Sam Altman Says Your PC’s RAM Is a Waste. Why He’s Not Wrong (But Many Are Angry).

If you’ve recently checked the price of a new laptop or desktop, you’ve likely felt the sting of rising costs. OpenAI CEO Sam Altman recently added fuel to the fire, arguing that the memory sitting in personal computers is largely a waste of a limited global resource. In a candid podcast appearance, Altman suggested that the world’s scarce DRAM (dynamic random-access memory) supply would be better allocated to AI data centers rather than consumer devices where it sits idle most of the day. His comments, coming at a time when memory prices have skyrocketed, have ignited a fierce debate about the future of computing, the ethics of the AI boom, and what it means for your wallet.

The Backlash to Altman’s ‘Local AI’ Take

On John Collison’s “Cheeky Pint” podcast, Altman addressed the idea of “local inference”—running AI models on personal devices to reduce reliance on cloud data centers. He argued the opposite: when memory is this scarce, it belongs in cloud servers running at full tilt rather than spread across home machines that sit idle most of the day. He described the current situation as “quite terrible to have everybody have a bunch of the world’s limited supply” of DRAM in their PCs, noting that average usage is around 5%. He even suggested that the “invisible hand is failing right now” in terms of market allocation.

While a server shared by thousands of customers is technically more efficient than a single home PC, Altman’s remarks rubbed many the wrong way. Home AI enthusiasts value local processing for privacy, offline capability, and avoiding subscription fees. Furthermore, critics quickly pointed out the irony: OpenAI is one of the largest buyers of the very memory Altman claims is in short supply, with the company’s massive data center plans arguably driving the shortage in the first place. The backlash on social media was swift, ranging from criticism to threats, with even Forbes’ Paul Tassi quipping, “You know what I don’t think I like this guy very much.”

Why DRAM Prices Are in a Pressure Cooker

Altman’s comments arrive amid a severe memory shortage. According to TrendForce, conventional DRAM contract prices surged by 93% to 98% in the first quarter of 2026 alone. At retail, a 32GB kit of DDR5 memory now costs around $500, up from $400 in early April—a markup dubbed the “cloud tax” by Tom’s Guide. Analysts warn the shortage could persist through 2027 as production shifts toward high-margin server memory.

The primary driver of this demand is, unsurprisingly, AI. In October 2025, Altman signed letters of intent with Samsung and SK Hynix to supply memory for OpenAI’s “Stargate” data center project. This deal could require up to 900,000 DRAM wafers a month, representing as much as 40% of global output. With such massive consumption, memory makers have little incentive to expand consumer-focused production, especially given the industry’s memory of past busts, like Samsung’s 96% profit drop in 2023 that led to production cuts.

OpenAI’s Valuation Scales Back Amid Funding Challenges

Adding to the narrative, OpenAI faced difficulties in its latest funding round, highlighting the massive capital requirements of the AI buildout. The company initially sought at least $30 billion at a pre-money valuation of $1.4 trillion. By October, the valuation had slipped to $1.25 trillion, and the round was ultimately cut to roughly $20 billion, with reports indicating they raised less than half of the original target. This scaling back occurred amidst a tech-stock selloff over AI spending and competitive pressure from Anthropic’s anticipated IPO. While a $1.25 trillion valuation remains massive, the cuts underscore the financial strain of Altman’s ambitious data center plans.

Navigating the Memory Market as a Consumer

For everyday shoppers, the practical takeaway is straightforward: be smart about your next purchase. With PC brands trimming storage in mainstream models to keep costs down, comparing specs line-by-line is more important than ever. If your current machine still meets your needs, holding off may save you money, though with supply constraints expected to continue into 2027, waiting isn’t a guaranteed win. Most importantly, unless you’re actively running local AI models or playing memory-intensive games, there’s no reason to pay a premium for top-tier memory you’ll never use.

The Bigger Picture: AI’s Hidden Tax on Consumers

Sam Altman’s comments may have been tactless, but they highlight a harsh economic reality: the AI revolution is reshaping global supply chains, and consumers are feeling the pressure. As memory becomes a luxury commodity, the era of cheap, over-provisioned consumer PCs may be waning. For now, buyers must navigate this tight market with caution, prioritizing actual needs over marketing hype.

Business, Computers, Electronics, Finance, Internet, News, People, Science, Shopping, Society, Technology, Artificial Intelligence