Flexium Interconnect chairman Walter Cheng said the company has emerged from the toughest operating downturn of the past two years and expects AI orders for smart glasses and infrastructure to accelerate in 2027, which is also expected to lower consumer electronics revenue to below 50%. The flexible printed circuit (FPC) maker expects its polytetrafluoroethylene (PTFE) products to enter mass production in 2028 and is targeting a return to profitability in the first quarter of 2027.

Qualcomm's weakening smartphone business is accelerating its transformation into an AI infrastructure supplier, with custom accelerators, server CPUs, and optical connectivity emerging as growth priorities. Following a 26% annual decline in handset chip revenue in fiscal 4Q26, the US chipmaker is targeting US$5 billion in data center revenue in fiscal 2027 and US$15 billion by fiscal 2029. The question is whether these businesses can scale quickly enough to offset pressures in its traditional market.

South Korean AI software developer Moreh says AMD Instinct MI355X GPUs running its inference software delivered more than twice the throughput of Nvidia B200 GPUs in company tests, as South Korea moves ahead with plans to build AI infrastructure combining AMD processors with locally developed chips.

South Korean cloud services provider Naver Cloud now runs AMD EPYC processors in about 35,000 servers and plans to deploy 144-core fifth-generation EPYC chips in its public cloud services in the second half of 2026. The company turned to AMD after high prices for Intel's higher-core-count processors slowed its server expansion.

Chinese AI developers are losing a key route to advanced US chips, renting them in data centers outside China, as Washington and US companies stretch American controls over compute well beyond US borders. The squeeze falls heavily on Southeast Asian data center operators that built capacity around Chinese demand. Beijing, meanwhile, is pressing its own tech giants to rely less on American hardware.

TECO Electric and Machinery is extending its motor and drive-control expertise into rotary actuators, developing joint modules for humanoid and quadruped robots through in-house motor, drive, and mechanical integration, alongside Taiwan-based reducer suppliers. The company said it has already shipped robot-dog joint products developed with Taiwan's Industrial Technology Research Institute (ITRI) and is now pursuing smaller joint modules while evaluating the dexterous-hand market.

Inventec reported October revenue of NT$79.97 billion (approx. US$2.44 billion), down 18.9% from September but up 46.0% from October 2025, setting a record high for the same period. Revenue for the first 10 months totaled NT$822.23 billion, up 43.1% year-over-year, also marking a record high for the period.

As concerns over AI deepen and negative headlines continue to emerge, companies across the AI server supply chain say customer orders remain strong. Industry sources said Nvidia GPU servers are entering a transition phase between old and new platforms. However, ASIC servers are filling the gap, supporting a positive outlook for the fourth quarter of 2026 as Nvidia's Vera Rubin shipments gradually ramp up in the first quarter of 2027.

Google's plan to invest US$15 billion in Malaysia over 2027 and 2028 deepens its bet on what AFP calls Southeast Asia's fastest-growing data center hub, and ties the expansion to commitments on power, water, and skills as Malaysian authorities scrutinize the facilities' demands on national resources.

Airoha, a MediaTek subsidiary, expects its optical communications business to deliver another multiple-fold increase in 2027 as 100G pulse-amplitude modulation (PAM4) digital signal processor (DSP) shipments ramp and its 200G product targets tape-out by the end of the year. The company also reported record quarterly revenue and profit in the third quarter of 2026, driven by its wired networking infrastructure and wireless edge AI businesses.

As AI data center construction continues to scale up, shortening build times and bringing computing capacity online faster have become key priorities for cloud service providers (CSPs) and AI infrastructure companies. Nvidia's recently launched DSX POD reference design seeks to move equipment installation, cabling, piping, and system testing from data center sites to factories, pushing AI data centers toward prefabricated, modular delivery.

Voters in Sunbury, Ohio, approved a ban on large data centers in the November 3 election, raising a new obstacle to Amazon's proposed US$2 billion project as communities across the US move to restrict AI infrastructure development.

Asian stocks wavered on November 4 in response to the US midterm election results. While the final votes are still being counted, current results show the Democratic Party winning majorities in both the House of Representatives and Senate, a clear indication of voters' discontent about data center development in their localities, which could usher in more regulation of the AI industry.

Taiwan has taken a major step toward regulating its virtual asset market, with the Financial Supervisory Commission (FSC) announcing supporting rules for the Virtual Asset Services Act on November 3. As the regulatory framework takes shape, stablecoins, digital asset custody and blockchain-based financial services are emerging as potential drivers of financial innovation, although industry leaders and regulators warn that security and risk management remain critical to market development.

Supermicro expressed confidence to analysts during its earnings call for the first quarter of fiscal year 2027 regarding quicker turnaround times for orders and improved cash flow. The company cited greater demand for its Data Center Building Block Solutions (DCBBS), which accelerate data center infrastructure deployments, but kept a cautious full-year sales outlook to account for component shortages.

SpaceX, which only entered the AI infrastructure business this year through its merger with xAI, may now be the fastest data center builder in the world, according to research firm SemiAnalysis. The firm argues this pace will make the rocket company the world's biggest AI cloud.

Cloud and AI services provider Acer E-Enabling Service Business (AEB) reported 3Q26 revenue of NT$3.822 billion (approx. US$116.7 million), up 70% year-over-year, as AI application demand drove simultaneous growth in revenue and profit.

Anthropic is demonstrating extraordinary demand for frontier AI, with third-quarter revenue reaching about US$17 billion, nearly twelve times the US$1.4 billion recorded a year earlier. But as the company prepares for one of the largest technology IPOs in history, its latest filing also exposes the growing financial burden required to sustain that expansion.

Samsung Electronics' and SK Hynix's dependence on sustained AI infrastructure spending is bringing customer financing risk into focus. As demand for high-bandwidth memory and server DRAM lifts their businesses, industry discussion of possible financial support for AI customers raises a further issue: protecting future chip orders could also expose suppliers to customers' investment risks.

AMD is pursuing a larger role in AI infrastructure by combining its growing server CPU franchise with rack-scale systems, broadening its offering beyond individual processors. The company's earnings call highlighted demand for CPUs that run agents and support accelerators, alongside the first revenue shipments of Helios systems built around its newest AI GPUs.

AMD expects company-wide revenue to grow more than 80% in 2027, an expansion increasingly dependent on delivering AI systems and server processors at scale. Data centers already generate nearly two-thirds of sales, but management's outlook leaves execution and product mix central to how much demand converts into revenue and profit.

AMD is preparing for another difficult year in PCs even as demand for its data center products exceeds available supply. The company's earnings call exposed the widening gap between its businesses: higher component costs are weighing on PC purchases, gaming revenue is contracting, and embedded processors are providing a separate source of growth. Chief executive Lisa Su nevertheless expects company-wide revenue to grow more than 80% in 2027.

Supermicro's latest results show its expansion into complete AI data center systems gaining scale, while leaving open how much of that growth can translate into durable profitability and cash. Record sales lifted earnings sharply, but the company retained less of each sales dollar than in the preceding quarter and continued to consume operating cash.

As US export controls make it increasingly difficult for China and other restricted nations to obtain advanced chips and AI servers, Taiwan is taking proactive measures to tighten its own export controls and prevent the nation from becoming a transshipment loophole. Under the Agreement on Reciprocal Trade (ART) between Taiwan and the US, Taiwan's Ministry of Economic Affairs (MOEA) plans to align local regulations with the US Foreign Direct Product Rule (FDPR).

Foxconn is leveraging its precision manufacturing and vertical integration capabilities to strengthen its position in AI infrastructure, where rapid server upgrades increasingly demand tighter component tolerances and faster system delivery. At its recent Technology Day, the company demonstrated how its expertise in materials processing, optical communications assembly, AI racks, and modular data centers can shorten the path from product design to deployment.