The era of humanoid robots is fast approaching, and artificial intelligence (AI) is finally making it possible to program machines for general-purpose tasks, according to Nvidia’s Rev Lebaredian.
Speaking to Euronews Next during the Computex technology fair in Taiwan, Lebaredian, vice president of Omniverse and simulation technology at Nvidia, described robotics as the “next phase” of AI — a development poised to help ease global labour shortages, especially in industrial sectors.
“For decades, robotics has been the stuff of science fiction,” Lebaredian said. “We’ve long been able to build the physical machines, but the programming part has always been the challenge. AI changes that.”
Companies like Tesla have already made headway, with its Optimus robot reportedly able to carry out household chores. But Nvidia believes true progress lies in virtual training. According to Lebaredian, humanoid robots should first learn in simulated environments — both for safety and efficiency.
“AI is data hungry. Large language models can be trained on vast amounts of online data. But robots don’t have that advantage — there isn’t a massive repository of physical-world data,” he said. “So we must simulate it.”
Simulated environments allow developers to feed robots “renewable” data, creating countless experiences without real-world risks. Once a robot performs well in simulation, it can then be deployed in the real world — much like a graduate entering the workforce, who then trains on specific, company-related knowledge.
The first real-world applications for humanoid robots, Lebaredian believes, will be in factories and warehouses, where workforce shortages are most acute. With many countries facing aging populations and a shrinking pool of workers, particularly in physically demanding or hazardous jobs, robots could play a vital role in sustaining productivity.
“Industrial use will come first because the need is real,” he said. “In every country, skilled workers are retiring and not enough young people are replacing them.”
Taiwan has already announced a five-year plan to invest in robotics to combat its own population challenges, highlighting a growing global trend.
Looking ahead, Lebaredian sees potential roles for robots in retail, mining, hazardous environments like nuclear reactors, and even in caregiving roles for the elderly — if public demand aligns.
Despite the excitement, concerns remain over safety and reliability. Lebaredian acknowledged that while AI models like chatbots still make mistakes, robotics offers a more measurable framework.
“Did the robot pick up the object and place it safely? That’s a binary outcome — and one we can test, measure, and improve,” he said. “We’ve built nuclear reactors safely. We can build safe robots, too.”
With AI-driven training, safety testing, and advancing simulation, the integration of humanoid robots into society may be closer than many think.