WCIT, the World Congress on Innovation and Technology, took place last year in Malaysia. This year, it was held earlier this month in Yerevan, Armenia.
Rev delivered a fascinating keynote that differed slightly from Jensen’s series of keynotes in 2025. His focus was on what he calls Physical AI and robotics.
Accelerated Computing
He began by educating the audience on the importance of GPUs for accelerated computing and AI to meet the growing demand. He emphasized the significance of the full-stack approach that Nvidia has adopted in recent years, which integrates systems, data centers, software libraries, and algorithms. Nvidia was an early pioneer in AI, launching ahead of much of its competition in 2012, with an earlier breakthrough in image recognition through AlexNet.
The Future: Physical AI and Robotics
This segment was particularly intriguing. While Jensen has touched on this topic often, Rev delved deeper into the details. Today, most AI applications focus on generative models and chatbots that assist with various tasks. However, Rev believes the real opportunity lies in applying AI to the physical world through robotics. Nvidia is already using AI to train robots, simulate driving, and develop autonomous and assisted vehicles. Robots are becoming a critical part of the company’s future.
Lebaredian broadly defines robots as systems that can perceive, reason, and interact with the world. This definition includes not only humanoid robots but also self-driving cars, automated systems, and even smart buildings.
According to Lebaredian, robots require three types of computers. The first is the AI computer, which trains the robot’s brain using Nvidia GPUs. The second is the simulation computer, which creates virtual worlds for the robot to learn in, utilizing Nvidia's Omniverse platform. Omniverse enables real-time collaboration between individuals or teams on 3D designs and simulations, which trained robots can also leverage.
Finally, the robot's own computer runs the AI brain and controls its actions as it interacts with the world. It sends commands to motors to generate specific actions, such as walking or grabbing objects. Humanoid robots, in particular, are well-suited for interacting with human environments and can learn by observing human demonstrations. This represents a significant opportunity that many are investing in.
The primary challenge in building humanoid robots lies in the complexity of improving perception, dexterity, balance, and safety. Recent AI breakthroughs are making these advancements possible.
Ultimately, Lebaredian argues that we are on the brink of a new industrial revolution driven by AI and robotics, with Nvidia leading the way by providing the tools and technologies to make this a reality. He also envisions a bright future for Armenia in this revolution, leveraging its talent and focusing on key areas within AI.