There is no stopping Jensen Huang. After sovereign AI, NVIDIA’s chief has a newfound obsession, mostly revolving around self-driving cars and AI agents. Speaking at CES 2025, he highlighted the potential of robotaxis and autonomous vehicle (AV) technologies to transform global mobility and logistics.
“I predict that this will likely be the first multi-trillion dollar robotics industry,” he said in his keynote. This claim comes amid a growing wave of innovation, investment, and competition in the autonomous driving landscape.
Nvidia CEO Jensen Huang tonight on robotaxis and self-driving: “I predict that this is going to be the first multi-trillion dollar robotics industry.” pic.twitter.com/QVTnwhdMUY
— Sawyer Merritt (@SawyerMerritt) January 7, 2025
Robotaxis show promise in reducing traffic congestion, lowering emissions, and providing mobility solutions that are accessible to millions worldwide.
Moreover, the integration of autonomous vehicles into public and private fleets is expected to create millions of new jobs in AI development, system integration, and maintenance, further cementing the industry’s economic impact.
Breakthroughs in AI platforms, sensor technologies, and cloud computing are backing the push for autonomous mobility. Companies like NVIDIA have introduced platforms such as DRIVE Hyperion and Cosmos to streamline the development of AV systems.
DRIVE Hyperion enables autonomous vehicle manufacturers to handle perception, mapping, and decision-making efficiently, while the recently launched Cosmos platform generates synthetic driving environments for training AV algorithms.
Using tools like AI traffic generators and neural reconstruction engines, Cosmos creates high-fidelity 4D simulations, turning hundreds of real-world drives into billions of effective miles for training data.
Real vs Synthetic Data Debate
The role of synthetic data in training AV systems has sparked debate among industry experts. Sawyer Merritt, an investor at Tesla, referred to NVIDIA’s Cosmos and pointed out that synthetic data, while innovative, cannot replace the reliability of real-world video driving data.
“Synthetic driving data is like using ChatGPT—you might trust what you see is true, but you often can’t be entirely certain without further validation,” he said.
Merritt emphasised Tesla’s unmatched advantage: over 7.1 million vehicles on the road worldwide, collectively driving upwards of 75 billion miles annually, with more than 56 million onboard cameras capturing real-world video data.
The company’s cars capture real-world driving scenarios, offering unmatched insights for self-driving.
In a recent conversation with AIM, Sami Atiya, president of the robotics & discrete automation business area at ABB, also expressed his views on the subject. He believes synthetic data will play a huge role in robotics as the company is already using it for “arm simulations and complex path-planning” for its in-house robots.
Synthetic data opens up endless possibilities without ever needing to touch the robot. But he also reminded us to be wary of any biases or misleading elements in the data. “The main expertise of the people who actually use these AI systems will become much more crucial to know the right input and output of data that is not biased,” Atiya said.
He agreed with Ilya Sutskever on the end of traditional pre-training due to data limitations, emphasising AI’s reliance on scaling models and exploring agents and synthetic data as the future of AI innovation. “We will reach a plateau, and we are about to see more capacity being thrown at systems,” he said.
An Industry on the Move
Waymo, Alphabet’s self-driving subsidiary, has been conducting extensive real-world trials and is expanding its operations in cities across the United States and in Tokyo early this year.
Meanwhile, Cruise, a subsidiary of General Motors, has been deploying autonomous taxis in select locations, including San Francisco and Phoenix. At the same time, Mobileye, an Intel company, launched a unique sensor technology for layered visuals, which generates 3D perception for a reliable understanding of the environment.
Mobileye’s latest system on a chip, the EyeQ6, powers this advanced processing, as announced by founder and CEO Amnon Shashua in his keynote address at CES 2025. The tech, which offers high-resolution sensing capabilities that address camera weak spots, will enter production in 2026.
NVIDIA also recently announced collaborations that will shape the future of autonomous vehicles. Toyota, the world’s largest automaker, is building its next-generation vehicles on NVIDIA DRIVE AGX Orin, running the safety-certified NVIDIA DriveOS operating system.
These vehicles will offer functionally safe, advanced driving assistance capabilities. Also, partnerships with companies like Aurora and Continental highlight the widespread adoption of these technologies across legacy automakers.
Other leaders using Cosmos to build physical AI for AVs include Fortellix, Uber, Waabi and Wayve. Such partnerships aim to overcome challenges in autonomous driving and ensure rapid deployment of robotaxis and self-driving fleets worldwide.
For instance, HERE Technologies and AWS recently announced a $1 billion partnership to develop AI mapping solutions critical for precise navigation.
Meanwhile, Uber and NVIDIA recently announced a partnership to support the development of AI-powered autonomous driving technology. More details on this are expected later this year.
Dara Khosrowshahi, CEO of Uber, said in the official announcement, “Generative AI will power the future of mobility, requiring both rich data and very powerful compute.”
With Uber completing millions of trips every day, Khosrowshahi hopes to create safe and scalable autonomous driving solutions for the industry. Now, major companies look forward to pairing up with the NVIDIA Cosmos platform and NVIDIA DGX Cloud to help build stronger AV partners.
Additionally, Amazon and Qualcomm have also announced a collaboration to revolutionise in-car experiences by combining Qualcomm’s Snapdragon Digital Cockpit platform with Amazon’s AI and cloud services.
Last month, Volvo Cars CEO Jim Rowan met Qualcomm president & CEO Cristiano Amon for a lap in the eX90 and a chat about the car as the new computing space.
The Landscape in India, Starting in Bengaluru
Bengaluru, known for its traffic and tech innovation, is emerging as a key player in India’s autonomous vehicle landscape.
The Bengaluru Traffic Police’s exploration of a digital twin for traffic management signals a tech-forward approach crucial for enabling robotaxis and autonomous vehicles.
Bengaluru’s police commissioner envisions a future where AI boosts enforcement and management, serving as a ‘force multiplier’ to meet the city’s unique needs. However, infrastructure upgrades, better rule adherence, and two-wheeler-focused technologies remain critical.
AI-based cameras reduce violation processing time from 300 seconds to 5 seconds. While enforcement leverages rule-based AI effectively, traffic management faces challenges like unpredictable road behaviour, requiring real-time adaptability and instant decision-making for optimal impact.
The post Robotaxis and Self-Driving Cars to be the ‘First Multi-Trillion Dollar Industry’ appeared first on Analytics India Magazine.
