CONNECT WITH US
Sep 1
Nvidia's MediaTek investment signals broader push to shape AI infrastructure
Nvidia's US$3.5 billion investment in MediaTek's overseas convertible bonds marks more than a financial bet. It underscores how the chip leader is trying to anchor its AI infrastructure strategy as cloud giants and model developers expand their use of custom silicon and rack-scale systems.
Dell Technologies closed its fiscal second quarter with revenue of US$47.0 billion, up 58% year over year, but the more telling number is what happened below the top line: operating income more than tripled, and the operating margin nearly doubled to 11.5% from 6.0% a year earlier. AI infrastructure did not just add revenue this quarter — it pulled the entire company's profitability up with it.
Apple has unveiled new Mac mini and Mac Studio models, equipped respectively with its first 2-nanometer chip, the M6, and the four-die M5 Ultra. Most notably, the latest 2-nanometer process is used only in the entry-level Mac mini, while the 3-nanometer M5 Ultra is officially billed as the "most powerful M-series chip yet."
Qisda has been expanding its display, commercial, and industrial businesses beyond traditional desktop monitors into medical, rugged, automotive, unmanned vehicle, drone, and low-Earth-orbit (LEO) satellite applications. Its Commercial and Industrial Group (CIG) has now made physical AI a core priority. CIG aims to extend Qisda's products from underwater, across land, and into the sky and outer space, building a full physical AI ecosystem.
Gudeng Precision Industrial chairman Bill Chiu said on August 31 that the company has begun building out its global manufacturing footprint. Its Japan plant is expected to complete further expansion in the second half of 2027 and gradually enter mass production, while the company has begun using an existing leased facility in the US for basic processing and manufacturing preparations.
Taiwan's limited land and heavy dependence on imported energy have made power a scarce and strategic resource in the compute era, putting AI data center (AIDC) site selection on the radar of industry and government. Taoyuan's Department of Economic Development chief Chang Cheng said on August 31 that the city is positioning itself as a northern AIDC demonstration base, with Tatan Power Plant and CPC Corp.'s Third Liquefied Natural Gas Receiving Terminal serving as key advantages.
Japanese electronic component shipments rose in the first half of 2026 as demand from AI servers drove strong growth in capacitors and other parts. According to Nikkei and Dempa, the Japan Electronics and Information Technology Industries Association (JEITA) said on August 31 that global shipments by about 50 member companies increased 4% from a year earlier to JPY2.2865 trillion (US$15.45 billion) in the first six months of 2026.
AI's early-stage growth is creating a short-term squeeze on land, talent, capital, and power in Taiwan, but there is no such thing as a world that cannot do without Taiwan, ASE Technology CEO and SEMI Global Board Executive Committee Chair Tien Wu said. He made the remarks at a pre-show conference on August 31 ahead of SEMICON Taiwan 2026, which runs from September 2 to 4.
South Korea's copper-clad laminate, or CCL, exports jumped sharply in the first seven months of 2026 as AI infrastructure investment and larger semiconductor packaging lifted demand, boosting revenue at suppliers including Doosan Electronic BG, LG Chem and Lotte Energy Materials.
Speaking at the Silicon Photonics Global Summit ahead of Semicon Taiwan 2026, Marvell Technology CTO Radha Nagarajan highlighted that the core focus for upgrading connectivity in the AI data center era has shifted firmly to power consumption. He noted that the acceleration of co-packaged optics (CPO) adoption is primarily driven by two critical bottlenecks: power and unit density.
Tmytek is preparing to list on the Taiwan Innovation Board as its millimeter-wave platforms approach commercial scale. For global telecom, satellite, and defense customers, the timing may matter because it could shape supply availability, manufacturing capacity, and the speed at which next-generation connectivity reaches the market.
DIGITIMES observed that as AI systems continued to scale, interconnect requirements inside AI data centers had expanded from chips, packages, and boards to racks, clusters, and eventually inter-data-center connections.