Jensen Huang just told the world that Nvidia's trillion-dollar AI chip projection is the floor, not the ceiling.
The Signal
One day after Huang pegged Nvidia's AI chip sales trajectory at $1 trillion, he clarified that figure doesn't include the rest of what Nvidia sells. Both Bloomberg sources confirm the same core story: the AI chip number is a subset, and total revenue will climb higher as Nvidia expands beyond its GPU fortress.
This matters because Nvidia isn't just riding the AI training wave anymore. The company that became a $2 trillion market cap by selling pickaxes to the AI gold rush is now diversifying the product line. Huang's language is deliberate. He's signaling new revenue streams before Wall Street starts worrying about GPU saturation. The AI chip market alone hitting $1 trillion would already make Nvidia larger than most countries' GDP. Now add networking, automotive AI, robotics inference chips, and whatever else they're cooking.
What's not in these sources but matters: Nvidia's margins. If they're pushing into "new markets," that likely means lower-margin products than their current AI accelerators. The trillion-dollar headline sounds massive until you realize it might mean thinner profits per dollar. But Huang knows something about market timing. When the GPU king says he's building adjacent kingdoms, people building agent infrastructure should pay attention. Every new Nvidia product line is a bet on where compute demand goes next.
The Implication
Watch what Nvidia launches in the next six months. If you're building in the agent economy, those product announcements are your roadmap to where the infrastructure layer is headed. Huang doesn't telegraph moves this publicly without conviction. The compute requirements for agentic AI are different from training LLMs. Nvidia's "new markets" push is them front-running that shift.
Sources: Bloomberg Tech | Bloomberg Tech