355
Audio & Video Production344
Automation & Workflow224
Software Development250
Marketing & Growth192
AI Infrastructure & MLOps173
Writing & Content Creation203
Data & Analytics140
Design & Creative169
Customer Support130
Photography & Imaging156
Sales & Outreach125
Voice & Speech135
Operations & Admin87
Education & Learning131
Huawei is adjusting how it designs AI chips as chip-making improvements slow down, a shift that could affect the US lead in advanced chips.
In short: Huawei is adapting its AI chip plans as Moore’s Law, the long running trend of chips getting quickly faster and cheaper, slows down.
Moore’s Law is the idea that computer chips tend to improve rapidly over time, often by packing more tiny parts onto the same space. For decades, this helped phones, laptops, and servers get better every few years without big changes in how they were built.
According to WIRED, Huawei is adjusting its chip strategy because those easy gains are getting harder. Making chips smaller and denser has become more expensive and more difficult, and the improvements are not as predictable as they used to be.
This matters a lot for AI chips, which are specialized processors used to train and run AI systems. You can think of them like the engines in a factory that makes AI, faster engines mean more AI work can be done for the same cost and electricity.
The WIRED report says Huawei’s approach could complicate the current balance of power in chips, where US companies have held a strong lead in top tier AI hardware. It also highlights how companies may try to keep performance rising even when the usual path, shrinking the chip parts, is no longer enough.
Watch for more signs that chipmakers are chasing progress in other ways, like using new chip layouts, combining multiple chips into one package (like snapping Lego pieces together), or improving the software that controls the chips. Also watch how US rules that limit access to certain advanced chip tools and parts shape what Chinese companies build and how quickly they can scale.
Source: Wired