NVIDIA Unveils AI Factory Taiwan Headquarters
At COMPUTEX 2025, NVIDIA founder and CEO Jensen Huang revealed exciting advances and partnerships across artificial intelligence, quantum computing, cloud services, and data center infrastructure. These innovations aim to accelerate AI development and adoption worldwide. Central to the announcement was the unveiling of NVIDIA Constellation, the company’s future headquarters in Taiwan, located at the Beitou-Shilin Technology Park in Taipei.
This new AI factory Taiwan facility underscores the region’s skilled workforce and strong semiconductor supply chain. It is expected to attract top talent from around the globe, reinforcing Taiwan as a vibrant hub for AI innovation.
Strengthening AI Factory Taiwan Partnerships
NVIDIA is deepening its collaboration with Foxconn Hon Hai Technology Group and the Taiwan government to build a supercomputer-based AI factory. This facility will deliver NVIDIA Blackwell infrastructure to researchers, startups, and numerous industries. Foxconn’s subsidiary, Big Innovation Company, will supply the AI infrastructure as an official NVIDIA Cloud Partner.
Equipped with 10,000 NVIDIA Blackwell GPUs, this AI factory Taiwan initiative will greatly expand AI accessibility and spark new innovations across Taiwan’s research and business sectors. The factory will feature cutting-edge NVIDIA Blackwell Ultra systems, including the GB300 NVL72 rack-scale solution, NVIDIA NVLink, Quantum InfiniBand, and Spectrum-X Ethernet.
Supporting Taiwan’s Technology Ecosystem
The Taiwan National Science and Technology Council will utilize the Big Innovation supercomputer to offer AI cloud computing resources, accelerating AI adoption throughout the local tech ecosystem. Notably, Taiwan Semiconductor Manufacturing Company (TSMC) researchers will also harness this powerful supercomputer to boost their R&D efforts.
Enabling Custom AI Infrastructure with NVLink Fusion
NVIDIA introduced a new silicon technology, NVLink Fusion, that empowers industries to build semi-custom AI infrastructure. Early adopters include MediaTek, Marvell, Alchip Technologies, Astera Labs, Synopsys, and Cadence. This silicon allows CPUs from Fujitsu and Qualcomm to seamlessly integrate with NVIDIA GPUs, creating powerful AI factories.
Additionally, NVLink Fusion supports cloud providers in scaling AI factories to millions of GPUs. It achieves this by combining any ASIC with NVIDIA’s rack-scale systems and networking platform, delivering throughput up to 800 Gb/s. Key components include NVIDIA ConnectX-8 SuperNICs, Spectrum-X Ethernet, and Quantum-X800 InfiniBand switches, with co-packaged optics arriving soon.
Accelerating Enterprise AI Factory Transition
To help the trillion-dollar IT infrastructure sector pivot to enterprise AI factories, NVIDIA introduced RTX PRO Servers and a new validated design called NVIDIA Enterprise AI Factory. Powered by RTX PRO 6000 Blackwell Server Edition GPUs, these servers offer exceptional performance and energy efficiency.
The Enterprise AI Factory design enables partners to build on-premises infrastructure with RTX PRO Servers, Spectrum-X Ethernet networking, BlueField DPUs, certified storage systems, and AI Enterprise software. This setup speeds up product design and engineering simulations, transforming how data centers operate.
AI-First DGX PCs Launch in Taiwan
NVIDIA announced that top Taiwanese system manufacturers will produce DGX Spark and DGX Station systems, empowering developers, data scientists, and researchers worldwide. Featuring the GB10 Grace Blackwell Superchip and fifth-generation Tensor Cores, the DGX Spark delivers up to 1 petaflop of AI compute with 128GB unified memory.
The DGX Spark allows seamless model export to NVIDIA DGX Cloud or any accelerated data center infrastructure. Available from leading brands like Acer, ASUS, Dell, GIGABYTE, HP, Lenovo, and MSI starting July, it represents a major step forward.
Meanwhile, the DGX Station, equipped with the GB300 Grace Blackwell Ultra Desktop Superchip, offers up to 20 petaflops of AI power and 784GB unified system memory. It supports networking speeds up to 800 Gb/s via NVIDIA ConnectX-8 SuperNIC for high-speed connectivity and multi-station scaling. This system will launch later this year through ASUS, Dell, GIGABYTE, HP, and MSI.
Powering the World’s Largest Quantum Research Supercomputer
NVIDIA also announced the opening of the Global Research and Development Center for Business by Quantum-AI Technology (G-QuAT). This center hosts ABCI-Q, the world’s largest quantum research supercomputer, delivered by Japan’s National Institute of Advanced Industrial Science and Technology.
ABCI-Q features 2,020 NVIDIA H100 GPUs interconnected via the NVIDIA Quantum-2 InfiniBand platform. It integrates NVIDIA CUDA-Q, an open-source hybrid computing platform that orchestrates hardware and software to run complex quantum computing applications.
Introducing DGX Cloud Lepton Marketplace
Finally, NVIDIA launched DGX Cloud Lepton, an AI platform with a compute marketplace connecting developers building agentic and physical AI applications with tens of thousands of GPUs globally. Cloud partners like CoreWeave, Crusoe, Firmus, Foxconn, GMI Cloud, Lambda, Nebius, Nscale, Softbank, and Yotta Data Services will offer NVIDIA Blackwell and other GPU architectures through this marketplace.
This platform enables developers to access GPU compute capacity in specific regions, supporting strategic and sovereign AI needs. Early access registration for DGX Cloud Lepton is now open.
For more news and updates on AI factory Taiwan, visit Filipinokami.com.