Apple's expected iPhone price increase signals the memory crunch has hit consumer hardware

When Apple unveils its next iPhones this week, the unwelcome news may be on the price tag. According to The Verge, a price increase from the industry's most powerful buyer would be the clearest indication yet that surging memory costs have become unavoidable across the hardware supply chain, and that the shortage driving them has years left to run.

The numbers behind the crunch

The component at the center of the story is DRAM, the working memory a phone relies on to open apps, render pages and run software. Counterpoint estimates that DRAM prices for smartphones rose roughly 56 percent in the first quarter of 2026 compared with the quarter before, then around 83 percent in the second quarter. For 16GB of smartphone DRAM specifically, the firm puts the cost at about $42 in the second quarter of 2025 and roughly $181 a year later, an increase of more than 300 percent.

As The Verge points out, those figures are not necessarily what Apple pays; few companies have more leverage over suppliers. But the swing shows how radically the economics of building a high-end phone have shifted in twelve months.

AI rewrote the memory market

One factor is concentration. Three manufacturers control roughly 90 percent of the memory market, according to Counterpoint, which estimates Samsung held 39 percent of it in the second quarter of 2026, followed by SK Hynix at 26 percent and Micron at 25 percent.

The bigger factor is AI. Data centers do not buy the same memory as phones: their accelerators depend on high-bandwidth memory, or HBM, which stacks chips together to move data far faster. It is harder to produce and considerably more profitable to sell, and Micron estimates that making a given amount of HBM consumes roughly three times as many wafers as the same amount of conventional DRAM. With buyers like Nvidia, AMD, Meta and Microsoft willing to sign multiyear commitments, capacity is tilting toward AI. Samsung executive vice president Jaejune Kim said on the company's latest earnings call that Samsung prioritizes customers who can guarantee committed future demand.

The profitability is unmistakable. SK Hynix posted a record 76 percent operating margin last quarter, up from 41 percent a year earlier. Micron's adjusted gross margin reached a record 85 percent, and Samsung's semiconductor profits jumped roughly 250-fold year over year.

AI is also inflating demand for ordinary DRAM. Device makers want to run smaller AI models directly on phones and PCs, which requires more conventional memory, and more advanced types of it. SK Hynix president Song Hyun-jong described a structural shift in which AI memory and conventional memory are growing at the same time.

The shortage predates ChatGPT

Counterpoint's associate director David Naranjo told The Verge that the situation is not simply a case of data centers eating the RAM supply, because the products involved are different. The squeeze also started before generative AI took off. Micron concluded back in 2021 that shrinking transistor-level gains could no longer deliver enough capacity on their own, and that the industry needed entirely new fabrication plants. Then the pandemic-era electronics boom reversed, leaving memory makers with excess inventory and losses, and expansion plans were slowed, just as AI demand arrived with force.

Relief is years away

The long-term fix is capacity, and capacity takes time. In July, Micron poured the first concrete for a manufacturing complex near Syracuse, New York, planned to include 2.4 million square feet of cleanroom space, which would make it the largest semiconductor manufacturing site in US history by that measure. Micron does not expect meaningful output from the site until 2030, according to The Verge.

Why it matters

Memory has gone from a quietly cheap commodity to a first-order cost constraint for anyone building or programming hardware. For developers, rising DRAM prices raise the bill of materials for phones, laptops and consoles, while the push to run AI models on-device pushes minimum memory requirements upward. For consumers, the decades-long pattern of devices getting more capable without getting dramatically more expensive is breaking, and Apple raising prices would confirm that even the largest buyer cannot escape it. With three companies controlling about 90 percent of supply and new factories not producing until 2030, the term "memory shortage" appearing in 473 company transcripts last quarter, per AlphaSense data cited by The Verge, looks less like a talking point and more like a multiyear operating condition.