The air at Taipei Music Center crackled with anticipation as Nvidia CEO Jensen Huang took to the stage at Computex 2025, announcing a series of strategic moves that firmly cement the tech giant's unwavering dominance in the artificial intelligence landscape. Amidst the buzz of new technologies, Huang's announcements carried a significant undertone, addressing the delicate balance between innovation and the rising tide of protectionist sentiments, particularly from the US.

A key highlight of Huang's keynote was the launch of NVLink Fusion, a groundbreaking technology that breaks down Nvidia's traditional walled garden approach to AI infrastructure.

For the first time, Nvidia will sell this technology to other chip designers, to help build powerful custom AI systems with multiple chips linked together.

This move acknowledges the growing trend of custom silicon development by major cloud providers and positions Nvidia as the crucial interconnectivity layer in increasingly heterogeneous AI systems. Notably, MediaTek and Marvell have already announced their plans to adopt NVLink Fusion for their custom chip efforts, signaling a significant shift in the industry. This technology, an extension of Nvidia's high-speed NVLink, is designed to accelerate chip-to-chip communication, a critical factor in building and deploying advanced AI tools.

In another significant announcement, Huang unveiled the DGX Spark, a desktop AI PC specifically targeting AI researchers. This powerful machine, now in full production and expected to be ready in a few weeks, aims to put supercomputing capabilities directly into the hands of developers and scientists . While specific technical details were not immediately available, the focus on a desktop form factor suggests a move to democratise access to high-performance AI computing for individual researchers.

Adding a layer of geopolitical significance to his address, Huang disclosed Nvidia's plan to establish a new headquarters in Taiwan, located in the northern suburbs of Taipei. This announcement underscores Nvidia's deep-rooted ties with Taiwan, a critical hub in the global semiconductor supply chain, particularly for manufacturing its advanced chips. Huang's emphasis on Taiwan's importance as an "ideal nexus for AI advancement" comes at a time when the US, under President Trump, has been vocal about imposing tariffs to encourage companies to increase production within the United States.

While Huang did not directly address the tariff issue in detail during his keynote, the backdrop of Trump's rhetoric and the “Make in USA” push loomed large. Nvidia, while strengthening its presence in Taiwan, has also made moves to align with the US government's objectives. Earlier this year, the company announced plans to manufacture AI supercomputers in the US for the first time, at TSMC facilities in Arizona and Texas. This dual approach – investing in both Taiwan and the US – suggests a calculated strategy to navigate the complexities of global trade and geopolitical pressures.

Huang's keynote at Computex 2025 painted a picture of Nvidia not just as a chipmaker, but as an infrastructure provider for the burgeoning AI revolution. By opening its NVLink technology, empowering researchers with desktop AI power, and solidifying its presence in Taiwan, Nvidia is strategically positioning itself at the heart of the AI ecosystem, even as the global trade landscape continues to evolve under the shadow of potential tariffs and calls for domestic manufacturing. The coming weeks and months will reveal how these announcements translate into tangible impacts on the AI industry and Nvidia's continued leadership.

Computex 2025, expected to have over 1,400 exhibitors and over 80,00 attendees begins on May 20, and will be the first major gathering of computer and chip executives in Asia since U.S. President Donald Trump threatened to impose tariffs to compel companies to increase production in the United States. Fortune India will be present at the event and covering all significant developments across multimedia.