Taipei, Dec 12, 2025 — Taiwan has officially launched a major new cloud computing and AI hub powered by Nvidia high-performance chips, marking a significant shift in its technology strategy and affirming its central role in the global AI ecosystem.
President Lai Ching-te inaugurated the state-of-the-art facility in Tainan, part of Taiwan’s broader strategy to build a “sovereign AI” infrastructure aimed at reducing dependency on foreign cloud services and boosting domestic tech capability.
- The centre is built with a 15-megawatt power capacity to support heavy computing workloads. It houses Taiwan’s most advanced supercomputer, “Nano 4,” featuring 1,760 Nvidia H200 chips and 144 Nvidia Blackwell GPUs — among the most powerful AI processors currently available.
This new facility isn’t just another data centre — officials describe it as the cornerstone of Taiwan’s transition from a hardware manufacturing hub to a high-performance cloud and AI services leader.
The cloud hub is a key element of Taiwan’s “Ten Major AI Infrastructure Projects”, launched mid-2025 to strengthen its position in AI research, high-performance computing, telecommunications, cloud services, and digital content development.
President Lai framed the launch as a milestone in transforming Taiwan into an “AI island” with technological sovereignty — one capable of hosting globally competitive AI workloads independent of external providers.
U.S. chip designer Nvidia plays a central role in this project, with its GPUs providing the computational backbone. These chips are largely manufactured by Taiwan Semiconductor Manufacturing Company (TSMC), reinforcing Taiwan’s deep integration into global AI supply chains. TradingView
This hub reflects broader market confidence in the island’s strategic position:
- Taiwan’s stock markets are tracking toward record highs, supported by strong demand for AI and semiconductor-related equities — underscoring investor belief in Taiwan’s long-term tech leadership. The launch underscores Taiwan’s leverage not only in hardware manufacturing but advanced system integration and high-performance computing — capabilities that go beyond chips to complete compute platforms. News.az
GMI Cloud, a U.S.-based provider, is building a $500 million AI data centre in Taiwan, expected to come online by March 2026 with roughly 7,000 Nvidia GB300 GPUs across 96 high-density racks — further expanding local cloud and AI compute capacity.
Taiwan’s tech ecosystem, notably via partnerships with global leaders like Nvidia, continues attracting advanced installations and investments aimed at capturing AI workloads and innovation cycles regionally.
This development highlights several broader trends:
- AI infrastructure is now a national strategic priority, not just a commercial interest. Taiwanese policymakers and industry are accelerating efforts to anchor advanced compute power domestically. Nvidia’s technology acts as a focal point for enabling these ambitions, reinforcing Taiwan’s synergy with U.S. tech players.
GMI Cloud to build $500 million AI data centre in Taiwan with Nvidia chips
U.S.-based cloud services provider GMI Cloud said on Monday it will build a $500 million artificial intelligence data centre in Taiwan with the support of U.S. chipmaker Nvidia .
www.reuters.com