He Tingbo (何庭波) unveils Huawei’s (华为) “Taoding (韬, τ) law” and names internal “Moye (莫邪)” working group after seven years of hardship

The τ-law and a new architecture push

He Tingbo (何庭波), director and president of Huawei’s (华为) semiconductor business, on May 25 unveiled what she called the “Taoding (韬, τ) law” — a new design principle Huawei says will guide chip performance growth now that geometric transistor scaling has slowed. She framed the idea simply: if transistors cannot keep shrinking, then speed must be gained by cutting time — reducing waiting, transmission, synchronization and computation across transistor, circuit, chip and data‑center layers. “Think of turning a flat city into a three‑dimensional one and installing millions of elevators,” she said. It has been reported that Huawei claims this layered, “logical folding” approach has already yielded 381 designed and mass‑produced chips over the past six years and gives a roadmap to a 1.4nm‑equivalent transistor density by 2031.

Moye (莫邪): the secretive, large‑scale effort

He also confirmed what had been long whispered inside the company: the existence of an internal “Moye (莫邪)” working group created after the 2019 U.S. sanctions. It has been reported that He’s 2019 internal letter declared Huawei’s chip “backup” program and that the Moye team — described internally as a “small group” but reportedly numbering in the tens of thousands — carried the effort through seven years of hard work and self‑sacrifice. The name references an ancient sword‑forging legend; the analogy is deliberate, emphasizing long, difficult foundational work rather than quick wins.

Product claims and wider stakes

He said Huawei will ship a Kirin (麒麟) chip this autumn that fully implements logical folding. According to the figures released, the 2026 Kirin chip achieves a 53.5% jump in transistor density to 238 MTr/mm², a 41% improvement in P‑core energy efficiency and a peak frequency rise of 12.7% to about 3.1GHz; further roadmaps reportedly target 400+ MTr/mm² and 5.0GHz by 2031. Huawei also asserts these gains are achieved without depending on next‑generation lithography tools — a point of geopolitical significance given export controls and limited access to extreme ultraviolet (EUV) equipment.

Why this matters

For Western readers: Huawei is not just a smartphone maker but a pillar of China’s telecom and cloud infrastructure, and its semiconductor progress is closely watched as a barometer of how Chinese firms adapt to tighter U.S. controls on chip materials, tools and designs. If Huawei can materially advance performance through architecture and system‑level optimization rather than raw process node access, it changes the contest lines in the global semiconductor supply chain. But many of the most consequential claims remain company‑sourced or reported; independent verification will be needed before the industry can fully gauge the impact. Will other Chinese firms follow Huawei’s τ‑law path? That may determine whether this is an isolated engineering feat or the start of a broader strategic realignment.