The most important thing today is Nvidia's GTC conference, practically an AI version of a brief history of humankind.

Today’s biggest event is NVIDIA GTC, basically an AI version of A Brief History of Humankind.

Jensen Huang hasn’t even taken the stage yet, but the pre-release information alone is enough to fill a book.

Tonight, I’ve summarized three main highlights. Let’s go, friends, follow me.

1. AI Computing Power Costs Cut in Half

The previous generation Blackwell was already impressive, right? Soon, the new Vera Rubin chips will go into mass production.

What makes Vera Rubin so powerful? Simply put, two words: cheap.

Running the same AI models, chip count reduced to a quarter, inference computation costs cut by 90%. Ninety percent reduction, friends. AWS, Microsoft, and Google’s top cloud providers are already on board.

2. The $20 Billion Acquisition of Groq Last Year, Now Delivering Results

Previously, Jensen Huang said at the earnings call that Groq would be integrated into NVIDIA’s ecosystem as an expansion architecture, just like Mellanox was to enhance networking capabilities.

Groq’s LPU (Low Power Unit) works alongside NVIDIA GPUs in the same data center—GPUs handle understanding, LPU quickly delivers answers.

The division of labor between the two chips, combined with agent scenarios, directly reduces latency.

AI agents do the work for people. A task might require dozens of model adjustments back and forth, each burning inference power, and users are waiting. A slower experience could cause a crash.

Inference involves two steps: first understanding your question, then outputting the answer word by word.

GPUs excel at the first step, but for the second—speaking the words—the speed and stability of Groq’s LPU are better.

Is $20 billion expensive?

Think about it—every company in the future will run hundreds of agents, each adjusting models thousands of times a day.

3. NVIDIA’s OpenClaw Launches, Named NemoClaw

It’s an open-source platform that companies can install to deploy AI workers to handle workflows, process data, and manage projects. They’re reportedly already talking with Salesforce and Adobe.

What’s interesting is that NemoClaw doesn’t require NVIDIA chips. Think about this logic. Selling chips only earns hardware-level profits; setting rules allows earning from the entire chain. Jensen Huang has a clear grasp of this.

4. Jensen Huang Says He Will Showcase “Chips Never Seen Before”

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

There’s also an obscure rumor I find quite interesting.

NVIDIA has released laptop processors, two models, aimed at gaming. The company that sells graphics cards is now competing for CPU market share.

Tonight, I feel Jensen Huang is destined to become a great figure in history.