Huawei has introduced the "Tau Law," an alternative approach to Moore's Law that shifts focus from transistor miniaturization to reducing signal propagation delay through three-dimensional stacking and circuit folding. While the company demonstrates architectural innovation in semiconductor technology, analysts remain skeptical about its ability to close the gap with leading competitors. Despite these efforts, Huawei faces persistent equipment shortages and supply chain constraints that limit near-term progress. Industry assessments suggest the company could still trail international competitors by six to eight years by 2031, though Huawei claims its new approach could increase chip density by fifty-five percent and achieve comparable performance levels to advanced processes.

Huawei and Intel have fundamentally shifted the semiconductor competition from traditional transistor scaling to comprehensive systems optimization. Huawei's Tau Scaling Law and LogicFolding architecture prioritize signal path efficiency across multiple hierarchical levels, with mass production of 381 chips and consumer products planned for late 2026. Intel's 18A-P node demonstrates continued lithographic relevance through advanced power delivery, new transistor options, and thermal management. Rather than a geopolitical rivalry, this represents a strategic disagreement about how to measure semiconductor progress. Both companies acknowledge that future advancement depends on integrating density improvements with data movement, power efficiency, heat dissipation, and system-level co-optimization—transforming chip development from isolated transistor innovation into holistic platform engineering.

Huawei's Tau Law proposes using time compression and LogicFolding architectures to sustain semiconductor performance gains.

Les satellites artificiels vont ils être trop nombreux et nous poser plus de problèmes qu’ils nous apportent de services ? Actuellement, les lancements de fusées relarguent dans l’atmosphère près 1 000 tonnes de suie par an. Dans la stratosphère (20-50km d’altitude) cette suie, constituée de particule de carbone aussi appelées carbone noir absorbe l’énergie solaire et provoque une augmentation des températures. D’après les recherches menées en 2022 par la National Oceanographic and Atmospheric Administration américaine (NOAA) le décuplement des lancements de fusées est attendu au cours des 20 prochaines années injecterait dans la stratosphère assez de carbone noir pour accroître les températures stratosphériques mondiales de 0,5 à 2 °C.

China Telecom is testing a new approach to packaging artificial intelligence services alongside its network offerings, reflecting a broader shift in how telecom operators are experimenting with AI-driven infrastructure and monetisation models. MEF CEO Dario Betti explores the implications for operators and how this development may reshape telecom business models.

LEGO a construit une Koenigsegg Sadair's Spear grandeur nature en briques qui a battu un record à 111 km/h à Goodwood.

Huawei's newly proposed "Tao Law" semiconductor technology has generated significant discussion regarding its potential to challenge TSMC's technological dominance. The approach leverages three-dimensional stacking, advanced packaging, and architectural integration to enhance chip performance and transistor density without requiring extreme ultraviolet lithography equipment. While Huawei claims the technology could achieve computing capabilities equivalent to 1.4-nanometer processes, NVIDIA CEO Jensen Huang contends the innovation, though noteworthy, poses no threat to TSMC. Huang emphasized that TSMC has invested nearly a decade in developing comparable die stacking and hybrid bonding technologies, maintaining a substantial technological advantage. TSMC's advanced packaging capabilities, including CoWoS and SoIC platforms, continue supporting leading-edge AI and high-performance computing applications from major manufacturers, reinforcing its market position.

This research addresses absorptive capacity in maintaining technological advantage amidst Emerging and Disruptive Technologies (EDTs). While traditional models focus on incremental integration, this study introduces the construct of transformative capability to account for the systemic reconfiguration necessitated by Artificial Intelligence and Big Data as two out of nine EDTs. Utilizing a structured policy research methodology, the study evaluates a purposive sample of 56 NATO policy documents, triangulated with semi-structured stakeholder interviews to identify high-priority catalysts for adoption. Findings identify four interconnected factors—technological, human, organizational, and societal—functioning as a recursive, systemic feedback loop. A distinctive contribution is the identification of the psychological unimaginability of EDTs as a primary barrier to diffusion, necessitating cognitive anchors through robust standardization, ethical governance and social institutionalization. By positioning ethics and societal assimilation as functional prerequisites, this research offers a multidimensional framework for fostering double-loop learning and organizational resilience in 21st-century defense ecosystems. 

Huawei introduced the Tau (τ) Scaling Law at the 2026 IEEE International Symposium on Circuits and Systems, presenting a transformative approach to semiconductor evolution. As Moore's Law faces physical and economic limitations, this new principle replaces traditional geometric scaling with time-based scaling to guide future semiconductor development. Leveraging innovative technologies such as LogicFolding and multi-level co-optimization mechanisms spanning devices, circuits, chips, and systems, the τ Scaling Law systematically reduces time constants to enhance performance, energy efficiency, and transistor density. This approach addresses the semiconductor industry's critical challenge of overcoming physical constraints while meeting escalating computing demands, offering a sustainable path forward for continued technological advancement.

TSMC has been pioneering advanced 3D packaging, chiplet stacking, and hybrid bonding for nearly a decade. Huawei is simply applying these existing concepts ...

Huawei has announced an ambitious semiconductor roadmap targeting 1.4-nanometer chip design by 2031, as revealed by HiSilicon President He Tingbo at the 2026 IEEE ISCAS conference in Shanghai. To circumvent U.S. technology sanctions restricting access to advanced EUV lithography equipment, Huawei introduced the Tau Scaling Law, which prioritizes system-level efficiency and latency reduction over traditional geometric transistor scaling. The company's LogicFolding architecture optimizes internal interconnects and silicon stacking to achieve performance equivalent to advanced nodes without requiring cutting-edge lithography. This strategic pivot represents China's significant effort to maintain semiconductor independence while advancing chip capabilities under strict export restrictions.

Computational lithography, transistor simulation, process control and wafer inspection now require massive-scale simulation and real-time optimization, and AI ...

On May 25, 2026, Huawei presented LogicFolding, a novel 3D chip architecture designed to circumvent dependence on EUV lithography technology. Unveiled by semiconductor division president He Tingbo at the IEEE ISCAS conference, this approach claims 55% superior transistor density compared to conventional planar designs through temporal scaling rather than geometric scaling. The technology targets equivalence to 1.4-nanometer processes by 2031 via architectural innovation rather than advanced etching. Huawei expects first Kirin chips featuring LogicFolding by fall 2026, with plans to extend the technology to Ascend AI accelerators by 2030. While the company claims six years of development and mass production of 381 chips, these claims remain unverified. The announcement triggered a 7.6% stock surge for Chinese foundry SMIC, reflecting industry confidence in potential alternatives to restricted advanced manufacturing capabilities.

In 1932, American industrialist Eben Byers died after his jaw physically collapsed from drinking “Radithor”, an over-the-counter energy drink infused with raw radium. In the early 20th century, society was so mesmerized by the superficial “glow” of radiation that brands blindly put thorium and radium into cosmetics, toothpaste, and water. It wasn’t until 1938, when the lethal, cellular-level destruction of radiation became undeniable, that governments stepped in, stripped radioactive elements from consumer shelves, and locked nuclear energy behind strict regulatory walls.Today, we are repeating that exact historical error.

NVIDIA and TSMC are leveraging artificial intelligence and accelerated computing to revolutionize semiconductor manufacturing. NVIDIA's CUDA-X libraries, Metropolis platform, and TAO Toolkit are being deployed across TSMC's production operations to enhance lithography simulation, transistor modeling, process control, and wafer inspection. These advanced AI systems enable automated defect detection at the nanometer scale while reducing manual labeling requirements. By integrating AI throughout the semiconductor design and manufacturing lifecycle, TSMC achieves improved turnaround times, enhanced energy efficiency, increased yield rates, and optimized fab operations. This collaboration between the two technology leaders addresses complex computational challenges inherent in advancing chip production to increasingly sophisticated nodes.

NVIDIA and TSMC are leveraging accelerated computing and artificial intelligence to revolutionize semiconductor design and manufacturing processes. As chip technology advances to more complex nodes, the industry faces unprecedented computational challenges in areas including computational lithography, transistor simulation, process control, and wafer inspection. TSMC is deploying NVIDIA's AI and accelerated computing solutions across its fab operations to enhance turnaround times, energy efficiency, yield rates, and overall operational productivity. This strategic partnership, spanning nearly three decades, represents a significant advancement in addressing the world's most complex design and manufacturing challenges through simulation, optimization, and AI-driven systems capable of supporting physics-based and image analysis applications simultaneously.

Photonics represents a transformative frontier in advancing next-generation optical communication networks. This open-access perspective article, published in May 2026, explores emerging technologies and innovations that leverage photonic principles to enhance data transmission capabilities. The research examines how photonic systems can address current limitations in optical communications, enabling faster, more efficient information transfer. By integrating cutting-edge photonic solutions, these networks promise substantial improvements in bandwidth capacity and system reliability. The article provides valuable insights for professionals seeking to understand the technological evolution of communication infrastructure and its implications for future digital connectivity.

Quantum networks represent a transformative technological advancement that harnesses quantum mechanical principles to revolutionize data transmission and computational capacity. Unlike classical systems using binary bits, quantum networks employ quantum bits or qubits, which exploit superposition to exist in multiple states simultaneously, and leverage entanglement to create instantaneous connections across distances. By linking quantum computers and ultraprecise sensors, these networks enable geographically distributed systems to operate in concert, delivering unprecedented measurement accuracy and processing power. Applications span earthquake prediction, agricultural monitoring, autonomous GPS-free navigation for emergency responders in challenging environments, and accelerated drug discovery through molecular simulation. Quantum networks fundamentally transcend the limitations of conventional binary-based architectures, offering transformative potential for scientific discovery and practical problem-solving across multiple sectors.

Huawei has unveiled the Tau Scaling Law and LogicFolding architecture, representing China's ambitious response to US chip export restrictions. Rather than pursuing traditional transistor miniaturization governed by Moore's Law, this framework shifts focus to time scaling—compressing signal travel duration through chips and systems. Presented by Huawei's semiconductor president at the 2026 IEEE symposium, the approach operates across four integrated levels: device, circuit, chip, and system. LogicFolding architecture optimizes critical-path wiring and resistance while UnifiedBus reduces latency in large-scale AI deployments. Huawei claims this methodology will enable high-end chips achieving transistor density equivalent to advanced nodes, effectively circumventing lithography equipment constraints imposed by international restrictions.

C'est du jamais vu dans l'histoire de l'intelligence artificielle et un immense séisme pour la tech. Lancé lundi, désactivé jeudi soir : le gouvernement américain a ordonné à Anthropic de couper l'accès à Fable 5 et Mythos 5, ses deux modèles d'IA les plus avancés, pour l'ensemble de ses clients dans le monde entier, sur fond de faille de sécurité contestée.