 billion figure OpenAI announced in March following a $122 billion fundraising effort, though it sits $150 billion below the valuation the company was pitching investors a month prior.
For consumers shopping for new hardware, Altman told Collison he hopes computing costs will eventually fall, though he emphasized that supply must expand dramatically first—a prospect TrendForce views as unlikely for PCs through 2027. In the interim, analysts advise shoppers to compare hardware specifications line by line rather than relying on sticker prices, as PC brands increasingly trim storage and memory configurations in mainstream models to manage costs. Consumers whose current machines remain functional may save money by holding off on purchases, though persistent supply constraints mean waiting offers no guaranteed relief from inflated component prices.
