Tim Cook's final months as Apple CEO are ending the way a lot of consumer-tech executives' tenures end: with a price increase. In an interview with the Wall Street Journal, Cook said that memory and storage chip costs have roughly quadrupled year-over-year, driven by AI hardware demand, and that passing those costs to buyers of iPhones, Macs, and iPads is now "unavoidable", even after Apple's efforts to absorb them.
The supply squeeze has a nickname in chip-industry circles: "RAMageddon." It describes the collision of two surges. AI server buildouts are soaking up the world's supply of DRAM (the working memory that lets a device juggle many tasks at once) and NAND flash (the storage that holds photos, apps, and files), pushing prices to roughly four times last year's levels. The chips that used to be a quiet line item in a phone bill of materials are now the most contested commodity in electronics.
Apple is the first household-name case where that bill is reaching the register. Cook did not name which products will see higher prices or when, and Apple declined to elaborate beyond the WSJ interview. The next iPhone is expected to launch in September, which is the most realistic window for any pass-through. TechInsights, the teardown and cost-modeling firm, estimated that to keep its current profit margin on the next iPhone Pro, Apple would need to add roughly $270 to the price tag. That figure is one analyst's margin-preservation math, not Apple guidance. The iPhone 17 Pro starts at $1,099.
What makes Apple's case unusual is that its own AI strategy is part of the pressure. The company paid a nine-figure sum earlier this year to settle a false-advertising lawsuit over a failed Apple Intelligence feature, an admission that its AI delivery has slipped. At WWDC earlier this month, Apple showed a more sober AI roadmap: more on-device processing, a Siri overhaul, and tighter integration across the device family. On-device AI is a privacy and latency win, and a marketing story. It is also a memory story. Running models locally means more DRAM and more flash in every device. The same RAMageddon cost curve that is squeezing Apple's bill of materials is, on Apple's own roadmap, the right answer to its AI ambitions.
The collision is sharp enough that Cook raised the alarm in April, after a record quarter, warning that the cost pressure would show up in the next set of results. The incoming CEO, John Ternus, echoed the warning in public remarks that same month. The transition, formally still ahead, has not changed the message.
The forward question is whether Apple can engineer its way around the squeeze. Three levers are visible. Silicon: Apple's in-house chips are optimized for memory bandwidth in ways commodity designs are not, and the next generation could squeeze more performance per dollar of DRAM. Model compression: smaller, more efficient on-device models reduce the memory footprint of the AI features Apple is promoting. Memory architecture: stacking, packaging, and on-die integration can lower the effective cost of a given capacity. Each lever is real, and none is free.
The cheaper scenario is that Apple eats the margin hit to defend the premium price point and signals discipline. The more uncomfortable scenario is a price hike at the September launch, framed by the company as the cost of staying on the AI roadmap. Cook's WSJ comments suggest Apple is no longer pretending that the third option, absorbing the increase forever, remains open. The register will tell the story next.