As Nvidia CEO Jensen Huang announced on May 19, 2025, at the Computex conference in Taipei, Taiwan, these measures represent nothing less than the next revolution in computing. Just recently, the company publicized intentions to fortify its foothold in Taiwan. They will open a new Taiwanese office and work on an ambitious AI supercomputer build together with Foxconn (Hon Hai Technology Group). Taken together, these projects demonstrate Nvidia’s dedication to offering the best deep learning platform available.
Nvidia just made their technological ecosystem stickier with the announcement of a new “NVLink Fusion” program. This new move allows customers to run and test non-Nvidia CPUs and GPUs on their systems without seamless integration. This program has the potential to transform AI infrastructures. More importantly, it will allow smooth interoperability between Nvidia drivers and numerous third-party application-specific integrated circuits (ASICs).
The NVLink Fusion program is proud to have industry-leading partners. These are MediaTek, Marvell, Alchip, Astera Labs, Synopsys and Cadence. Nvidia is indeed getting in very deep with both custom CPU developers and ASIC designers. This strategy is designed to maximize the inclusion of data centers that utilize ASIC technologies.
Ray Wang, a semiconductor and technology analyst, highlighted how critical NVLink Fusion is to Nvidia’s strategy. He stated, “If widely adopted, NVLink Fusion could broaden NVIDIA’s industry footprint by fostering deeper collaboration with custom CPU developers and ASIC designers in building the AI infrastructure of the future.”
Huang kicked off his keynote address with an introduction to Nvidia’s newly announced next-generation Grace Blackwell systems purpose-built for AI workloads. MTR Corporation’s new “GB300” series, expected to enter service in the 3Q of 2025, should enhance overall system performance even further.
The news of the introduction of the new NVIDIA DGX Cloud Lepton raised eyebrows. This new shared AI platform will power a new compute marketplace. It will enable US-based AI developers to connect to tens of thousands of NVIDIA GPUs from cloud providers around the world. Second, the platform addresses a fundamental challenge—the need to reliably get high-performance GPU resources. It does this by simplifying access to cloud AI services.
“DGX Cloud Lepton helps address the critical challenge of securing reliable, high-performance GPU resources by unifying access to cloud AI services and GPU capacity across the NVIDIA compute ecosystem.” – Nvidia
Under the NVLink Fusion framework, companies like Fujitsu and Qualcomm Technologies can connect their third-party CPUs with Nvidia’s GPUs within AI data centers. Huang explained some of the strategic benefits of this flexibility.
“At the system level, the added flexibility improves the competitiveness of Nvidia’s GPU-based solutions versus alternative emerging architectures, helping Nvidia to maintain its position at the center of AI computing.” – Rolf Bulk
He described how the initiative includes much, much more than chips.
“NV link fusion is so that you can build semi-custom AI infrastructure, not just semi-custom chips.” – Jensen Huang
Nvidia is working hard to do just that. The firm’s intentions are to bolster collaborations with Taiwanese tech companies and lead the area’s growth as a center for artificial intelligence advancement.
“We are delighted to partner with Foxconn and Taiwan to help build Taiwan’s AI infrastructure, and to support TSMC and other leading companies to advance innovation in the age of AI and robotics.” – Jensen Huang
The collaboration with Foxconn on the AI supercomputer project marks a significant step in Nvidia’s strategy to enhance its capabilities within Taiwan’s thriving technology sector. Nvidia is riding the wave of demand for all things AI and deserves credit. Their recent success is attributed to powerful recurring office launches, product invention, and collaborations.