17
Sep
Topview Optronics said shifts tied to artificial intelligence are helping it secure orders through the end of 2026, even as it continues to contend with memory shortages and rising component costs. The surveillance equipment maker said it is accelerating its adoption of AI and smart factory technology, and expects only slight full-year growth in 2026 compared with 2025.
17
Sep
CXMT cashes in on the DRAM shortage — and takes aim at HBM
ChangXin Memory Technologies (CXMT) is emerging from the memory shortage with sharply higher profits and greater pricing power, giving China’s largest DRAM maker more resources to expand into LPDDR6 and high-bandwidth memory (HBM) while China’s domestic memory supply chain moves further upstream.
17
Sep
TSMC reconsiders 12-inch equipment donation as Taiwan overhauls scope of research partnership
TSMC’s planned donation of 12-inch equipment is facing changes as the government adjusts the scope of a semiconductor collaboration project, while Taiwan’s National Science and Technology Council (NSTC) has not yet approved the related 2027 budget. NSTC Minister Cheng-Wen Wu said the partnership with Taiwan Semiconductor Research Institute (TSRI) and ITRI EOSL remains unchanged, although the location has shifted.
17
Sep
Memory LTAs lock up 50-70% of capacity
The memory industry remains in short supply, and major chipmakers are raising the share of capacity committed to long-term supply agreements (LTAs) while extending their duration from three to five years, with many in the industry expecting 2027 to be the tightest year yet. Supply chain sources said Kioxia has about 50% of its LTA capacity tied up, while South Korea’s two major vendors reserve about 70% of capacity for LTAs. Micron also has around 60% of its capacity locked into such agreements.
17
Sep
The Trump-Jensen speakerphone diplomacy: how Nvidia turned political pressure into a seat at the table
The relationship between Nvidia CEO Jensen Huang and US President Donald Trump has evolved from policy friction to a strategic alignment. The rapid shift in their dynamic reflects how deeply the AI sector has become intertwined with US-China technology rivalries and domestic manufacturing policies.
17
Sep
VLEO emerges as next orbital frontier as China warns of critical 2026–2030 window
Very low Earth orbit (VLEO) is moving from concept toward real-world deployment.
17
Sep
Humanoid robots could reshape MEMS demand as physical AI advances
Humanoid robots have not yet reached mass adoption, but their sensor needs are already reshaping supply-chain expectations worldwide. Analysts and industry players say the next wave of physical AI could lift demand for MEMS chips, edge computing, and advanced sensing, with broader implications for robotics makers, component suppliers, and global technology markets. Industry players say these systems need several times more sensors than conventional robots, creating a large opportunity for MEMS and sensor suppliers even if unit revenue per robot stays relatively modest.
17
Sep
Ardentec, Sigurd race into silicon photonics TIA testing as US client books capacity
The global optical communications module market continues to grow, and Taiwan’s test and packaging companies are moving to capture demand in silicon photonics (SiPh) transimpedance amplifier (TIA) testing as AI data-center networks shift toward faster optical interconnects. Ardentec and Sigurd are among the firms benefiting as a major US customer rushes to reserve capacity.
17
Sep
Tier IV open-sources self-driving AI chip design to challenge Nvidia
Japanese autonomous driving developer Tier IV is developing an AI chip and will release its chip design and related toolchain as open source, allowing semiconductor manufacturers to use them to develop autonomous driving chips. The company aims to break Nvidia’s dominant position in the autonomous driving chip market.
17
Sep
Huawei Connect 2026: OceanStor M900 turns PB-scale KV cache into Huawei’s next AI infrastructure layer
Huawei has unveiled the OceanStor M900 at Huawei Connect 2026, extending its AI infrastructure strategy into memory and storage with a petabyte-scale key-value (KV) cache system designed for AI agents and long-context inference.