The next time a phone company claims its newest device was “built for AI,” pay attention. That phrase now has a very practical meaning: it will cost a lot more.

For years, smartphone prices increased as cameras improved, screens grew larger, chips got faster, and devices felt more high-end. Now, an unseen and less glamorous part of the device is taking center stage: memory. This includes storage for photos and videos as well as the fast memory that transfers data quickly enough to run artificial intelligence.

That’s why your next smartphone will cost significantly more: the economics within the device have changed.

AI is causing a memory shortage from two sides at once. First, phones and computers require more memory to run AI directly on the device. Companies want AI features to operate faster, more privately, and with less reliance on cloud servers. This means the phone in your pocket has to do more of the work itself. A device that summarizes messages, edits photos, understands voice commands, and helps write emails can’t function well on yesterday’s memory budget.

Second, the global AI boom is redirecting memory chips from consumer electronics to data centers. The massive server farms behind chatbots, AI search tools, coding assistants, and image generators need vast amounts of memory and storage. Chipmakers naturally prioritize the largest and most profitable buyers. As a result, phones, laptops, appliances, and cars are now competing with AI data centers for parts that were once common.

The pressure extends beyond consumer electronics. Memory chips are also embedded in military systems, medical devices, factory equipment, and much of the infrastructure modern life depends on. When a small but essential component gets more expensive or harder to find, the cost ripples outward through the entire supply chain. What looks like a technical problem inside a smartphone can eventually become a price problem across many industries. Many of the world’s most advanced memory chips are manufactured in Taiwan, making this supply crunch a foreign-policy concern as well as an economic one.

Congress passed the CHIPS and Science Act in 2022, spending roughly $52 billion to incentivize domestic semiconductor manufacturing and reduce that dependence. Factories are being built, but they take years to come online and are not yet producing the most advanced chips at scale. The current shortage is a reasonable moment to ask whether that investment is working, whether it is working fast enough, and whether American consumers should expect any relief from it any time soon.

Apple’s own product plans illustrate how important memory has become. The company is preparing one of the biggest changes ever to its chip strategy. The reason is clear: Apple wants to speed up technology for on-device AI and more complex software. Memory bandwidth is now a key factor in whether a computer can handle AI tasks, because AI requires large amounts of data to move quickly through the system.

The hardest part for consumers is that memory is invisible. A better camera or larger screen feels like an upgrade. More memory bandwidth may seem like an insignificant feature buried in the product’s fine print. But this is more than a minor spec upgrade: it signals that the AI era is turning memory into a premium feature.

The consumer impact is already noticeable. Apple recently raised prices on Mac models, with some increases reaching as much as 25%. The company said rising demand from AI data centers had pushed memory and storage costs to levels it could no longer absorb. “We have never seen a component price increase this much, this quickly,” Apple said in a statement.

If Apple, one of the world’s most powerful buyers, must adjust its plans and product designs because memory is scarce and expensive, other device makers will feel it even more.

TrendForce, a semiconductor research firm, warned that surging memory prices could significantly raise consumer-device prices, with some devices potentially seeing 30%+ increases. Tight supply may persist into late 2027, and manufacturers are responding by raising prices or reducing specs to offset the costs.

Most consumers have been told to expect the cost of AI to show up in water and electric bills because data centers consume huge amounts of both, but the costs won’t stop at the utility meter. Consumers will likely pay twice for the AI boom: first through higher utility bills from data centers, and then again as the boom drives up the prices of everyday devices.

The frustration is that many consumers don’t even want these AI features, at least not enough to pay hundreds more for them. Plenty of people use their phones to text, take photos, check email, shop, and scroll. They are not seeking a device that summarizes their inbox or rewrites their messages. But the industry has already redesigned its hardware around AI, and there are few versions of the phone without it.

When the largest technology companies race to build AI infrastructure, the cost doesn’t stay in Silicon Valley. It shows up in the price of just about everything, paid by ordinary buyers who may never use a single AI feature. Most people won’t notice until they reach the checkout, where the sticker shock will be very real.