Robbie Whelan

But now that we're in this new world where training is less emphasized and there are suddenly dozens and dozens of competitors who can provide solutions that are sufficient or that are comparable to what Emidia has, I don't see how they can maintain their margins for forever.

Yeah, I got a message from an investor who I know, and he said, Vera stole the show in that call.

And what he's talking about is a type of chip called the Vera chip.

It's a CPU.

And to back up for a minute, a CPU stands for central processing unit.

It's the sort of most basic computer brain behind pretty much any computer in the world.

But what NVIDIA has historically specialized in is not a CPU.

It's something called a GPU, a graphics processing unit.

And that's what allows people to operate these gigantic models that are sort of the bedrock of all the AI technology we use today.

But there's this funny thing that's happened, which is that there's a new type of AI called agentic AI.

And that means designing these tools called agents that can go schedule your dry cleaning pickup for you or be a sort of digital personal assistant for you or even write software code while you're out taking a yoga class or whatever.

These AI agents, they require many more of these basic chips called the CPUs.

NVIDIA does design some of these and sell them, but they're a lot cheaper than their complex GPUs.

They're a lot less expensive.

They produce a lot less revenue.

That's yet another shift underway right now that's very dramatic and very important for the future of AI companies and for chip companies.

Yeah.

So what those three companies all have in common, besides the fact that they're all about to go public and raise a huge bundle of money, is that they'll buy chips from NVIDIA.

So when we see this news about Anthropic racing to profitability a lot faster than anyone expected, OpenAI and SpaceX, all three of them really moving towards IPO.

If those companies are going great guns, that's great news for NVIDIA.