All eyes are on Nvidia (NVDA -0.02%) this week as the company is hosting its AI Summit in Washington, D.C.
The three-day event will give Nvidia its latest opportunity to show off its AI innovations, including its brand-new Blackwell platform, which is now in production.
The event will feature live demos and hands-on workshops in areas like generative AI, robotics, and industrial digitalization.
The event also seems likely to focus on Nvidia's vision for the future, which centers around "AI Factories," the phrase CEO Jensen Huang likes to use in his vision of the data center of the future, and Nvidia's key role in running those AI factories. Nvidia laid out its plans in a recent investor presentation ahead of the AI Summit.
With Nvidia's market cap now at $3.1 trillion, selling the market on his vision for AI factories and making that transition happen could drive the AI chip star to a valuation of $4 trillion, making it the first company ever to achieve that milestone.
Let's take a closer look at Nvidia's plan to get there.
Image source: Nvidia.
What are AI factories?
Nvidia envisions a new, more advanced form of data center it calls AI factories, the purpose of which is to take data, process it, convert it into models, and finally convert it into usable pieces of information Nvidia calls "tokens." These AI factories will run on Nvidia's accelerated-computing platform.
In this evolving paradigm, companies will produce digital intelligence that serves needs such as customer-service agents, drug discovery, autonomous vehicles, and weather predictions.
Additionally, sovereign AI is becoming another large market for Nvidia, and nations are producing their own artificial intelligence to improve infrastructure, security, and communications.
Nvidia thinks these AI factories will have a significant impact on all industries, from manufacturing to education to financial services. If this turns out to be true, the company is all but set to stay on top.
Why Nvidia can dominate the AI era
Nvidia has a dominant share of the data-center graphics processing unit (GPU) market at an estimated 98% in 2023.
While competition is heating up in that market as AMD launched its Mi300 accelerator and Intel recently introduced the Gaudi 3, Nvidia continues to think expansively with its new Blackwell platform. This platform is four times faster and also decreases power consumption by four times, versus the previous Hopper generation.
Nvidia is also the only chip company that offers a full-stack solution for AI and accelerated computing thanks to its CUDA library of pre-built functions and software that runs on its GPUs. Finally, it's built the leading inference platform, which is accelerated by OpenAI's new o1, the newly released advanced model from the AI start-up.
Can Nvidia keep gaining?
Unrelated to its AI summit this week, Nvidia got some good news from close partner-server provider Super Micro Computer, which said it's currently shipping more than 100,000 GPUs per quarter and noted that its ultradense servers include 8 Nvidia HGX GPUs. It also said its highly regarded liquid-cooling solution can hold up to 96 Nvidia B200 GPUs per rack.
Nvidia shares rose on the news as it reaffirmed that demand for its products remains strong.
While there's been concern among some investors that the cloud-infrastructure giants like Microsoft and Alphabet are overspending on AI, the facts on the ground continue to indicate otherwise, and innovations and tech accelerations like Blackwell should continue to incentivize AI investment.
OpenAI, the leading generative AI start-up, also just nearly doubled its valuation to $157 billion in its latest funding round, in which Nvidia participated, which also shows that the market for AI technology continues to grow.
Huang's vision of AI factories is becoming a reality as companies like OpenAI race to build artificial general intelligence. The AI chip maker could reach a $4 trillion valuation faster than you think.
