Much of Asia's manufacturing base is now subject to new US import duties of 10% to 12.5% after the White House finalized a Section 301 action against 60 economies, citing their failure to prohibit or effectively enforce a ban on goods made with forced labor. The presidential memorandum, dated July 23, 2026, took effect at 12:01 a.m. on July 24, according to the Los Angeles Times and the New York Post, and lands squarely on the region's textile, electronics, and assembly exporters.

AMD CEO Lisa Su said demand from cloud service providers and enterprise customers continued to climb rapidly, reinforcing AMD's view that the artificial intelligence (AI) market still has substantial room to grow. The CEO said AI adoption is now moving beyond efficiency gains and into new products, services, and business models.

AMD has formally introduced Helios, a next-generation rack-scale AI infrastructure platform designed for the fast-growing market of global AI factories and hyperscale data centers. The system is designed to boost compute density, lower deployment complexity, and offer an open alternative to competing rack-scale architectures for training and inference worldwide.

AMD unveiled its Instinct MI400 series at its Advancing AI 2026 conference, broadening its push into data center chips with products aimed at different users. For global readers, the move signals how AI infrastructure is becoming more specialized, with implications for cloud builders, governments, researchers, and supercomputing operators worldwide.

Taiwanese companies are strengthening operational technology cybersecurity as smart manufacturing and critical infrastructure become increasingly digital, with manufacturers generally further ahead than other industries because of international supply chain audit requirements.

Taiwan is turning to artificial intelligence, robotics and connected-care systems to manage one of the world's fastest demographic transitions, while seeking to turn the challenge into a new growth industry.

Wistron said its revenue and profit continue to rise, and its global manufacturing platform is opening new growth avenues for readers watching the electronics, AI, and industrial supply chains. The Taiwan-based company is pushing into higher-margin businesses while demand remains strong across key markets worldwide.

SemiAnalysis recently forecast that large-scale deployment of Nvidia's native single-ended 800V high-voltage direct current power architecture could slip from the market's previous expectation of 2027 to beyond 2028, raising concerns over growth prospects for the AI power supply chain.

A springtime source-code exposure reframed the AI race around "harness" and memory engineering rather than raw model scale — and a wave of Chinese models built around exactly those ideas has since arrived within striking distance of Anthropic's flagship.

Chinese officials have embraced Kimi K3 as evidence of indigenous innovation and a maturing open-model ecosystem. Some Chinese developers and technology commentators are far less convinced.

OpenAI has increased its projected spending on compute from US$600 billion to US$750 billion by 2030. The 25% increase is meant to give its models an edge during a time of intense competition for compute power, although some worry about that the start-up's investments are oversized.