Morgan Stanley: The Tau Law Supports Exponential Growth in Demand for AI Optical Transceivers

Overview

Morgan Stanley released a commentary on the tech hardware industry in Greater China, focusing primarily on Huawei's "t (Tau) Law" unveiled at ISCAS 2026. The research report argues that this law provides a solid theoretical foundation for the exponential growth of the AI optical transceiver industry and, based on this, reaffirms a positive outlook on the industry. The report points out that as geometric scaling faces physical limits, the industry's focus is shifting from simply reducing the size of individual chips to optimizing system-level efficiency, which includes interconnects, thermal management, and signal integrity.

Core views

The core logic of the research report is grounded in Huawei’s T (Tau) Law, which defines a time-scaling principle for multi‑layer electronic systems. By optimizing a unified time constant (t), this law seeks to address the computational‑power bottleneck that has emerged as traditional Moore’s Law encounters physical limits at advanced geometries. Unlike previous approaches that focused solely on transistor scaling, the T Law emphasizes system‑wide coordination across all computational layers—from individual transistors to data centers—prioritizing interconnects, signal integrity, power efficiency, and thermal management. This paradigm shift is viewed as the underlying driver behind the sustained surge in demand for AI optical transceivers, as more efficient optical interconnects are critical to achieving system‑level time optimization. On the technical implementation front, the report highlights an innovative “logic folding” approach based on the T Law. By vertically stacking the active layers of digital, analog, and memory circuits, this technique delivers a 55% increase in transistor density and a 41% improvement in energy efficiency—without relying on further advances in process nodes. This demonstrates that, even amid slowing process‑node scaling, architectural innovation can still yield substantial performance gains, directly benefiting the associated hardware supply chain. Looking ahead, the report notes that the T Law, combined with a unified memory‑semantic bus interconnect architecture, near‑package optics such as Hi‑ONE (high‑speed optoelectronic interconnects within the package), and edge‑to‑face 3D folding technologies, will fundamentally reshape the hardware landscape of AI clusters. The firm projects that, driven by these advancements, AI cluster hardware integration could grow by more than 100-fold by 2035. This long‑term quantitative target underscores the significant growth potential over the next decade for AI optical transceivers and related advanced packaging and interconnect technologies.

Analysis framework

The research report employs an analytical framework that links industry trends to theoretical paradigms. Rather than focusing solely on short-term order or shipment data, it begins with the generational shifts in underlying technological theory to substantiate the sustainability of sector growth. The firm first identifies the key challenges confronting the industry following the apparent end of Moore’s Law, then introduces Huawei’s “t‑law” as a new explanatory variable. By examining the specific technological pathways outlined by this theory—such as logic folding and Hi‑ONE—it derives the rigid demand for certain hardware components, notably optical transceivers, and ultimately constructs long‑term quantitative forecasts. This first‑principles approach enables investors to discern structural opportunities that transcend short‑term cyclical fluctuations.

Methodology notes

• Industry/Industrial Analysis FrameworkOthers

  t (Tau) Law / System-Level Time Scaling Principle

  This is the core technical analysis framework cited in the research report, which posits that, as Moore’s Law approaches its geometric scaling limits, optimizing a unified time constant (signal delay) across hierarchical layers can enhance overall system efficiency. In this paper, it serves as a theoretical anchor for projecting the long-term demand trajectory of AI optical transceivers, replacing the conventional approach of tracking process nodes.

• Industry/Industrial Analysis FrameworkPenetration Rate S-Curve

  The Substitution of New Technology Paradigms for Old Paradigms and Long-Term Growth Forecasts

  The research report positions T‑Law as the successor to Moore’s Law and sets a long-term target of achieving a 100‑fold increase in hardware integration by 2035. This reflects the underlying logic of the technology S‑curve: as legacy technologies—such as geometric scaling—reach their plateau, emerging innovations—like system‑level optimization—usher in a new phase of exponential growth, enabling investors to pinpoint the current stage of industry development.

Key data

• Logic Folded Transistor Density Enhancement55%Under fixed process nodes, density gain is achieved through the vertical stacking of active layers.
• Logical folding for energy efficiency enhancement41%Improvement in Power Efficiency Under the Same Process Node
• Expected increase in hardware integration by 2035>100xLong-term forecasts of AI cluster integration based on Moore’s Law and its supporting interconnect technologies

Impact & implications

The research report argues that the establishment of T-law signifies a shift in the rules of the AI computing power race—from simply chasing advanced process technologies toward system-level architectural innovation. For the industrial chain, this reinforces the strategic importance of segments such as optical interconnects, advanced packaging, and heterogeneous integration. In particular, technologies like Co-Packaged Optics (CPO/Hi-ONE) are no longer merely optional performance enhancements but have become essential components for achieving system efficiency under the new paradigm. Although the industry rating remains "In-Line," the renewed positive emphasis on the AI optical transceiver sub-sector suggests that this segment possesses growth potential exceeding the industry average over the long term.