Nvidia doubles down on platform role
Nvidia is hosting GTC Taipei and framing itself as a neutral platform that won’t ‘pick winners’, while highlighting ecosystem co‑innovation across enterprise, robotics and edge AI. Jensen Huang’s Taipei appearances and keynote agenda signal a push to monetise ecosystem gravity rather than favour individual model vendors. ( )
Nvidia is turning its Taipei event into a pitch for itself as the platform underneath the artificial intelligence industry, not a backer of one model winner over another. (nvidia.com, businessinsider.com) Nvidia says GTC Taipei at Computex 2026 will run June 1-5, with Jensen Huang speaking live at Taipei Music Center and conference sessions at the Taipei International Convention Center. Nvidia’s event page says the agenda spans “physical AI,” compute, infrastructure, robotics, and training sessions with partners. (nvidia.com, nvidia.com) Huang said on the “Dwarkesh” podcast, released April 15, that Nvidia tries to invest in “all” major foundation-model companies and “don’t pick winners.” Business Insider reported April 16 that he tied that approach to Nvidia’s early history, when dozens of graphics startups competed and Nvidia itself was not the obvious survivor. (businessinsider.com) That posture fits Nvidia’s Taiwan message. In Taipei on February 2, Huang said Nvidia is “the only company that works with every AI company” and is present in cloud data centers, enterprise systems, robots, and cars. (focustaiwan.tw) Taiwan is central to that claim because Nvidia’s platform business depends on suppliers as much as software. Huang said on February 2 that Nvidia is hiring more engineers in Taiwan as its supply chain grows, and that the company plans to expand its presence there with a larger site. (focustaiwan.tw) The event programming shows how Nvidia wants to sell that role. Its GTC Taipei page highlights sessions on agentic and reasoning artificial intelligence, “AI factories” for large-scale computing, robotics, and infrastructure, while separate Computex forum talks feature Nvidia executives on robotics, edge AI, networking, and development. (nvidia.com) That is a broader pitch than selling chips alone. Huang said in February that Nvidia now builds graphics processors, central processors, networking chips, switches, and other products, and the company’s GTC materials frame the conference around the “full AI stack,” from chips and infrastructure to models and applications. (focustaiwan.tw, nvidia.com) Nvidia has also backed companies across that stack. Business Insider reported that Nvidia has stakes in CoreWeave, Intel, Synopsys, Nokia, Mistral AI, Wayve, Scale AI, and Figure AI, and said Huang described OpenAI and Anthropic as likely nearing the point where Nvidia would stop adding to those positions because they are expected to go public. (businessinsider.com) Huang has been signaling for months that Taipei would be a major stop. He told reporters on February 2 that he would return for Computex, make multiple announcements, and deliver a keynote, then said again at the airport that the June event would include “several new announcements.” (focustaiwan.tw, focustaiwan.tw) The result is a simple Nvidia story for June 1 in Taipei: if every model company, cloud provider, robot maker, and enterprise wants a common base layer, Nvidia wants to be the company selling it. (nvidia.com, businessinsider.com, focustaiwan.tw)