Taiwan’s AI Server Giants Post Record Q3 Results as Demand Soars
Taiwan’s AI server manufacturing giants delivered a spectacular third quarter of 2025, shattering previous records, rapidly ramping shipments, and firmly cementing their dominance in global next-generation server design and production.
At the forefront was Foxconn (Hon Hai), the world’s largest electronics manufacturing service provider. The company recorded a September all-time-high revenue of NT$837.1 billion (about US$27.55 billion). For the third quarter, revenue soared to NT$2.57 trillion (US$84.7 billion based on a Q3 average exchange rate of ~30.3 NT$/US$), marking an increase of nearly 11% year-on-year and 14.5% compared to the previous quarter. Notably, Foxconn’s Cloud and Networking Products division, which includes AI server racks, outperformed its flagship iPhone business for the first time ever, and accounts for over half of Foxconn’s total server sales. This shift underscores the company’s successful transition toward AI and data infrastructure.
Company executives highlighted exceptionally robust demand for AI rack shipments heading into Q4 2025 and beyond, underpinned by deep partnerships with Nvidia. Foxconn also leveraged its early lead in new platforms such as Vera Rubin and GB300, both of which are expected to mark significant future growth milestones. While exchange rate fluctuations and typical consumer product seasonality presented headwinds, Foxconn’s long-term bets on AI server innovation and its ongoing expansion of U.S. manufacturing have begun to unlock substantial new growth.
Quanta, another manufacturing giant, reported record-breaking revenues of NT$184.1 billion (about US$6.06 billion) for September and a third-quarter total of NT$495.3 billion (US$16.3 billion). By September’s end, its year-to-date revenues had reached NT$1.49 trillion (about US$49 billion). While Quanta’s notebook shipments peaked at 12.7 million units in the quarter, it is AI rack shipments, especially as the company transitions from supplying Nvidia’s GB200 to GB300 platforms, that are increasingly driving growth. Company executives expect this trend to continue powering revenues into 2026.
Wistron posted outstanding third-quarter results, with revenues surging 108% year-on-year to NT$567.8 billion (about US$18.7 billion), including a 110% jump for September alone. While PC and monitor shipments held steady at about 2.45 million units each, it is AI server shipments where Wistron anticipates record volumes in the months ahead, spurred by escalating demand from enterprise clients.
Wiwynn, a specialist in large-scale data center equipment, surged to NT$86.3 billion (US$2.84 billion) in September revenue, up 150% year-on-year. Q3 sales rose 173% to NT$266.8 billion (US$8.8 billion), and the company’s year-to-date total hit NT$658.2 billion (US$21.7 billion), a 169% increase from the previous year. This strong momentum cements Wiwynn’s position as a critical supply partner for the world’s largest AI infrastructure deployments.
Mitac also posted a new high in September, crossing the NT$10 billion (US$330 million) revenue mark for the first time, fueled by high-value shipments for cloud data centers and advanced automotive electronics. Its quarterly revenue grew more than 63% year-on-year, while digital spinoff Mitac Digital Technology delivered a 16% annual revenue boost, benefiting from smart solutions for industrial and cloud applications.
The supply chain’s strength is evident upstream as well. Component maker Lite-On reported cloud and server-related revenues growing 80% year-on-year, reflecting sustained investment and fast design cycles throughout the ecosystem.
The industry’s competitive landscape is particularly noteworthy. Foxconn currently maintains an edge in new product introduction (NPI) for Nvidia’s most advanced platforms, including Vera Rubin and the 600kW liquid-cooled “Kyber” racks due from 2026. Quanta and Wistron are racing to catch up, actively seeking future design wins, but Foxconn’s established leadership gives it a strong edge.
Despite the sector’s record-setting quarter, several challenges loom. Supply constraints in high-bandwidth memory (HBM), power subsystems, and liquid cooling technology remain critical bottlenecks as complexity ramps. Executives warn that longer component lead times and increased capital requirements could pressure operational flexibility across the supply chain. Furthermore, global currency market volatility and geopolitical uncertainty introduce ongoing cost and allocation risks. In response, many manufacturers are expanding their capacity footprints in the United States and Mexico to better meet localization requirements from cloud customers and to manage macro and policy risks.
The real story behind these numbers is that Taiwan has moved from quietly enabling PC and smartphone growth to standing center stage as the world’s AI and data center hardware nexus. Today, the AI server, not the consumer gadget, is the flagship technology, and Taiwan’s technical know-how, supply chain depth, and world-class engineering are its core differentiators.
Looking ahead, the AI server surge shows no signs of tapering. Every leap forward in generative AI and hyperscale cloud is matched by new challenges for power, thermal management, and integration. With their ability to innovate, scale, and overcome tough technical challenges, Taiwan’s AI server manufacturers are strongly positioned to capitalize on this opportunity.
