Sam Altman argued on the Cheeky Pint podcast that, with DRAM supplies limited, it may be more efficient to keep memory in cloud data centers—where it can be used continuously—than in personal computers that sit idle part of the time. He was responding to interest in running AI models locally and acknowledged that people want local inference; his remarks were about allocating scarce memory, not a plan to remove RAM from PCs or target gaming computers.

The hardware crunch is being driven by strong AI demand, low inventories, geopolitical tariffs and manufacturers directing advanced production toward enterprise servers instead of consumer memory. The article argues that AI companies are a major contributor to the scarcity, while consumers are bearing higher costs for PCs, phones and consoles.

The scale of AI demand is underscored by reports that OpenAI has pursued deals for about 40% of global RAM production—roughly 900,000 DRAM wafers a month—from suppliers including SK Hynix and Samsung, making Altman’s comments sound especially tone-deaf to critics.

Memory manufacturers are increasingly serving data-center customers: Micron says it can use the same production lines for different markets and forecast fiscal first-quarter 2027 revenue of $61.5 billion, plus or minus $1.5 billion. More than 75% of its shipments were committed for the year, while its data-center business reported an 87% non-GAAP gross margin last quarter.

The shortage is delivering record profits for suppliers. SK Hynix reported quarterly operating profit of 83.3 trillion won, said more than half of its 2027 and 2028 production was already booked, and warned tight supply could last beyond 2030; Samsung reported a preliminary operating profit of about 107.4 trillion won for July through September, its fourth consecutive record quarter.

Altman said manufacturers are hesitant to build capacity because previous overinvestment led to oversupply and financial losses; Samsung, for example, cut production in 2023 after first-quarter operating profit fell roughly 96%. That caution, alongside manufacturers’ focus on returns, makes quick relief for gadget buyers unlikely.

Consumers are already seeing the impact: OnePlus, Samsung, Oppo and other brands have raised smartphone prices in India as mobile-memory costs climb, putting phones below Rs 10,000 under pressure. PlayStation price increases and a $1,300 Xbox handheld with ROG also highlight the squeeze on gaming hardware, while the next console generation faces cost concerns.

A February 2026 Gartner forecast projected that combined DRAM and SSD prices could rise 130% by year-end, PC prices could increase 17% from 2025, and global PC shipments could fall 10.4%; these are projections, not final results. Weak demand is another concern: Microsoft reported a 22% decline in Windows OEM and Devices revenue in its fiscal first quarter, and Intel expected the PC slowdown to continue through at least the first half of 2027.

Micron shares fell 5.30% on October 30 to $884.74 amid a broader technology-stock pullback. Its estimated, unconfirmed December 16 earnings report is a key checkpoint, particularly if revenue misses guidance or data-center growth fails to offset weakness in Mobile and Client.

There are also signs of investor caution around OpenAI: its fundraising plan was reportedly cut from at least $30 billion at a pre-money valuation of about $1.4 trillion to roughly $20 billion at a $1.25 trillion valuation, and The Wall Street Journal reported on October 30 that the company had raised less than half its original target. Broader strains cited in the article include AI stock sell-offs, Amazon’s data-center expansion weighing on free cash flow, and concern that Google’s AI features could reduce search-ad revenue.

Anthropic’s 2026 IPO filing illustrates the scale and risks of the boom: it reported $16.3 billion in first-half revenue, nearly $9 billion in AI infrastructure spending, an $80.7 billion net loss and $800 billion in compute commitments over ten years. The filing also reportedly warned that models could sabotage code, manipulate information or behave as if blackmailing people.

The article sees investor skepticism and customer cost-cutting as possible signals of slower AI spending, citing a 40% drop in China’s STAR50 index that erased about $480 billion in market value, falling U.S. stocks, and organizations limiting or abandoning AI over costs, modest productivity gains, errors or customer dissatisfaction. At the same time, adoption is broad but uneven: a June 2026 Pew survey found 49% of 5,119 U.S. adults had used chatbots and 24% used them daily, while Stanford’s 2026 AI Index found 88% of surveyed organizations used AI in at least one business function in 2025, but AI-agent deployment remained in the single digits across nearly all functions.