In a rare interview, Huawei's top chip scientist, Liao Heng, warns of the looming physical limits Western chipmakers face and lauds Huawei's Tau Scaling Law
Huawei Technologies Co.'s top semiconductor scientist warned of the looming physical limits Western chipmaking giants including Nvidia Corp. face …
Context & Ripple Effects
Huawei’s chip strategy has been framed as a response to export restrictions that cut it off from advanced US technology. Its stated plan to pursue 1.4nm chips through LogicFolding and the Tau Scaling Law gave that strategy a named technical path.
The interview extends a competitive narrative already visible in Nvidia’s acknowledgment that Huawei and other challengers are significant rivals. It also arrives as China’s procurement incentives favor domestic alternatives to Nvidia hardware.
First-order effects
- Huawei gains a higher-profile platform for positioning Tau Scaling Law as an alternative to continued conventional transistor scaling, while Liao Heng directly challenges the durability of Nvidia-led chip roadmaps.
- Nvidia and other Western chipmakers face renewed scrutiny in China over whether their technology edge can remain decisive as physical scaling becomes harder.
Second-order effects
- Huawei’s customers and domestic partners have more reason to evaluate Ascend-based systems and locally sourced components alongside imported accelerators, reinforcing the procurement shift associated with China’s push for semiconductor independence.
- Competing AI-chip vendors will be pressed to differentiate not only on leading-edge process nodes, but also on packaging, system design, software support, and availability under export constraints.
Third-order effects
- If Huawei can translate its scaling claims into competitive products, AI hardware competition could shift from a race centered on the smallest process node toward alternative architectures and vertically integrated domestic supply chains.
- The longer-term constraint may become the ability to turn novel chip designs into dependable volume production and usable AI systems—an outcome that remains unproven by an interview or technical claim alone.
The trend: This is one data point in the AI hardware strategy split, where geopolitical access limits and slowing conventional scaling are pushing chipmakers toward distinct architectural and supply-chain paths.