Just realized something interesting about Nvidia that most people are probably sleeping on. Everyone's obsessed with the data center dominance, but there's this whole physical AI angle that could be absolutely massive.

So here's the thing - Nvidia just reported $215.9 billion in fiscal 2026 revenue, and yeah, data centers are carrying basically the entire company at 89.7% of that. But then Colette Kress (their CFO) casually drops that physical AI already hit north of $6 billion in revenue. That's less than 3% of total revenue, but that number is way more interesting than it sounds at first.

They're not just talking about this theoretically either. Waymo, Tesla, Uber, WeRide, Zoox - all these robotaxi companies are already running on Nvidia's chips. And we're talking thousands of vehicles scaling to millions over the next decade. The compute demand alone for that is insane.

What caught my attention was the Rubin architecture rollout. This isn't just an incremental chip update - it's specifically designed for what Jensen Huang called the "agentic AI inflection point." The Vera Rubin NVL72 is basically a plug-and-play supercomputer with 72 GPUs, 36 CPUs, and full networking infrastructure. They're building hardware specifically for autonomous systems, not just general compute.

The way I see it, this is the strategic play. Generative AI got them to where they are now, agentic AI will keep them growing, but physical AI is where the greater long-term potential lives. Hundreds of billions in revenue potential isn't just hype - it's what happens when you need exponentially more compute for millions of autonomous vehicles plus industrial robotics across every sector.

The beauty is they don't even need physical AI to explode soon. They're already printing cash from data centers and agentic AI adoption. But if robotaxis and general robotics actually scale the way everyone expects, Nvidia's positioned to capture that entire wave. That's the kind of optionality that makes this thesis compelling for the next decade.

Worth keeping on the radar if you're thinking long-term growth plays.