Today’s most important event is NVIDIA GTC Conference, which is basically an AI version of A Brief History of Humankind.

The most important thing today is the NVIDIA GTC conference—basically an AI version of a history of humankind.

Even before Huang Renxun takes the stage, the leaked information is already enough to fill a whole book.

Wenwen pulled together three big highlights—come on, buddies, follow me.

1）AI compute costs are cut directly to one tenth

The previous-generation Blackwell was already pretty powerful, right?
Soon, they’re going to announce mass production of the next-generation Vera Rubin chips.

What’s so powerful about Vera Rubin? To put it simply: it’s cheap.

Run the same AI model,
the number of chips drops to one quarter, and inference compute costs fall by 90%.
By 90%, my friends.
AWS, Microsoft, and Google—the three biggest cloud providers—will be among the first to get on board.

2）The Groq they bought for $20 billion last year turns in homework today

At a previous earnings call, Huang Renxun said Groq would be connected to the NVIDIA ecosystem as an expansion architecture—like back then when buying Mellanox helped round out network capabilities.

Groq’s LPU and NVIDIA GPUs are in the same data center. The GPU understands the problem, while the LPU rapidly spits out the answers.

With these two kinds of chips working in tandem, the latency in Agent scenarios drops directly.

AI Agents do tasks for people. One task can loop back and forth and tweak the model dozens of times; each round burns inference compute, and the user is waiting there. If it’s slower, the experience collapses.

Inference is done in two steps: first, understand your question, and then output the answer one word at a time.

GPUs are good at the first step, but for the second step—the speed and stability of “spitting out words”—Groq’s LPU is stronger.

Is $20 billion expensive?

Just think about it: later, every company will run hundreds of Agents. Each Agent will adjust models thousands of times every day.

3）The NVIDIA version of OpenClaw launches—called NemoClaw

It’s an open-source platform. Businesses install it and can deploy AI employees to run workflows for real people, handle data, and manage projects.
It’s said to already be in talks with Salesforce and Adobe.

What’s interesting is that NemoClaw doesn’t require you to use NVIDIA chips.
Think about that logic.
Selling chips earns you only the hardware layer. Only by setting the rules can you make money across the entire chain. Huang Renxun has clearly already done the math.

4）Huang Renxun says they’re going to demonstrate “a chip the world has never seen before”

It’s highly likely that the next-next generation architecture, Feynman, will make its first appearance—mass production in 2028—with TSMC’s most advanced 1.6nm process.

And there’s another niche piece of intel I think is pretty interesting.

NVIDIA has made laptop computer processors—two of them—focused on gaming.
The company selling graphics cards is coming to grab the CPU market’s dinner.

Wenwen, I feel like Huang Renxun is going to become a great figure of an era in the future.