CONNECT WITH US
Wednesday 19 November 2025
AI and chip demand drives Taiwan air freight rates higher
Taiwan's air cargo market is bracing for further increases in shipping prices as sustained demand for AI servers and semiconductor components drives record export volumes. Freight carriers expect rates to keep climbing through late 2025, reinforced by robust North America-bound shipments despite rising geopolitical and tariff-related uncertainties.
Wednesday 19 November 2025
RISC-V reaches data-center tier in China with StarFive's 'Lion Rock' breakthrough

China's AI server vendors are accelerating their shift toward self-developed architectures, bringing RISC-V from edge devices into data-center workloads. StarFive Technology has entered this space with the Lion Rock chip, its first RISC-V–based data-center management processor and one of the earliest large-scale commercial deployments of the ISA in China's server market.

Wednesday 19 November 2025
Power chip startups drive vertical power delivery trend
Silicon Valley venture firm Playground recently showcased seven semiconductor startups at a media event in Taiwan. Beyond the well-known Ayar Labs, d-Matrix, and NextSilicon, two HPC power component startups—PowerLattice and Vertical Semiconductor—stood out.
Wednesday 19 November 2025
DFI bets big on edge AI, predicts drone demand surge in 2026

DFI is sharpening its focus on edge AI and long-lifecycle industrial markets as global supply chains shift and AI deployments accelerate. At the company's investor briefing on November 17, DFI president Claire Tien said the Taiwanese industrial computer supplier plans to build its next phase of growth around manufacturing strength, localized production and long-term project stability.

Wednesday 19 November 2025
AMD eyes US$20B data center GPU revenue in 2 years, bets on rack-level MI400 growth
AMD CEO Lisa Su revealed at a recent Financial Analyst Day in New York that the company expects to generate about US$34 billion in revenue in 2025, with approximately US$16 billion coming from its data center segment.
Wednesday 19 November 2025
Taiwan suppliers will determine silicon photonics breakout
Silicon Valley venture firm Playground recently held a media event in Taiwan featuring a fireside chat with Ayar Labs, a silicon photonics startup backed by leading chipmakers including Nvidia and AMD. Founder and CEO Mark Wade and former Intel CEO Pat Gelsinger discussed the future of silicon photonics.
Wednesday 19 November 2025
Microsoft and Nvidia form multi-billion partnership with Anthropic
Three AI giants have formed a strategic partnership that will once again reshape the AI industry. Announced just before Microsoft Ignite 2025, Microsoft and Nvidia have committed to investing US$5 billion and US$10 billion, respectively, into Anthropic. In turn, Anthropic will purchase US$30 billion in computing capacity on Microsoft's Azure cloud computing platform. This deal will deepen Microsoft's shift toward a multi-model strategy, strengthen Nvidia's competitiveness against rival OpenAI, and provide Anthropic with the computing scale needed to leverage next-generation AI systems.
Wednesday 19 November 2025
SMIC resurges in memory and controller orders amid deepening AI-era crunch
The global semiconductor sector is at a critical juncture as demand for AI servers and high-performance computing (HPC) accelerates. A structural memory shortage is emerging across the supply chain, pushing SMIC back into a prominent role as a foundry.
Wednesday 19 November 2025
Santa Clara data centers remain idle due to local power shortages despite completion
Two newly constructed data centers in Santa Clara, California, have yet to begin operations because of insufficient power supply from the local grid. The facilities, developed by Digital Realty and Stack Infrastructure, offer a combined capacity of approximately 100MW but remain largely unused, highlighting ongoing energy constraints in the region.
Wednesday 19 November 2025
Iron Force expands into AI cooling systems with new factory
In recent years, Taiwan's automotive component makers have aggressively expanded into adjacent high-growth sectors, including thermal modules for AI servers and parts for semiconductor manufacturing equipment. Their diversification strategies are beginning to crystallize.
Wednesday 19 November 2025
Taiwan makes TOP500 global supercomputer list
The 66th edition of the global TOP500 supercomputer rankings was officially released at the 2025 International Conference for High Performance Computing, Networking, Storage, and Analysis (SC25). The dominance of the US in high-performance computing (HPC) was yet again put into the spotlight during the conference, with other notable highlights including Europe's Jupiter becoming the first exascale system on the continent. Taiwan also earned a spot on the list.
Tuesday 18 November 2025
Lagis opens second plant with triple capacity, targets own-brand and OEM growth
Taiwan's minimally invasive surgical device maker Lagis Enterprise has completed its second factory, three times the size of its original facility. Chairman Cheng-hung Chen said the plant has completed setup and is undergoing GMP certification, with production slated to begin by late November. The expanded capacity is expected to be a key growth driver in 2026.
Tuesday 18 November 2025
Taiwan connector maker Simula forecasts strong 2026 on cloud, drone, robotics, and EV demand surge
Qisda's connector manufacturing subsidiary Simula Technology expects three major growth drivers in 2026—cloud deployment, drones and humanoid robots, and automotive applications—supported by strong customer orders. The company remains highly optimistic about its 2026 outlook.
Tuesday 18 November 2025
GMI Cloud to build US$500 million AI factory in Taiwan with 7,000 Blackwell Ultra GPUs
GMI Cloud, an Nvidia Cloud Partner (NCP) and GPU-as-a-Service provider, has announced plans to build a US$500 million AI Factory in Taiwan, adding new large-scale compute capacity to the region's expanding AI infrastructure. The company said the facility is intended to support enterprises developing and deploying large AI models, forming part of a broader push to establish locally controlled, high-performance computing capabilities while maintaining access to advanced US technology.
Tuesday 18 November 2025
Ennoconn eyes stronger growth in 2026 with global service and brand integration
Amid accelerating consolidation in the global artificial intelligence of things (AIoT), smart manufacturing, and industrial control markets, the supply value chain is shifting from pure hardware manufacturing to integrated hardware-software solutions and comprehensive service offerings. Ennoconn chairman Fu-chuan Chu stated during the company's earnings call on November 14, 2025, that with the global localization of supply chains, and as the company's internal organization, products, and technologies are successfully restructured and its overseas deployments gradually take shape, the company expects operations in 2026 will be significantly stronger and show a higher growth trajectory compared to 2025.
Tuesday 18 November 2025
Etron benefits from booming edge AI market, shortages extend through 2026
AI has pushed the memory industry into a growth cycle. Etron Technology President Elvis Deng stated that memory supply is now in short supply across the board. DDR4, LPDDR4, and DDR3 shortages will be difficult to ease in the short term, and are expected to persist into the second half of 2026—2027.
Tuesday 18 November 2025
Tesla unveils three-phase blueprint to scale Optimus robot
The head of Tesla's artificial-intelligence division recently warned employees that 2026 will be the most demanding year of their lives, urging the AI organization to brace for an unprecedented pace of work, according to Business Insider.
Tuesday 18 November 2025
PDG announces US$700 million data-center investment in South Korea

Princeton Digital Group (PDG) has announced plans to enter the South Korean market with a US$700 million data-center investment, marking the company's seventh country of operation after Singapore, Japan, India, Indonesia, China, and Malaysia.

Tuesday 18 November 2025
Arm plugs into NVLink — and suddenly the AI data-center map looks less x86-shaped
Arm and Nvidia announced new steps to expand their collaboration on AI data center infrastructure, highlighting the growing demand for energy-efficient computing as AI workloads scale. Companies across the sector are prioritizing "intelligence per watt," as power constraints increasingly limit data center capacity and performance.
Tuesday 18 November 2025
Microsoft benefits from OpenAI tech sharing for ASIC development
Microsoft CEO Satya Nadella recently stated in an interview that Microsoft can access OpenAI's model research outcomes. System-level innovations developed by OpenAI can also be fully leveraged by Microsoft, which can implement these designs first for OpenAI and then extend or adapt them to its own purposes. Going forward, Microsoft will be able to share OpenAI's proprietary chip-related technologies and use them as a foundation to adjust or expand its own application-specific integrated circuit (ASIC) chip development.
Monday 17 November 2025
Taiwan MCU makers expand into power and cooling to mitigate AI data center power consumption surge
As rack-level power consumption and cooling requirements rise in data centers, Taiwanese microcontroller (MCU) suppliers are innovating in power-supply and cooling-fan controls. Nuvoton Technology recently introduced an integrated baseboard management controller (BMC). Additionally, Holtek is upgrading server-fan solutions and expects major growth in brushless DC motor (BLDC) applications.
Monday 17 November 2025
Reducer demand rises as humanoid robot competition intensifies across Taiwan and China
As the integration of production technology advances, the demand for humanoid robots and robotic dogs is growing, with the reducer—a key component known for its high gross margin and technical challenges—emerging as a competitive focus among Taiwanese and Chinese manufacturers. This competition underscores quality control and mass production capabilities.
Monday 17 November 2025
Microsoft launches AI superfactory to mitigate grid load
Microsoft has activated its data center in Atlanta, Georgia, connecting it with another Fairwater project data center in Wisconsin. This will support OpenAI and its own LLM training, using two-story buildings and liquid cooling to improve energy efficiency.
Monday 17 November 2025
Auras Technology reports record highs amid strong liquid cooling demand
The demand for liquid cooling in AI server racks is soaring, driving significant operational growth for Taiwanese cooling module manufacturer Auras Technology. Chairman Yu-shen Lin said that liquid cooling will be a major growth driver in 2026, with rising needs not only from Nvidia but also Advanced Micro Devices (AMD) and application-specific integrated circuit (ASIC) servers. The company estimates that revenue will grow by more than 50% in 2026.
Monday 17 November 2025
Nvidia tightens grip on Vera Rubin cooling supply chain, squeezing margins
Nvidia is preparing a major shift in how it manages the supply chain for its next-generation Vera Rubin server platform, tightening control over both assembly and key thermal components as power consumption rises and liquid cooling becomes mandatory.