AI is making your gadgets more expensive and less powerful

We’re fast approaching the release of new hardware products from many of the world’s biggest tech companies. Apple just announced new Mac minis and Mac Studios powered by its latest M6 chips, while we’re expecting the iPhone 18 Pro and Pro Max, Apple’s first foldable iPhone, and new Apple Watch models to be announced at the company’s annual iPhone event in the first half of September. While the Mac minis and Studios have high-quality, cutting-edge tech powering them, we’re entering a strange new world. Consumer hardware, and the technology powering the tools we rely on more and more, is starting to get worse rather than better. One example can be seen in Nvidia’s forthcoming Vera CPU. JPMorgan says Nvidia has halved the planned amount of SOCAMM memory attached to each chip, from 1.5TB to 768GB. It has also reduced the planned HBM memory used in its Rubin Ultra systems, moving from 16-high HBM4E stacks to 8-high or 12-high versions. The bank describes those changes as “content optimization as a coping mechanism” for constrained memory supply, rather than a sign that the systems suddenly need less memory. “To this scale, we’ve never seen anything like it before,” says Runar Bjorhovd…