When AMD CEO Lisa Su arrived in Taiwan on May 20, she announced plans to invest more than US$10 billion with local supply-chain partners and the island's broader semiconductor ecosystem. The goal, she said, was to help secure a long-term supply of advanced artificial intelligence (AI) chips.
Facing expanding opportunities in both optical communications and satellite connectivity, Dennis Chen, chairman of WIN Semiconductors, said the company is actively advancing a range of products, including driver ICs, continuous-wave (CW) lasers, and photodiodes, while also ramping up capacity in anticipation of growing demand from next-generation networks.
Samsung Electronics is reportedly preparing to allocate much of its Pyeongtaek P4 cleanroom capacity to next-generation high-bandwidth memory (HBM) in 2027, a move that could tighten the supply of general-purpose DRAM as memory makers shift more production toward higher-value AI server products.
Intel is looking to broaden its foundry strategy beyond the race for advanced process nodes, placing greater emphasis on advanced packaging and glass substrate technologies as it positions Rio Rancho — its New Mexico site — as a global hub for next-generation packaging production.
Following GlobalWafers' shareholders' meeting on May 25, Chairperson Doris Hsu stated that the company's core compound semiconductor business, gallium nitride (GaN), is addressing strong demand for high-efficiency power solutions in AI servers. The company is also beginning to see emerging demand from diversified applications such as AI robotics. As a result, production capacity in 2026 has already entered a state of supply shortage. To meet strong demand from Japanese IDM customers, GlobalWafers is launching a continuous "30% plus 20%" expansion plan.
Huawei has released its Data Storage 2030 white paper, setting out a technology roadmap for the global storage industry over the next five to 10 years, as AI large language models drive data creation into what the company calls the yottabyte era.
Grace Wang, vice president and general manager of ASML in Taiwan, said that ASML plans to deploy extreme ultraviolet (EUV) and next-generation high numerical aperture (high-NA) EUV technologies to reduce energy consumption per wafer at the process level, while continuing to deepen its presence in Taiwan. The company also plans to hire around 1,000 new employees in Taiwan in 2026 to support customer expansion and rising global capacity demand.
As the global semiconductor industry approaches the physical limits of transistor scaling, Huawei has proposed a new framework for the post-Moore era through its recently introduced "Tau (τ) Law" and a related time-scaling theory.
GlobalWafers held its shareholders' meeting on May 25, where CEO Doris Hsu stated that the semiconductor market in 2026 has gradually moved beyond 2025's two extremes, when only AI and advanced process technologies dominated growth. In 2026, non-AI and traditional application markets began to recover, making market conditions thrive.
Heat management has become a critical challenge as HBM technology advances, with higher stacking and faster speeds to meet surging global demand for AI data processing. Efficient control of power density in the die-to-die physical layer between HBM and GPUs is now central to next-generation HBM competitiveness and data center reliability.
Following the conclusion of the Trump-Xi meeting and amid continued delays in China approving imports of Nvidia H20 GPUs, China's National Development and Reform Commission (NDRC) on May 22 sent a strong policy signal on artificial intelligence (AI) self-sufficiency, explicitly calling for greater efforts to pair domestic large language models with domestically developed computing chips.
More coverage