At its annual Nvidia GTC, CEO Jensen Huang laid out an ambitious roadmap for the future of artificial intelligence, centered on increasingly powerful chips and massive computing infrastructure. The event, often called the “Super Bowl of AI,” showcased how Nvidia plans to stay at the heart of the AI boom by supplying the hardware that powers everything from chatbots to robotics.
A key focus of Huang’s presentation was the shift from training AI models to running them in real-world applications—known as inference. Nvidia introduced new chip designs and systems aimed at handling this next phase, where AI tools are deployed at scale across industries. Huang described this moment as a major turning point, signaling that AI is moving beyond experimentation into widespread, everyday use.
The company also revealed its long-term chip roadmap, including upcoming architectures like Rubin and future systems designed to deliver dramatic performance gains. These chips are part of a broader strategy that connects processors, data centers, and software into what Huang calls a full AI “stack”—an ecosystem needed to support the rapid expansion of AI across the global economy.
Perhaps most striking was the scale of Nvidia’s ambition. Huang suggested that demand for AI infrastructure could generate trillions of dollars in economic activity, with chip revenue alone potentially reaching around $1 trillion in the coming years. This reflects a broader belief that AI is not just another tech trend, but a foundational shift comparable to the rise of the internet—reshaping industries, economies, and the future of computing itself.
