AI Eats Your Laptop
Apple raised prices on MacBooks, iPads, and the Apple TV this week. The increases range from $30 to $300. A 14-inch MacBook Pro that cost $1,700 now sells for $2,000. The iPad Air went from $600 to $750. The Apple TV streaming box jumped from $130 to $200.[1]
The reason, according to Apple, is memory and storage. AI data centers are gobbling up chips so fast that component prices have surged at least 50% per quarter since late 2025. An Apple spokesperson called it a component price increase unlike anything the company has seen before. Tim Cook, in his final months as CEO, told The Wall Street Journal last week that price hikes were "unavoidable" and described the spike as a "hundred-year flood."[2]
The iPhone, Apple's main revenue engine, was spared for now. But the message is clear: the AI infrastructure boom is not just burning through GPU shipments and data center budgets. It is leaking into the supply chain for every electronic device that uses memory or storage, which is essentially all of them.
This is the part of the AI cost curve that gets less attention than trillion-dollar valuations and mega-training runs. The secondary effects. NAND flash and DRAM were already tight markets with limited producers. When hyperscalers place orders that dwarf consumer electronics demand, the ripples reach everyone. You do not need to use ChatGPT to pay the ChatGPT tax. You just need a laptop.
Apple noted $416 billion in revenue last fiscal year and said it had "shielded our customers from these increases so far." That shield has now dropped. Whether competitors follow is almost not a question. Dell, HP, Lenovo, and every other PC manufacturer buy from the same memory suppliers. If Samsung, SK Hynix, and Micron are charging Apple more, they are charging everyone more.
The timing is awkward. PC sales had just started recovering from a post-pandemic slump. Chromebook refresh cycles, Windows 11 migration, and the first wave of "AI PCs" with NPUs were supposed to drive upgrades. Instead, the AI boom that created the NPU marketing angle is also making the machines more expensive to build.
Cook's successor, John Ternus, takes over on September 1, days before the next iPhone unveiling. He inherits a pricing problem that no amount of supply chain wizardry will solve quickly. Memory prices are set by quarterly contracts and demand forecasts, not by executive preference.
For anyone building or buying hardware right now, the practical takeaway is simple. If you were already planning a laptop or tablet purchase, the price you see today may be better than the price you see in six months. The AI data center buildout is not slowing down. Neither, apparently, is the bill for it.