LAS VEGAS – NVIDIA Corp. CEO Jensen Huang, the de facto grand marshal of CES 2026, officially kicked off the Super Bowl of tech events here Monday with a keynote heavily steeped in artificial intelligence (AI) predictions and products highlighted by new autonomous vehicle software called Alpamayo and a robotics platform.
The software helps self-driving cars make navigation decisions while creating documentation trails for engineers. Huang said NVIDIA will open-source both the models and training data to enable automaker evaluations and build trust in how the AI systems were developed.
“Not only do we open-source the models, we also open-source the data that we use to train those models, because only in that way can you truly trust how the models came to be,” Huang said before about 3,000 gathered at the Fontainebleau Hotel. “It is impossible to collect every possible scenario for data on driving.”
Alpamayo uses what NVIDIA describes as chain-of-thought reasoning combined with vision language action modeling. The technology enables vehicles to identify unexpected driving conditions — such as a malfunctioning traffic light at an intersection — analyze the situation, and determine an appropriate response.
NVIDIA is positioning Alpamayo as a foundational teaching tool that developers can customize and integrate into their autonomous driving systems. The company said Lucid, Uber Technologies Inc., and Berkeley DeepDrive have expressed interest in the technology.
NVIDIA said its Drive AV platform will power assisted driving features in the 2026 Mercedes-Benz CLA, marking the system’s entry into production vehicles. The CLA earned the highest safety rating from Euro NCAP among all vehicles tested in 2025. According to NVIDIA, the vehicle will offer Level 2 automated driving capabilities, including point-to-point urban navigation that lets drivers input specific addresses as destinations. The technology relies on what NVIDIA calls its Hyperion architecture, a compute-and-sensor system designed to provide redundant safety features.
The company projects the CLA will support hands-free driving on U.S. roadways before the end of this year.
The launch addresses persistent challenges in the self-driving industry. While companies like Google’s Waymo have deployed autonomous vehicles commercially, the technology continues to encounter problems, including traffic disruptions and confusion in complex driving scenarios.
NVIDIA is emphasizing virtual training environments as a solution, allowing developers to refine their AI systems through simulation rather than relying solely on real-world road testing. The approach could accelerate development while reducing the risks and costs associated with on-road experimentation.
Additionally, NVIDIA unveiled a comprehensive robotics platform in an ambitious play to dominate the emerging humanoid robot industry. Huang said major manufacturers Boston Dynamics, Caterpillar, LG Electronics, and NEURA Robotics have adopted NVIDIA’s technologies to develop their robotic systems.
The announcement positions NVIDIA at the center of the tech industry’s accelerating push to commercialize humanoid robots for industrial applications. The company estimates the potential market in manufacturing and logistics alone at $50 trillion.
The company’s robotics offering combines new AI models for training robots to navigate real-world environments with specialized computing hardware to power them. The integrated approach allows the company to capture revenue across multiple segments of the robotics supply chain. The company has branded its technology physical AI to distinguish systems that interact with the physical world from software-only artificial intelligence applications.
The announcements come as the AI chip-making behemoth faces competitive pressure on multiple fronts. Last month, the company acquired talent and chip technology from startup Groq, including executives who helped Alphabet’s Google design proprietary AI chips. Google’s chips now represent a significant threat to NVIDIA’s dominance, particularly as Google collaborates with Meta Platforms Inc. and others to challenge NVIDIA’s AI market position.
Meanwhile, NVIDIA is working to demonstrate that its latest products outperform older chips like the H200, which President Donald Trump has permitted for export to China. Reuters reported strong Chinese demand for the H200 — predecessor to NVIDIA’s current Blackwell flagship chip — is raising concerns among China hawks across the U.S. political spectrum.
“We believe the robotics and autonomous technology market represents an incremental market opportunity that NVIDIA can tap into which speaks to our view that this company will exceed $5 trillion market cap in the near-term and ultimately could be a $6 trillion market cap,” Wedbush Securities analyst Dan Ives said in a note to investors following Huang’s eagerly awaited keynote address.
