When Apple unveils its next generation of iPhones this week, consumers should brace for an unwelcome surprise: a higher price tag. A price increase from the world's most valuable tech company would be the clearest signal yet that soaring memory costs have become impossible to ignore. And according to industry analysts, there is no end to the memory crunch in sight.

Welcome to what some in the industry have dubbed “chipflation” or “RAMageddon.” After decades of steadily falling memory prices that made consumer electronics more powerful without making them dramatically more expensive, the trend has abruptly reversed. Corporate transcripts now contain frequent references to memory prices and shortages, reflecting a problem that is pinching every corner of the technology industry, from smartphones and laptops to game consoles and AI servers.

The AI connection isn't the whole story

The obvious culprit is artificial intelligence. AI's rapid expansion has created enormous demand for memory chips, both in data centers and in consumer devices that now run on-device AI models. But the shortage was already brewing before ChatGPT took the world by storm. AI simply poured gasoline on a fire that memory manufacturers had been struggling to contain for years.

The current market is the result of a complicated and highly profitable reshuffling of the memory industry. For years, manufacturers could increase output by packing more chips onto each silicon wafer through technological advances. But those gains began shrinking and taking longer to achieve. By 2021, executives at Micron, one of the three largest memory makers, concluded that technology alone would no longer generate enough capacity to meet long-term demand. The industry needed to build massive new fabrication plants.

Then the memory market collapsed. Pandemic-era demand for computers, tablets, and phones had pulled future sales forward, and when consumer spending weakened, manufacturers were stuck with excess inventory. They lost money, cut production, and slowed their expansion plans. Just as the market began recovering, generative AI ignited a wave of demand far larger and more memory-intensive than anyone had anticipated.

Three companies control nearly all memory supply

The memory business is extraordinarily concentrated. Samsung, SK Hynix, and Micron together control about 90 percent of the global memory market, according to estimates from Counterpoint. In the second quarter of 2026, Samsung held roughly 39 percent, SK Hynix had 26 percent, and Micron had 25 percent. That leaves the entire world dependent on a handful of companies deciding how to allocate limited capacity between AI infrastructure and consumer electronics.

Memory inside phones and computers generally falls into two categories. DRAM, or dynamic random-access memory, provides temporary storage for running applications and loading data. NAND flash memory offers longer-term storage for photos, files, and software. Both are in short supply, but DRAM has become the more critical bottleneck.

AI data centers need a special form of DRAM called high-bandwidth memory, or HBM. Unlike the DRAM in your phone, HBM is designed to move enormous quantities of data at blazing speeds, which is essential for training and running complex AI models. It is made by stacking memory chips together with advanced packaging, and it consumes significantly more manufacturing capacity than conventional DRAM.

“It's not as simple as saying data centers are consuming RAM,” says David Naranjo, an analyst at Counterpoint. “The RAM is not the same.”

Micron estimates that producing a given amount of HBM requires roughly three times as many wafers as producing the same amount of conventional DRAM. HBM is also more lucrative to sell, making it easy for manufacturers to prioritize AI customers like Nvidia, AMD, Meta, and Microsoft, who are willing to sign multiyear commitments and pay premium prices.

“The three big memory guys are basically allocating the capacity that they have to these companies,” Naranjo says.

Why this means higher phone prices

As memory makers shift capacity toward HBM, less conventional DRAM is available for phones, PCs, and other consumer devices. At the same time, those devices need more memory than ever because they are increasingly running AI features directly on-device, which requires larger and more sophisticated DRAM configurations.

The numbers are staggering. Counterpoint estimates that smartphone DRAM prices jumped about 56 percent in the first quarter of 2026, and then surged another 83 percent in the second quarter. For a 16GB DRAM module used in a smartphone, the cost went from roughly $42 in the second quarter of 2025 to about $181 a year later, an increase of more than 300 percent.

Not every company pays those prices. Apple has extraordinary leverage with suppliers and may negotiate better terms. But the figures illustrate how dramatically component costs have changed across the industry.

The impact is already visible. Microsoft recently raised Xbox prices by another $100 to $150, leaving some models as much as $300 more expensive than at launch. It also increased prices on some Surface laptops and tablets by $500 over their original starting prices. Meta added $100 to the price of its Quest 3 headset. Samsung and Google have each raised the starting prices of some new phones by $100 this year.

Apple faces its biggest test yet

Apple is perhaps the most revealing case study because it should be best positioned to weather the storm. Its scale gives it enormous leverage with suppliers, and its customers have proven willing to pay premium prices. Yet Apple has already raised prices across its Mac and iPad lineups, and now the iPhone is under pressure.

Rather than launching its full lineup at once this fall, Apple is expected to introduce the iPhone 18 Pro and Pro Max alongside its first foldable iPhone, while holding the standard iPhone 18 and a refreshed Air model until spring 2027, according to recent reports. That concentrates its fall releases at the most expensive end of the market, exactly where there is the most room to protect margins.

Counterpoint's supply-chain tracking shows Apple preparing roughly 10 percent more iPhone 18 Pro and Pro Max units for September through December than it allocated to the equivalent iPhone 17 models during the same period last year. On Apple's most recent earnings call, then-CEO Tim Cook described memory pricing as a “100-year flood,” citing exponential increases that forced the company to raise prices.

The iPhone 18 Pro could start at $1,299, according to calculations from TechInsights cited by The Wall Street Journal. That would be $200 more than the iPhone 17 Pro. Analysts suggest Apple may be able to pull off these higher prices because its phones behave more like luxury goods than interchangeable electronics. The company has already gained market share by keeping prices steady while lower-end competitors raised theirs.

New fabs are coming, but very slowly

The obvious long-term solution is to make more memory. But building a new chip factory takes years, even under the most accelerated timelines.

Micron is currently constructing a massive manufacturing complex near Syracuse, New York. The site, once completed, will contain 2.4 million square feet of cleanroom space, making it the largest semiconductor manufacturing site in US history and Micron's biggest facility anywhere in the world. Yet the cleanrooms will occupy only a fraction of the campus, which will span the size of roughly 350 football fields once fully built.

The complexity is difficult to overstate. Micron executives describe the challenge of keeping microscopic particles off wafers as akin to building all of New York State without allowing a single ant anywhere in the entire territory. The facility requires enormous systems for water treatment, power distribution, air handling, and vibration control, all working together in perfect precision.

Micron broke ground in January 2026 and poured the first concrete in July, ahead of schedule. But essential systems won't be fully activated until the end of 2028, and the company doesn't expect meaningful output until 2030. Even an accelerated timeline means more than six years from permitting to production. The company's new Idaho fab is further along and is expected to begin wafer output in mid-2027.

No relief until 2028 at the earliest

Micron has roughly doubled its capital expenditure this year, exceeding $25 billion, but it cannot identify a clear point when supply will catch up with demand. Demand continues to grow almost as quickly as manufacturers can expand production. Micron executives say the shortage will extend beyond 2027, with no clear end in sight.

Counterpoint expects capacity to catch up with demand in the best-case scenario by late 2027 or early 2028. IDC similarly forecasts the shortage will last through 2027 and well into early 2028 before meaningful relief arrives. But relief will likely not mean a return to old prices.

“Not to say prices will come down to 2025 levels, but the price increases will stabilize at a new normal,” says Nabila Popal, a senior director at IDC. That new normal, she adds, is “at least triple what they used to be.”

What happens next

PC and smartphone manufacturers cannot absorb component costs that have multiplied several times over indefinitely. They can raise prices, reduce device counts, include less memory, or focus exclusively on high-end models where profit margins are large enough to absorb the added cost. Industry analysts are already seeing smartphone companies reduce expected shipment volumes while shifting toward more expensive models with premium features.

This means more OLED displays, larger storage configurations, advanced camera systems, and other upgrades designed to justify higher prices. As a result, industry revenue could remain flat or even rise while unit sales decline. The memory shortage isn't just making electronics more expensive; it is reshaping which electronics companies choose to build in the first place.

Samsung, SK Hynix, and Micron are all investing heavily in additional capacity. SK Hynix has accelerated the completion target for its massive Yongin Semiconductor Cluster from 2045 to 2033 and says it will double capacity within five years, yet even that may not keep pace with AI-driven demand. Samsung and SK Hynix recently announced plans to invest a combined 800 trillion won, roughly $588 billion, in four new memory factories in South Korea.

A potentially important fourth player is emerging in China. CXMT is building additional fabrication plants, though its technology lags behind the three dominant manufacturers. Apple has reportedly been testing CXMT chips, but would need approval from the Trump administration to do business with the company on a large scale.

The other possibility is that demand collapses before supply catches up. If the AI investment bubble bursts, memory manufacturers would suddenly face a very different competitive landscape. But for now, long-term contracts signed by AI companies with memory suppliers make the immediate future relatively predictable. As one analyst put it, “2027 is pretty locked and loaded.”

Yet even if price increases slow from their current breathtaking pace, they are unlikely to reverse. New fabs take years to come online, AI data centers continue to expand, and consumer devices require ever more memory. Memory manufacturers have little economic incentive to prioritize cheaper consumer electronics when they can lock in massive contracts from the world's wealthiest companies. The era of ever-cheaper, ever-more-powerful gadgets may be over for the foreseeable future.

Source: The Verge News