RAM is getting expensive fast. That’s terrible news if you’re planning to buy or build a PC in 2026—but it may finally put a lid on two years of breathless “AI PC” marketing.
Driven by AI data centers hoovering up memory, RAM and flash prices have spiked. Ben Yeh, principal analyst at Omdia, says “mainstream PC memory and storage costs rose by 40 percent to 70 percent” in 2025, with those costs pushed straight onto buyers.
Yet despite the hit to component budgets, global PC shipments actually grew last year: Omdia pegs 2025 growth at 9.2 percent over 2024, while IDC puts it at 9.6 percent. The party may not last. IDC’s Jean Philippe Bouchard warns that “the year ahead is shaping up to be extremely volatile.”
Volatile or not, one thing is already clear: the RAM crunch is colliding head‑on with the industry’s AI PC story.
Less RAM, less patience for AI PC buzzwords
Both Omdia and IDC expect PC makers to respond to the shortage in two ways: raising prices and cutting memory specs.
IDC expects PC prices to jump “15 to 20 percent” and for RAM configurations to “be lowered on average to preserve memory inventory on hand,” Bouchard says. Yeh sees “leaner mid to low-tier configurations to protect margins.” In plain language: entry‑level and midrange machines will come with less memory than you’d hope, at prices you won’t like.
That’s a problem for the current crop of “AI PCs,” which generally require at least 16GB of RAM to run on-device models comfortably. IDC’s Jitesh Ubrani tells Ars that “these RAM shortages will last beyond just 2026, and the cost-conscious part of the market is the one that will be most impacted.”
Vendors, he says, will “prioritize midrange and premium systems to offset higher component costs, especially memory.” If you were waiting for cheap AI‑branded laptops, don’t hold your breath.
Worse for OEM marketing teams, enthusiasm for AI branding was already cooling before the shortage.
“PC OEMs had trouble selling the on-device AI message even before the memory shortages,” Ubrani says. For many buyers, especially in IT, local AI isn’t a must‑have feature—it’s a nice‑to‑have if it’s free.
“General interest in AI PCs has been wavering for a while, since cloud-based options are widely available and the use cases for on-device AI have been limited,” Ubrani notes. And there’s a twist OEMs won’t like: “This indifference (between on-device and cloud-based) from a demand perspective might work in favor of PC OEMs, as they don’t need to provide large amounts of RAM.”
Translation: if customers don’t care whether AI runs locally or in the cloud, there’s no reason to ship big, expensive RAM configs just to tick an “AI PC” box on the spec sheet.
With Ubrani seeing RAM prices staying unstable until at least 2027, the industry may be stuck in this low‑RAM, low‑AI‑enthusiasm holding pattern for a while.
Dell’s XPS whiplash shows the pivot
We’re already seeing how quickly OEMs can pivot away from AI‑first branding when the economics shift.
In 2025, Dell actually killed off its long‑running XPS consumer line of laptops and desktops. One reason, according to Kevin Terwilliger, VP and GM of commercial, consumer, and gaming PCs at Dell, was that “the AI PC market is quickly evolving.” At the time, Dell was clearly trying to reposition around the new AI narrative.
Fast‑forward to CES 2026, and XPS is back. Dell resurrected the brand and quietly moved its messaging to things people actually understand and care about: build quality, battery life, display.
Terwilliger also changed his tune. “… what we’ve learned over the course of this year, especially from a consumer perspective, is they’re not buying based on AI,” he said at a press briefing ahead of the show. “In fact, I think AI probably confuses them more than it helps them understand a specific outcome.”
That’s a brutal admission for a sector that has spent two years insisting that the “AI PC” is the next upgrade cycle.
Even Microsoft knows the experience isn’t there yet
The hardware story would be stronger if the flagship AI experiences were clearly delivering. They aren’t.
In December, The Information reported that Microsoft CEO Satya Nadella sent an email to engineering leads expressing frustration with the consumer version of Copilot. He reportedly wrote that tools for connecting Copilot to Outlook and Gmail “for [the] most part don’t really work” and are “not smart.” The outlet also reported that Nadella reassigned some of his own responsibilities so he could spend more time on Copilot.
Microsoft declined to comment on the report, but if accurate, it underlines the gap between the AI PC pitch and what people actually see when they click the Copilot icon.
If the software layer feels half‑baked, it’s much harder to convince buyers they need a “next‑gen AI PC” rather than just a solid laptop with decent battery life.
From marketing slogan to meaningful feature
None of this means AI on PCs is doomed. Ubrani is explicit that the concept itself isn’t the issue—it’s the way it’s been sold.
“The idea of AI PCs isn’t inherently a bad thing,” Ars notes. The problem is that over the last two years the term has often felt like a solution in search of a problem. Vendors leaned on generative AI to sell “more or more powerful devices, even if they didn’t need them or local AI processing.”
The RAM crunch changes the calculus. When every extra gigabyte eats into already‑thin margins, PC makers have to justify it. They can’t just spray “AI PC” across the box and hope buyers upgrade for fear of missing the next wave.
As shortages drag on and customers remain unmoved by vague AI promises, there’s at least a small upside: less empty hype, more pressure to prove value. If AI PCs are going to stick around as a category, they’ll have to answer a basic question they’ve mostly dodged so far:
What real problem do they solve that a cheaper, leaner PC plus cloud services doesn’t?
