I last wrote about NVIDIA ((NVDA – Free Report) ) as the Bull of the Day on January 29 in the immediate aftermath of the DeepSeek invasion.
I explained why buying near $120 offered excellent risk/reward to accumulate more shares of the primary engine of the AI revolution. That play gave investors and traders a quick 15%+ gain on the NVDA rally back to $140.
I then fleshed out the long-term thesis in this article from Feb 18…
DeepSeek and the 5th Industrial Revolution
But then a bigger correction unfolded for technology stocks as the Nasdaq 100 plummeted 13% in just 3 weeks.
So what did I do? I added more NVDA shares on the dip under $110.
Because the biggest NVIDIA catalyst event of the year starts today with CEO Jensen Huang’s keynote address at the GPU Tech Conference (GTC) in San Jose.
While I could recap the broad investment case for owing the “engine of AI,” the above article has already done that with a 5-year vision.
So I thought it would be fun to profile a unique new product line that Jensen unveiled at CES in early January. We are certainly due to hear more about it today. Here are excerpts from a recent X thread I wrote @KevinBCook.
Introducing Project DIGITS
At CES, Jensen wowed the crowd with something amazing: a personal AI supercomputer for your desktop powered by Blackwell.
This magic little box, not much bigger than your favorite paperback novel, can build models of up to 200 billion parameters.
That means it’s more like a few bars of palladium on your desk. And it’s only $3,000. Let’s dive in!
If you see a picture of Project DIGITS, it actually looks like a mini DGX box. But it’s locked and loaded with the new GB10 Grace Blackwell Superchip. Plus CUDA software and libraries.
And it wields a petaflop of AI computing performance for prototyping, fine-tuning & running large AI models.
“AI will be mainstream in every application for every industry. With Project DIGITS, the Grace Blackwell Superchip comes to millions of developers,” said Jensen Huang, founder and CEO of NVIDIA.
At the big CES reveal, you couldn’t even buy it yet. I’m betting the wait list has grown long and today we could find out when they ship.
“Placing an AI supercomputer on the desks of every data scientist, AI researcher and student empowers them to engage and shape the age of AI,” Jensen added.
Tools for the New World of Physical AI
On his roadmap in January, Jensen explained the NVIDIA vision going from Generative-AI to Agentic-AI to Physical-AI.
In my “5th Industrial Revolution” article, I give some examples of Agentic AI where we can use software tools that work, build, and create for us automatically, thereby multiplying our time, energy, and creative potential.
But Physical AI gets even crazier as robotic systems can embody AI tools and software to conduct physical tasks for us. This takes the best of what NVIDIA offers to create training environments for, as an example, humanoid robots.
Once developers simulate a real-world situation, NVIDIA Omniverse and Cosmos can train the robot on billions of iterations to perform complex tasks for us… like folding laundry or even cooking a gourmet meal.
Now here’s where it gets exciting for the potential of DIGITS. Each Project DIGITS features 128GB of unified, coherent memory and up to 4TB of NVMe storage. And you just plug it in to a wall outlet.
I have to believe that that DIGITS will give the power to small developers to build their own models for Physical AI. We may learn more about that today.
What I know for sure, is how DIGITS will empower those building Agentic AI applications from home or lab to go beyond most platforms.
From the company press release on Jan 6 from CES:
“To build agentic AI applications, DIGITS users can also harness NVIDIA Blueprints and NVIDIA NIM™ microservices, which are available for research, development and testing via the NVIDIA Developer Program.”
A big key to NVIDIA success for the past two decades is evangelizing developers by giving them power tools in the form of GPU hardware and CUDA software to build just about anything they can imagine.
And that support only keeps expanding. With DIGITS, when AI applications are ready to move from experimentation to production environments, the NVIDIA AI Enterprise license provides enterprise-grade security, support & product releases of NVIDIA AI software.
Be sure to catch Jensen’s keynote at 10 am Pacific today. You won’t be disappointed when the Wizard of AI shows off his latest toys and magic.
Financial Market Newsflash
No financial news published today. Check back later.