Tau Law

Tau Law is the semiconductor and organization idea evaluated in 当华为抛出韬定律，我们该信它到哪一步？. The source describes it as Huawei’s attempt to reframe chip progress around tau, signal delay, and end-to-end system speed rather than only around smaller process nodes.

The hosts’ main position is cautious: Tau Law should not be read as a new natural law that replaces Moore’s Law. It is more defensible as a KPI-like engineering metric that can align devices, process, circuits, architecture, systems, and software around lower latency. Its credibility depends on whether Huawei can produce measurable, repeatable gains in performance, cost, and energy efficiency.

华为的「韬定律」，是创新还是噱头？｜ Bonus sharpens the implementation boundary through Zhang Haijun / 张海军’s explanation of Cell-to-Cell Logic Stacking. The bonus episode treats the time metric as meaningful and not merely marketing, but says full logic-cell stacking would need new EDA flows and public product proof; Huawei’s “381 chips” claim is therefore broader methodology evidence, not proof that the hardest cell-level route has shipped.

真正改变世界的技术，为什么一开始都不被看好？| S10E16 adds the historical analogy to Moore’s Law. 汪波 frames both law-like claims as industry-mobilizing predictions under uncertainty: early Moore’s Law faced yield, heat, and tooling doubts, while Tau Law faces similar questions around System-Level Semiconductor Optimization, EDA, thermal behavior, yield, and product validation.

Key Claims

• Tau is treated as a time-constant and delay-oriented metric: smaller tau means faster switching and shorter effective communication paths.
• The episode frames the first move as changing the goal from “smaller” to “faster.”
• Semiconductor 3D Stacking and “logic folding” are the technical routes named in the source, but the hosts say the details of logic folding remain unclear from public information.
• The What’s Next source narrows “logic folding” into Cell-to-Cell Logic Stacking, distinguishing it from die-to-die stacking such as HBM, CoWoS-style packaging, or cache stacking.
• The same source treats EDA readiness as a gating test for whether Tau Law can become a practical design flow rather than only a system-performance frame.
• Tau Law has an internal organization role: it can give HiSilicon and adjacent teams a shared measure across device, circuit, architecture, system, and software layers.
• The public word “law” is treated as a communication and mobilization choice, not proof that the idea has the status of physics.
• To become an industry law, it would need the kind of long-term, measurable, ecosystem-coordinating proof that Moore’s Law once provided.
• S10E16 strengthens the analogy to Moore’s Law but keeps the same burden of proof: mobilization value does not replace chip, tool, cost, and power evidence.

Connections

• Huawei — origin and main strategic context.
• HiSilicon — chip-design capability that would have to help implement the metric.
• Ren Zhengfei and Huawei Organizational Methodology — organization background for why a KPI-like technical doctrine fits Huawei.
• Semiconductor 3D Stacking — key technical route for reducing signal distance and latency.
• Cell-to-Cell Logic Stacking and Electronic Design Automation — What’s Next’s toolchain-specific implementation boundary.
• Constraint Driven Engineering Strategy — strategic reason Huawei would emphasize a route beyond pure lithography scaling.
• Large Company Organizational Inertia — contrast case: tau-law framing attempts to coordinate a large organization around one metric rather than let scale fragment effort.
• System-Level Semiconductor Optimization and Moore’s Law — S10E16’s historical comparison and post-Moore engineering extension.