19 May 2025, Taipei, Taiwan – Hon Hai Technology Group (“Foxconn”) (TWSE:2317) will kick off COMPUTEX 2025 with the industry's most complete set of solutions for enterprise, agentic and physical AI factories, including advanced superchip AI infrastructure systems featuring the NVIDIA GB300 NVL72 architecture,, as well as systems and solutions for humanoid robots and smart manufacturing platforms – all available from the world’s largest electronics manufacturing service provider.

A centerpiece of Foxconn’s industry leadership at this year’s COMPUTEX will be Tuesday’s keynote by Foxconn Chairman Young Liu, outlining the competitive edge generative artificial intelligence plus robotics offer to enterprises in their quest for innovation and for Foxconn as it optimizes its three intelligent platforms. Thanks to genAI driving strong growth momentum in the global robotics industry, Foxconn’s collaboration with key partners will be a highlight of its exhibits this year covering robotics from application to manufacturing.

“GenAI plus robotics can assist, augment and optimize the way humans can innovate,” said Chairman Liu. “Foxconn’s three smart platforms all are connected to NVIDIA’s brilliant technologies. With the progress of our Smart Manufacturing, Smart EV and Smart City, Foxconn will help accelerate AI factories to become the standard for enterprise IT transformation underpinned by the NVIDIA platform.”

On the COMPUTEX stage for the first time, the Foxconn keynote begins at 0930 a.m. local time on 20 May 2025 on 7F at the Taipei Nangang Exhibition Center Hall 2.

Foxconn continues to promote smart manufacturing at multiple locations around the world and is working with NVIDIA to optimize production processes through a variety of AI solutions. On the exhibit floor, at Foxconn Booth L0106, the physical display of NVIDIA GB300 NVL72 AI infrastructure for training and reasoning large language models (LLM) with mega-level parameters, represents the state-of-the-art computing core of the AI factory. Foxconn, which holds more than 40% of the global AI server market, is a major supplier of NVIDIA GB200 NVL72 and the upcoming GB300 NVL72 architecture.

Moreover, the NVIDIA MGX 4U, also on display, is a highly flexible air-cooled server designed to meet the diverse computing needs of modern data centers. The system can be equipped with up to eight NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, combining advanced AI acceleration capabilities with professional-grade graphics performance to provide breakthrough acceleration for multiple workloads such as generative AI, graphics rendering, and video processing. Foxconn will also showcase the latest generation NVIDIA HGX B300 system. As the latest in the HGX architecture, B300 is designed for the AI inference era and has more powerful computing performance and higher memory capacity to meet the growing needs of AI model deployment.

Continuing Foxconn's industry-leading vertical integration advantage in AI servers, the Group is also exhibiting a full range of high-speed signal, high-voltage power supply and liquid-cooled full-speed connection solutions, including: new generation AI chip slot solutions, motherboard PCIE high-speed connectors and cables, GB300 Power Clip, Busbar cable and other related products. Through these innovative solutions, Foxconn strengthens its AI server cooling integration capabilities and continues to provide efficient, stable and customizable thermal management services.

In the field of robotics and smart manufacturing platforms, Foxconn’s Semiconductor Hybrid Robot will take a prominent place on the exhibit floor, combining its advanced visual recognition and precision motion technology to provide efficient and safe intelligent transportation solutions for the semiconductor industry. Meanwhile, the AI-powered Nurabot, jointly developed by Foxconn, Taichung Veterans General Hospital and Kawasaki Heavy Industries, leverages all three of NVIDIA’s computers for developing and deploying physical AI. Nurabot’s AI models are trained using their FoxBrain AI Factory powered by NVIDIA DGX SuperPOD, simulated using NVIDIA Omniverse, and deployed on NVIDIA Jetson.  This collaborative nursing robot is expected to be deployed to partner hospitals in Taiwan later this year andis estimated to be able to reduce nursing hours by 30% and improve care efficiency and quality.