The advent of so-called physical artificial intelligence is seen by many as the next frontier in the AI investment cycle, opening up opportunities to support robot makers, automotive OEMs and technology companies. specialized semiconductors. Investors are looking for another way to play AI after huge runs in Nvidia and the other stocks powering and producing digital robots like ChatGPT. Search bots were the beginning and the technology has now moved to “agentic AI” like customer service representatives and search assistants. Now, investors analyzing the forefront of this technology are focusing on interacting in the physical world with devices that act independently thanks to artificial intelligence, such as nursing robots, drones and self-driving cars . And analysts believe that’s where the big money will be. “As we approach 2025, agentic AI is the next inflection point before we reach the moment of physical AI… like everything else in the world, you have to crawl before you can walk then run,” said CFRA senior equity. analyst Angelo Zino. “The AI generation hasn’t really seen a lot of monetization so far. So when you start thinking about agentic AI and robotics and everything else, there’s a real kind of belief – and d You’ll see these use cases start coming about this year, as well as Mobileye, a maker of driver assistance technologies, and Aptiv, an auto parts supplier, could benefit from this. futuristic trend. The rise of humanoid robots also has several big-name players, which we’ve already talked about, like Tesla and XPeng. And of course, Nvidia is also a big leader in physical AI, according to several Wall Street analysts. expects realistic AI agent applications to start coming out this year and physical AI innovations to start showing up later, most likely first in self-driving cars. Physical AI has created “an extra leg” for the AI development phase, said Toshiya Hari, an analyst at Goldman Sachs. “There is now an additional engine that augments what is happening today and AI in general.” Nvidia and the ‘chatGPT moment’ for robotics Nvidia has been the darling of AI for as long as investors can remember, as it owns the majority of the AI chip market and continues to be a notable name in the field of agentic AI. But its efforts in robotics have yet to catch up with Wall Street’s short-term estimates for the stock. Nvidia founder and CEO Jensen Huang’s keynote speech at CES on January 6 left investors and analysts optimistic about the company’s announcements regarding its hardware and software products. Huang touted during his keynote that “the ChatGPT moment for general robotics is fast approaching,” and said the company hopes to enable the development of three key robots in the coming years: agent agents, autonomous vehicles and humanoid robots. Huang devoted much of his talk to Nvidia’s development of agent models capable of performing tasks such as coding and chat support. He also announced Nvidia’s Cosmos platform for companies looking to train and develop physical AI systems. Core models of the Cosmos world can generate physics-based videos from multiple inputs, such as video, image text, as well as sensor and motion data from the robot, according to a recent release from Nvidia press. Prominent robotics and automotive companies, such as XPeng, ride-hailing giant Uber and Agile Robots, have already adopted the platform, the company said. Several humanoid manufacturers have also adopted Nvidia’s Project GR00T model. In an interview, Goldman’s Hari said he was “struck” by the Cosmos announcement and the pace at which Nvidia is innovating. His Buy rating and $165 price target are driven by sales of Nvidia’s Blackwell GPUs and the growth of generative AI-powered data centers, as well as continued cloud and enterprise spending, which he expects to see throughout 2025 and 2026. Although physical AI is not a large part of Hari’s thesis on Nvidia, the analyst said that “the concept “physical AI would lead to increased confidence in this company, growing year over year” in its 2027 report. and 2028 estimates. CFRA’s Zino and Bank of America senior analyst Vivek Arya also remain optimistic about the Nvidia’s broader footprint in the AI space, especially after Huang’s keynote. But they also reiterated that there are many unknowns about how long the company’s robotics efforts can translate into financial results. Arya said that Nvidia’s approach of being the brains behind physical AI, rather than the maker of true self-driving, humanoid cars, puts the company in a situation “where the biggest profit reserves lie,” as consumer device makers ultimately face lower margins and greater competition. Zino said Nvidia’s existing group of customers and partners should ultimately allow it to “continue to really gain traction as physical AI becomes more of an opportunity.” He added that “it’s too early to really say how this will all evolve, but Nvidia will clearly be a winner here.” Other emerging winners from ‘physical AI’ Several companies involved in the technologies that power physical AI systems, such as makers of sensors and control systems, are also expected to benefit from this trend. Oppenheimer analyst Colin Rusch named self-driving technology developers Aptiv and Mobileye, warehouse automation company Symbotic and high-energy lithium-ion battery maker Amprius Technologies among top picks. “We view physical AI as a nascent interdisciplinary field poised to transform the industrial complex through increased asset productivity and labor efficiency,” Rusch said in a recent note. “We see outsized ROI potential in the areas of sensing, sensor fusion, compute hardware and software architecture, connectivity, controls, and AI training strategies,” he said. he continued, adding that he expects maturing technology to benefit companies with strong balance sheets and access to progressive technologies. investment capital. Auto parts supplier Aptiv is an important player in the sensor space, as the company provides radar and sensor fusion technologies that help build “a truly comprehensive and reliable environmental model,” the analyst said. Sensor fusion technologies combine data from radar, lidar and camera sensors to create a detailed picture of the driving environment. Intel-owned Mobileye is also one of Oppenheimer’s top picks when it comes to physical AI software architecture. “Even though some vertically integrated players will likely emerge in an autonomous future, we believe MBLY will be the primary enabler for most existing OEMs offering autonomous capabilities,” Oppenheimer said in a note. Rusch highlighted Mobileye’s “cutting-edge” expertise in sensors and an AI-enabled system approach for autonomous driving systems. MBLY 1Y Mountain Mobileye stock performance over the past year. Amprius, another Oppenheimer physical AI pick, has seen its shares soar about 105% over the past month, fueled by the company’s Tuesday announcement of its SiCore cell, part of a platform that provides high-performance energy and electrical performance for drones. , electric vehicles and aeronautical applications. The company’s price target suggests a whopping upside of over 316% over the next year. There is no clear timeline for when the physical AI opportunity will emerge and most analysts are still struggling to incorporate it into their short- to medium-term models. “A year ago, most of these companies, most of these great minds were talking about reaching this point within three years or maybe sooner, some a little later,” Zino said . This is “where we start to see a multi-billion dollar TAM opportunity on the physical AI side.”
