Jensen Huang founded accelerated computing company Nvidia (NVDA -0.02%) in 1993, and has served as the CEO and president ever since. Nvidia has achieved many breakthroughs under his leadership, but the invention of the graphics processing unit (GPU) in 1999 was particularly momentous.
Nvidia GPUs have long been the gold standard in rendering graphics for 3D design and gaming applications. More recently, they have become the chips of choice for complex data center workloads like training large language models and running generative artificial intelligence applications.
Jensen Huang last week gave a keynote speech at CES 2025, an annual conference held in Las Vegas. CES focuses on innovations within the technology sector, and Huang's speech made it clear that Nvidia's product pipeline is still bursting with potential.
Generative AI is just the first phase of a technological revolution
Wall Street's interest in generative artificial intelligence can be traced back to the launch of ChatGPT in late 2022. The conversational application went viral almost immediately, setting in motion a series of events that caused a tremendous increase in demand for Nvidia GPUs. The company has reported triple-digit earnings growth in the last six quarters, and its share price has increased 840% in the last two years.
Some investors worry generative AI is a short-term catalyst or even a bubble, but nothing could be further from the truth. The internet has only become more essential since it was created, and artificial intelligence will be no different. In other words, the generative AI boom is just the first phase of a technological revolution that will continue indefinitely. And Nvidia sits at the heart of that revolution.
Physical AI is the next phase of the technological revolution
Jensen Huang at CES said, "The next frontier of AI is physical AI." Whereas generative AI can understand and generate media, physical AI can understand, navigate, and interact with the physical world. It will eventually power numerous types of autonomous robots, but the first type of intelligent robot most people engage will be an autonomous car, according to Huang.
Importantly, Nvidia has products that address all three layers of the autonomous vehicle computing stack: Its GPUs provide the supercomputing infrastructure needed to train AI models. Its Drive platform provides the software development tools required to build self-driving applications. And its AGX systems provide the in-vehicle computing power that lets cars navigate the physical world.
Jensen Huang in a recent interview with Yahoo Finance said Nvidia's autonomous driving products may reach a revenue run rate of $5 billion this year, up from $1.8 billion in the last quarter. Importantly, automotive and robotics is currently the company's smallest segment, but that could change quickly. Citigroup estimates the number of autonomous vehicles will increase almost 5-fold by 2030 and 14-fold by 2035.
Image source: Getty Images.
Nvidia is ideally positioned to be a leader in AI robotics
Jensen Huang at CES told the audience, "The ChatGPT moment for robotics is coming." He then introduced Cosmos, a suite of pretrained robotics models that can be fine-tuned by developers. Huang also explained that Nvidia has products that address all three layers of the robotics computing stack.
First, Nvidia GPUs provided the supercomputing infrastructure needed to train robotics models. Second, the Isaac platform includes code libraries and pretrained models that help engineers develop robotics applications across three use cases: industrial manipulation arms, autonomous mobile robots, and autonomous humanoid robots. Isaac also works as a simulation engine that supports synthetic data generation and the evaluation of robotics models.
Finally, Jetson embedded systems bring together GPUs, CPUs, and memory on a single chip, which provides the computing power robots require to interact with the real world. Nvidia is clearly well positioned to benefit as the AI boom evolves into a robotics revolution, and humanoid robots could be a massive opportunity for the company. Citigroup estimates spending will top $200 billion by 2035 and $1 trillion by 2040.
Nvidia stock is cheaper today than it was two years ago
Many investors assume Nvidia is a very expensive stock because it returned 840% in the last two years. But shares trade at 55 times earnings, a reasonable valuation given that Wall Street anticipates earnings growth of 38% annually in the next three years. Those figures give a price-to-earnings-to-growth (PEG) ratio of 1.4.
Comparatively, the stock traded at 63 times earnings two years ago, and Wall Street at the time anticipated earnings growth of 22% annually. Those figures give a significantly higher PEG ratio of 2.9.
So, Nvidia stock is actually much cheaper today than was before the generative AI boom started. And given the tailwinds in autonomous driving and robotics that Jensen Huang highlighted at CES, the stock is still a worthwhile long-term investment.
