In a world seemingly heading towards artificial intelligence, Jensen Huang's keynote address at the GTC Taipei event marked yet another critical moment for industry, as the Nvidia founder and CEO talked about the growing trends in the market, and how AI is reshaping itself.
Here are five key takeaways from Huang's keynote address at the GTC Taipei 2026.
Nvidia x Microsoft
Nvidia's partnership with Microsoft was by far the most eye-catching announcement from Huang's keynote speech, with the two tech giants set to build the next generation of PCs.
These computers will be powered by Nvidia's brand new RTX Spark superchip, powered by the Blackwell architecture and capable of running a version of Microsoft Windows that can offer agentic services.
The partnershiop between Nvidia and Microsoft is significant as it threatens to eat away market share from Intel, AMD, which power most windows devices to this day. Apple, too, could feel the heat as it faces a strong ARM-based CPU competition for laptops.
ALSO READ: Nvidia, Microsoft Team Up To Take On Apple In PC, Laptop Market — Here's What It Means For You
The time of Agentic AI
Much of Huang's keynote speech revolved around agentic AI, with the Nvidia founder arguing that the time for AI to think complex and help out an end user through its agentic capabilities is finally here.
Huang has integrated these agentic AI systems into most of the products announced at the GTC Taipei event, including the new RTX Spark superchip, which will power the Windows.
AI for profit
Huang further noted that AI is at a point where it has become profitable for companies, especially ones who are willing to show enterprise and utilise the agentic capabilities.
Huang added that AI can now power economies and thanks to his efficient taskforce, it can yield 3x its investment.
First CPU not made for humans
Apart from RTX Spark, Nvidia also unveiled Vera - a whole new CPU architecture that is designed not for humans but for AI agents.
These high-performance CPUs are designed to power a computer that is exclusively run by AI, rather than letting an user do menial tasks such as browsing a web, playing a game or editing a photo.
Now in full production, Vera is powered by Olympus, which is Nvidia's custom CPU core, and features 88 cores. VERA also features Spatial Multithreading and an LPDDR5X memory subsystem delivering up to 1.2TB/s of bandwidth.
Vera Rubin
From the company's perspective, the most exciting announcement was Vera Rubin, the next-generation of Nvidia AI infrastructure system that will look to succeed Blackwell.
Vera Rubin is Nvidia's first fully agentic AI-based compute infra that promises to feed not individuals but large companies with an equally large data centres and network demand.
Vera Rubin is in full production, with companies like Anthropic, Open AI and SpaceX already emerging as early adopters.
ALSO READ: Nvidia Announces Vera — World's First CPU Not Designed For Humans
Essential Business Intelligence, Continuous LIVE TV, Sharp Market Insights, Practical Personal Finance Advice and Latest Stories — On NDTV Profit.