New York, New York
— March 15, 2026

Microsoft Validated NVIDIA Vera Rubin in the Cloud. Now It’s GTC Week.

NVIDIA’s GPUs were the first to reach space. Now, its data center stack is scaling so fast it’s starting to look less like “chips” and more like critical infrastructure.

This week we got a fresh signal that NVIDIA’s next platform is moving on schedule: Satya Nadella said Microsoft is the first cloud provider to bring up an NVIDIA Vera Rubin NVL72 system for validation.

If you’re an investor, this is exactly what you want to see. “Validation” isn’t mass deployment, but it’s a meaningful milestone in the rollout pipeline. It suggests the hardware is far enough along for hyperscalers to begin testing, tuning, and planning real capacity around it.

And that timing matters because Jensen has been pointing to Rubin deployments in the back half of the year, so Microsoft being early to the validation step supports the “on track” narrative heading into GTC.

---

GTC kicks off. Here’s what we’re watching.

Software: will Jensen lean into it harder?

NVIDIA has been pushing back on the idea that software is “dead” or that we’re headed into a SaaS-pocalypse. GTC is the perfect stage for Jensen to reframe the narrative: that software isn’t going away, it’s just changing shape into agentic workflows, copilots, and AI-native platforms.

Which ties directly into the next rumor…

NemoClaw: an agentic software platform for enterprises?

There’s credible reporting NVIDIA may debut an open-source AI agent platform called “NemoClaw,” positioned for enterprise use, with security and governance in mind.

If this shows up at GTC, the subtext is bigger than the product. It’s NVIDIA saying:

• “We’re not just selling GPUs.”
• “We want to own the developer surface area for agents.”
• “We’re going to play offense in software, even if the market thinks it’s all about infrastructure.”

Physical AI: robotics and autonomous vehicles

We also expect more emphasis on physical AI, meaning robots and autonomous systems that operate in the real world.

This is where the next demand wave can come from. The first mega-wave was LLM training and inference. The next wave is compute for agents and machines that act. Reuters has flagged that NVIDIA plans to highlight agentic AI, networking, and AI factory infrastructure at GTC, alongside new product and partnership updates.

Telco updates and the “plumbing”

And yes, we’re watching the less flashy stuff too: telco updates, networking, optical interconnect, and anything that expands NVIDIA’s role in the full AI factory (not just the GPU line item). That’s where moats get built.

---

Bottom line

Microsoft validating Vera Rubin NVL72 is a clean checkpoint: it suggests NVIDIA’s next platform is moving from roadmap to reality inside the biggest cloud ecosystems.

Now it’s GTC week, and we’re watching for NVIDIA to press its advantage across:

• Software (agents),
• Physical AI (robots + AVs),
• AI Infrastructure stack that powers it all.

We’ll have more after we hear from them. Be good!