MA-tek, a semiconductor testing and analysis company, reported August revenue of NT$565 million, up 17.78% from a year earlier and marking a record high for the sixth straight month. Revenue for the first eight months reached NT$4.196 billion, up 17.66% from the same period in 2025.
This week, DIGITIMES heads to Shenzhen to cover an expo showcasing the world of optoelectronics, a class of technologies that serves as the link between electricity and light. In today's era of pervasive digital connectivity, from smart glasses to cars to phones, this event is expected to give us a glimpse into the hardware supply chain behind these devices, and increasingly behind the global AI infrastructure build-out.
As AI server computing power and per-rack power density continue to rise, traditional 54VDC power architectures are hitting bottlenecks such as excessive current, higher copper usage, and greater transmission losses. This is pushing next-generation AI data center power designs toward 800V high-voltage direct current (HVDC).
The US-China space race is broadening beyond rocket launches and crewed lunar missions into a growing array of strategic arenas, including lunar resources, low-Earth-orbit (LEO) satellite constellations, space computing and military capabilities in orbit.
This year's SEMICON Taiwan saw companies worldwide flock to Taipei to learn new industry developments and score potential business opportunities. Many came from elsewhere in Asia, but some came from as far as California, where business leaders highlighted the complementary relationship between the state's research and start-up ecosystem and Taiwan's manufacturing capabilities.
STMicroelectronics said humanoid robots are still most likely to enter manufacturing first, as the industry works through key hurdles in hand dexterity, edge AI decision-making, and total system cost. The chipmaker made the remarks at a recent press event in Taiwan focused on sustainability, AI data centers, and robotics.
DeepSeek is reportedly planning to purchase 160,000 Huawei chips for its large 1 GW data center currently under construction in the Chinese province of Inner Mongolia. This would make it among the largest clusters of Huawei chips deployed to date, as the Chinese authorities attempt to wean the country off Nvidia chips, but still struggle with insufficient domestic manufacturing capacity.
HyperVault, a Tata Consultancy Services (TCS) subsidiary, has secured 264 acres in Hyderabad to build a large AI data center campus capable of reaching 1GW of capacity. The project could strengthen India's role in global artificial intelligence infrastructure, while influencing jobs, energy demand, and digital supply chains worldwide.
Foxconn said August sales rose sharply from a year earlier as demand for AI servers and related hardware continued to lift its cloud and networking business. The latest update suggests the global electronics supply chain remains supported, even as investors closely watch shifts in trade, policy, and consumer demand.
As AI data centers scale rapidly, Eaton is expanding beyond traditional power management into modular power deployment, next-generation DC conversion and liquid cooling, positioning itself to address infrastructure requirements from the electrical grid to AI chips.
Taiwanese automotive microcontroller (MCU) maker SiliconAuto showcased its products at SEMICON Taiwan 2026, with CEO Gene Liu saying the company is now spotlighting AI chiplet architecture for automotive and industrial edge AI as carmakers seek more flexible chip designs.
As AI computing power surges, the energy consumption of AI servers has become a major bottleneck that must be solved. The issue is spreading from chips into power delivery systems as Nvidia's AI racks ramp up, rapidly pushing up the power draw of a single AI rack.
I-Chiun Precision Industry is accelerating its shift from LED lead frames into high-end thermal solutions, with thermal products now accounting for 40-45% of revenue. The company expects the share to exceed 50% as early as the second half of 2026 or in 2027 and to rise above 60% over the medium to long term.
I-Chiun Precision Industry and Taiwan's Industrial Technology Research Institute (ITRI) have established Huizhi Advanced, the country's first startup focused on AI chip thermal design, targeting a global liquid-cooling market expected to reach US$19.2 billion by 2030.
AI and high-performance computing (HPC) demand is changing the semiconductor industry's technology and investment priorities. As chip power consumption, computing density, and data transfer requirements rise rapidly, competition is no longer limited to front-end process scaling; 3D stacking, advanced packaging, high-density interconnects, thermal management, and memory technologies are also becoming more important.
Quantum computing has a trust problem. When a quantum machine produces an answer no ordinary computer can feasibly produce, there is no obvious way to tell whether the answer is right.
System integration remains a key hurdle as the AI chip industry pushes for higher power efficiency and faster transmission speeds, drawing intense attention to when short-reach optical communications between racks and chips will enter practical use. Industry players say the maturity of system integration is still the main consideration for commercialization.
Lite-On Technology announced a strategic investment in Polish liquid cooling solutions provider DCX Polska, also known as DCX Liquid Cooling Systems. The deal will give Lite-On approximately a 25% stake in DCX upon completion, with a total transaction value of roughly US$176 million.
Taiwan is racing to expand its AI compute capacity as demand for AI chips, GPUs and AI servers surges, with compute infrastructure now central to national competitiveness. According to data provided by the National Science and Technology Council (NSTC), the government-backed mainframe Nano 5 is already fully loaded, while the National Institutes of Applied Research (NIAR) will launch three major investment projects.
Nvidia used IFA 2026 to argue that serious AI work can run locally as well as in the cloud, pairing an October launch date for its RTX Spark Windows PCs with free software that distributes independent inference requests across compatible machines already on a local network.
Nvidia's agreement to buy Hugging Face for roughly US$13 billion is being framed by analysts less as a move into software than as insurance against three ways its core business could deteriorate: cheaper open-weight models eroding the frontier labs that buy its chips, those same labs diversifying their silicon, and enterprises pulling workloads out of hyperscaler bundles. On each of those paths, owning the industry's main model repository pays off.
The mechanics of Nvidia's US$12.93 billion purchase of Hugging Face say as much about the deal's purpose as the strategic case does: a tenth of the consideration is set aside to stop the team from walking, part of the payout goes to Intel and AMD, and the whole transaction is framed against an open-weight model market in which the fastest-moving suppliers are Chinese.
Nvidia's agreement to acquire Hugging Face for US$12.93 billion pushes the world's dominant AI chip supplier past silicon and into the layer where developers actually pick their models — a position that gives it early sight of demand shifts, a structural hedge against customers designing their own accelerators, and a neutrality problem it has already had to answer for in writing.
Taiwan's National Science and Technology Council (NSTC) has budgeted NT$3.76 billion (US$116.58 million) for a quantum computer system project, while the Industrial Technology Research Institute (ITRI) will spend NT$14.6 billion with support from the Ministry of Economic Affairs (MOEA) on a 2027 technology program that includes building a key quantum computing technology verification platform.
