AI data center 結果共42筆

Light was one of the hottest topics of conversations at Computex 2026. Across the keynotes and the show floor, attention kept returning to optical interconnects, the technology that carries data between chips as photons instead of electrons. The reason is simple: rack performance increasingly depends on how quickly data can move between the chips inside it, and copper is running out of headroom.

The PC came to Computex 2026 as a guest in its own house. The keynotes belonged to server silicon, the biggest booths to rack builders, and the show’s center of gravity to the data center. Yet the week still produced the most consequential PC news in a decade, because the machine itself is being rebuilt around AI agents at the exact moment its economics are being rewritten by AI’s appetite for memory.

Some of the biggest crowds at Computex 2026 were not gathered around AI server racks but the liquid cooling and power delivery systems that keep them running. In a clear sign of how the AI boom is changing, the booths of Delta, Lite-On, Auras, Jentech, and other companies in the segment saw lines stretching through the aisles. For years, cooling pipes and power supplies were boring background parts that only mattered when something broke. Now they have become one of the main stories. As AI racks grow more powerful, delivering the electricity they demand and removing the heat they generate have become two of the toughest challenges in the data center.

At Computex 2026, the silicon industry did more than simply attend the show. It realigned. While crowds focused on server racks, dexterous robots, and high-speed interconnects, the deeper story appeared in the chip announcements and partnership deals. The old battle lines of the processor world are dissolving. Nvidia has entered the CPU market, Arm is building its own data-center processors, MediaTek is expanding from phones to AI servers, and even Intel and Nvidia have found reasons to cooperate. What emerged is a new map of the industry, shaped by the demands of agentic computing and held together by alliances as much as by competition. All of this continues to rely on Taiwan’s unmatched manufacturing ecosystem.

The crowds arrived early and never thinned. From the moment the gates opened each morning, the aisles of the Nangang Exhibition Center filled with a press of analysts, buyers, engineers, reporters, and tech enthusiasts that did not let up until the halls closed in the evening. Booths that in previous years drew a polite trickle were three deep in people leaning in to watch a robot arm sort parts or a coolant loop pulse through a transparent server rack. There was a charge in the air that trade shows rarely carry, the sense of being inside something while it is still happening rather than reading about it afterward. The banner overhead read ＂AI Together,＂ and for once the slogan matched the feeling on the floor.

Delta Electronics unveiled a prefabricated AI modular data center at Computex Taipei, promising to cut setup time by 60%. The company highlights energy management and AI integration in infrastructure.

At Computex 2026 in Taipei, two of Taiwan’s largest hardware makers converged on the same strategic bet from opposite directions. The next phase of AI data center growth will be prefabricated, modular, and shipped in containers rather than built room by room on a customer’s site.

Computex 2026: Marvell, Intel and Nvidia CEOs reveal AI infrastructure race has shifted from chips to connectivity. Taiwan emerges as strategic hub.

黃仁勳AI五層蛋糕理論最底層就是能源、電力，台達電作為電源與散熱管理大廠，在COMPUTEX2026大秀一系列專為AI資料中心（Data Center）打造的極致散熱與微電網技術。董事長鄭平指出，「固態變壓器（SST）」技術其實已經量產很久，只是應用領域不同，現在在CSP領域仍在測試階段，不過中國市場倒是很願意嘗試使用。

Most of the attention in the AI server boom goes to the companies building the servers. Delta Electronics makes the power and thermal layer they all depend on.The Taoyuan-based firm heads into Computex 2026 with its strongest first-quarter results ever and near-record April revenue. Chairman Ping Cheng recently told investors that customer demand is “very, very large” and current capacity “absolutely insufficient” – language Delta has rarely used in any prior cycle. Its booth theme, “Superior Efficiency, Shaping Sustainable AI,” reflects two decades of quiet mastery over the technologies now most critical to AI: high-efficiency power conversion, advanced cooling, and their tight integration.

輝達（NVIDIA）北士科總部預計2030年落成，輝達執行長黃仁勳則喊話「需要更多的電力」。經濟部能源署今（28）日表示，2034年之前預計規劃累計有10.2 GW（1020萬瓩）的燃氣電廠上線，因此2034年包含半導體、AI Data Center的供電都是無虞。

Computex 2026 opens on June 2 in Taipei under the banner “AI Together,” and for Taiwan’s server makers the timing is close to perfect. The first quarter delivered record results across the entire group. Hyperscaler orders for Nvidia’s next-generation GB300 rack systems are accelerating into the second half of the year. And a new category of compute that the industry has started calling Physical AI, is opening a fresh growth lane alongside the familiar workloads of training and inference.

AI demand reshapes Taiwan’s memory market, driving up costs and creating a digital divide. Local tech prices soar as manufacturers prioritize AI over consumer needs.

As AI servers grow exponentially more powerful, the infrastructure needed to feed them electricity and keep them from overheating has become one of the technology industry’s most critical bottlenecks. In Taiwan, the companies solving these problems are posting some of the fastest growth rates in the entire hardware supply chain, and the momentum heading into Q2 2026 shows no signs of slowing.

Chenbro Micom may not be a household name, but inside every major data center, its products are everywhere. The Taiwan company is one of the world’s leading makers of server chassis, the metal enclosures that house the processors, memory, and cooling systems powering the cloud. As AI infrastructure spending surges, Chenbro is posting record numbers and rapidly transforming from a chassis specialist into a full-service provider of rack-level mechanical solutions.

Inventec Corporation, one of Taiwan’s leading electronics manufacturers, posted record quarterly revenue in the first three months of 2026, powered by surging demand for AI servers. The results underscore the company’s successful pivot toward high-value AI infrastructure, even as its traditional notebook business holds steady.

Delta Electronics started out decades ago making power supplies for televisions. Today it is a US$100 billion company and the single most important supplier of power and cooling systems for the world’s AI data centers. As artificial intelligence drives the biggest infrastructure buildout in a generation, Delta sits at the center of it all.

Taiwan semiconductor firms and server manufacturers and are aggressively expanding into the AI server application-specific integrated circuit (ASIC) market, positioning themselves to capture a substantial portion of the hyperscaler-driven custom silicon boom. Companies like MediaTek, Alchip Technologies, and Global Unichip Corporation are securing major design wins with cloud giants, leveraging Taiwan’s advanced foundry ecosystem and proximity to TSMC to accelerate development and deployment. This pivot offers these players significant revenue diversification beyond traditional mobile and consumer segments, with the data center ASIC market now projected to reach $50-70 billion by 2028 and AI server compute ASIC shipments expected to triple between 2024 and 2027.

Every AI accelerator that powers today’s largest models relies on an advanced substrate to connect the chip to the outside world. These ABF substrates, the high-performance interconnect layers that bridge GPU dies and circuit boards inside advanced packages, can only be manufactured at the required specifications by Japan’s Ibiden and Taiwan’s Unimicron Technology. 

Nvidia’s fiscal fourth-quarter results, reported on February 25, did more than silence doubters of the AI spending boom. They reaffirmed Taiwan’s position as the indispensable backbone of the global artificial intelligence supply chain. With record revenue of US$68.1 billion, up 73% year over year, and a fiscal first-quarter 2027 outlook of US$78 billion that sailed past Wall Street expectations, Nvidia delivered the kind of numbers that ripple outward through the island’s dense ecosystem of chip foundries, server manufacturers, and component makers. Nvidia’s data center segment produced US$62.3 billion in quarterly revenue, accounting for roughly 91% of total sales and growing 75% year over year. Gross margins held near 75%, underscoring that the company is scaling without sacrificing profitability. CEO Jensen Huang framed the moment as an ＂agentic AI inflection point,＂ pointing to enterprise adoption of AI agents and multi-quarter infrastructure buildouts by cloud providers and hyperscalers.

When Nvidia CEO Jensen Huang energy as the foundational layer of “the largest infrastructure buildout in human history＂ at Davos earlier this year, he could have been referring to the business that Delta Electronics has spent more than five decades building.

The AI infrastructure story has entered its second act. The first wave of investment chased silicon chips and the fabs that produce them. Now the market is confronting a different set of constraints: the physical ones. Heat must be removed. Power must be delivered. And increasingly, the companies solving these problems are headquartered in Taiwan.

Hon Hai and OpenAI partner to design AI infrastructure hardware in the U.S., enhancing supply chains and accelerating AI deployment. Production will occur domestically.

Foxconn will unveil its Model A electric vehicle and showcase AI and smart tech at Hon Hai Tech Day, Nov. 21-22. Expect insights from global tech leaders.

Foxconn partners with NVIDIA to launch an 800 VDC power architecture in Kaohsiung, enhancing AI data centers. This collaboration aims to boost AI factory efficiency and sustainability.