If your next phone costs more than your last one, or ships with less memory than you expected, the reason is not in the phone. It is in a data center several states away. J.P. Morgan Global Research estimates that DRAM prices will have risen more than 400 percent from the start of 2024 to the end of 2026, and the cause is not a factory fire or a trade war. It is a deliberate choice by the three companies that make nearly all of the world’s memory to sell it to someone who pays more.
That someone is the AI industry. And the consumer electronics market is finding out what happens when it stops being the customer that matters.
A shortage nobody is short of capacity for
Most chip shortages are supply accidents. The pandemic shortage came from shuttered factories and scrambled logistics. This one is different in a way that matters for how long it lasts: the factories are running. They are just making something else.
Samsung, SK Hynix, and Micron build both the ordinary DRAM that goes into phones and laptops and the high-bandwidth memory, or HBM, stacked next to AI accelerators in data centers. They run on overlapping production lines, and HBM is far more lucrative. So, as IDC describes it, the major memory makers shifted production toward AI data-center memory instead of expanding the conventional chips consumer devices need. IDC expects 2026 supply growth for DRAM and NAND to run well below historical norms, at around 16 and 17 percent.
The scale of the reallocation is the striking part. Citing IDC research, analysts now forecast that data centers will consume roughly 70 percent of all memory chips produced worldwide in 2026, up from something like 20 to 30 percent as recently as 2022. In four years, consumer electronics went from the main event to the leftovers.
Follow the price per bit
The incentive is easy to see in the analyst numbers. Consensus estimates compiled by Visible Alpha project Samsung’s revenue per bit of conventional DRAM rising 116 percent year over year, with SK Hynix up 78 percent and Micron up 54 percent. Read that carefully: prices on the ordinary memory are soaring precisely because nobody is prioritizing it. The scarcity is a byproduct of chasing the premium product, and it is pure upside for the chipmakers on both sides of the line.
None of this is a scandal. It is exactly what a profit-seeking manufacturer should do when one customer class will pay dramatically more than another. But it does mean the usual reassurance — that shortages correct themselves when high prices pull new supply into the market — runs slower here. New supply takes years of fab construction, and when it arrives, the same incentive decides who gets it first. IEEE Spectrum notes that the biggest AI hardware companies are likely first in line for new memory, leaving phone and PC makers to take what is left.
Where it lands: your phone
Phone makers have three levers, and they are pulling all of them. They can raise prices, cut specifications, or absorb the cost and accept thinner margins. IDC says it plainly: as memory costs surge, manufacturers will likely have to raise prices significantly, cut specs, or both, with Android manufacturers most exposed.
The damage was visible early. Back in December, Counterpoint Research raised its forecast for the average selling price of smartphones to a 6.9 percent jump in 2026, nearly double its prior estimate, while flipping its shipment outlook from flat to a decline. Forecasts have generally moved in one direction since.
And the pain is not evenly spread. Apple and Samsung buy memory in enormous volumes under long-term contracts and have premium margins to absorb shocks. The companies with the least room are the ones selling $200 Android phones in price-sensitive markets, where memory is a large share of the bill of materials and a price increase directly costs sales. The shortage is, in effect, regressive: it hits hardest at the bottom of the market, among buyers who can least afford an upgrade.
The irony in your pocket
Here is the part that should make phone makers wince. The single biggest selling point in the smartphone industry right now is on-device AI — assistants, summarization, photo generation, all of it pitched as the reason to upgrade. And AI features running locally are memory-hungry. They need more RAM, not less.
So the industry is in a genuinely awkward position: the AI boom is making memory expensive at the exact moment the AI marketing pitch requires phones to carry more of it. A manufacturer trimming RAM to hold a price point is quietly trimming the features it is spending its advertising budget promoting. Expect to see this tension surface as vague asterisks — AI features “available on select models,” or reserved for the higher storage tiers where the margin can cover the memory.
What to actually do
For buyers, the practical advice is unusually clear for once. If your current phone works, this is a reasonable year to keep it; a structural shortage rewards patience more than a cyclical one does. If you need to buy, pay attention to RAM specifications rather than assuming they will hold steady from last year’s model, because this is precisely the spec manufacturers are tempted to shave. And treat any promised AI feature as something to verify on the specific model and configuration you are buying, not as a brand-wide guarantee.
The larger lesson is about where the technology industry’s center of gravity now sits. For two decades, consumer devices were the demand that set the terms for the entire semiconductor supply chain. That era is ending. The memory in your pocket is now the product chipmakers sell with whatever capacity AI does not want — and until that changes, you are paying the difference.
Related reading: Is AI a Bubble or a Revolution? The Honest Answer Is Both · AI’s Energy Use Is Both Wildly Overblown and Seriously Underrated
