Data from a16z points to a Jevons paradox: token prices keep falling while H100 GPU rental prices hold steady or rise, as cheaper AI drives demand faster than costs drop. If that demand levels off, the whole chain from chipmakers to cloud providers would be hit.
ZDNet reviews VZmore's AX9 Max, a compact PC aimed at people who need strong performance for running AI tasks locally but have little space.
In an episode of TechCrunch's Equity podcast, the hosts discuss what moratoriums on data center construction could mean. They also ask whether the AI industry is struggling to communicate, including whether dropping non-disclosure agreements could win over opponents.
The US International Trade Commission has opened a patent case against 20 companies, including Foxconn, Luxshare and Quanta, over power systems in AI servers. It follows a complaint from power-chip maker Vicor, which says their products infringe one of its patents.
Google has launched an experimental satellite carrying its own AI chips. The long-term aim is to assemble solar-powered constellations of computing centers in orbit.
An Apple research paper describes Minions, a system that pools the mostly idle GPUs of Apple Silicon Macs into a distributed cloud for machine learning inference. The idea targets the cost, rigidity and waiting times of centralized GPU infrastructure.