AI is making your gadgets more expensive and less powerful

As prominent tech firms prepare to unveil their latest hardware, a concerning trend emerges: gadgets are becoming both more expensive and less powerful. Apple's newest Mac minis and Mac Studios, powered by its cutting-edge M6 chips, are set to launch, followed by the iPhone 18 Pro, Pro Max, Apple Watch upgrades, and foldable iPhone at the September company event.

However, this era marks a shift where consumer hardware is getting worse instead of better. Nvidia's upcoming Vera CPU is a prime example, with JPMorgan reporting a reduction in SOCAMM memory from 1.5TB to 768GB and HBM memory in Rubin Ultra systems, decreasing from 16-high HBM4E stacks to 8-high or 12-high versions. These adjustments are seen as "content optimization" due to constrained memory supply, rather than a decrease in required memory.

According to Runar Bjorhovde, a senior analyst at Omdia, this scale of optimization has never been witnessed before. While innovation hasn't ceased and Moore's law still applies, high demand for memory and advanced components in AI infrastructure has driven prices up. Memory costs have quadrupled in the past year, accounting for more than half of the bill of materials in some cheaper devices, up from 10% to 15% previously.

Manufacturers are responding by trimming components, opting for 4G phones, reduced storage, lower-quality cameras, and older processor generations. High-end laptops remain unaffected, but at lower price points, consumers can expect less powerful devices. Samsung's Galaxy A16 4G, an inexpensive phone, remains its best-seller in Europe, while Xiaomi's revenue fell slightly despite a 25% drop in unit shipments.

This trend is expected to continue for years, with analysts predicting no significant price drops within the next 18 months and a structural shift lasting until at least mid-2028.