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 human history.

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

Wanwan has sorted out three big highlights. Come on, fat friends—follow me.

1）AI compute costs cut straight to one-tenth

The previous-gen Blackwell is already pretty powerful, right?
Next up, they’re about to announce mass production of the new-gen Vera Rubin chips.

Where is Vera Rubin strong? To put it bluntly: two words—cheap.

Run the same AI model,
the number of chips drops to a quarter, and inference compute cost falls by 90%.
Falls by 90%, my friends.
AWS, Microsoft, and Google—the three major cloud providers—will get on the first batch of vehicles.

2）Groq, bought for $20 billion last year—handing in the assignment today

At a previous earnings call, Huang Renxun said that Groq would be integrated into the Nvidia ecosystem as an expansion architecture—just like how they acquired Mellanox back then to round out their networking capabilities.

Groq’s LPU and Nvidia’s GPU sit in the same data center. The GPU understands the problem, and the LPU rapidly spits out the answer.

With these two chips working in tandem, Agent scenario latency drops straight down.

With AI Agents doing human work, a single task can go back and forth and involve dozens of model adjustments—every round burns inference compute, and the user is just sitting there waiting. If it’s even a bit slower, the experience collapses.

Inference happens in two steps: first understand your question, then output the answer word by word.

The GPU is good at the first step, but for the second step—the speed and stability of output—Groq’s LPU is stronger.

Is $20 billion expensive?

Think about it: later, every company will run hundreds of Agents, and each Agent will adjust the model thousands of times every day.

3）Nvidia’s OpenClaw version launches—called NemoClaw

It’s basically an open-source platform. Once enterprises install it, they can deploy AI employees to run processes for humans, handle data, and manage projects.
They say it’s already in talks with Salesforce and Adobe.

The interesting part is that NemoClaw doesn’t require you to use Nvidia chips.
Think about that logic.

Selling chips makes money only on the hardware layer—only by setting the rules can you earn across the whole chain. Huang Renxun has done the math perfectly.

4）Huang Renxun says he wants to showcase “chips the world has never seen before”

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

Also, there’s another less-talked-about tidbit I think is pretty interesting.

Nvidia has made laptop computer processors—two of them—focused on gaming.
People selling graphics cards are coming to grab the CPU’s share of the business.

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