The robot revolution is here, and it’s not just about sci-fi fantasies anymore. Imagine a world where machines don’t just think, but act—folding laundry, washing dishes, and even caring for the elderly. This is the promise of physical AI, and it’s closer than you think. But here’s where it gets controversial: as these intelligent machines step out of the digital realm and into our homes, workplaces, and streets, they’re sparking debates about job security, privacy, and the very nature of human labor. And this is the part most people miss: it’s not just about the technology itself, but how we adapt to living alongside it.
Physical AI, as Jensen Huang, CEO of Nvidia, puts it, is ‘AI that understands the laws of physics, AI that can work among us.’ It’s about machines that don’t just process data but interact with the physical world—driving cars, assisting in care homes, and performing tasks that once required human hands. Tech giants are pouring billions into this field, with Morgan Stanley predicting over a billion humanoid robots by 2050. Videos of Chinese-made androids dancing to pop music or effortlessly lifting heavy objects have only fueled the excitement. But beyond the flashy demos, the real question is: how soon will these robots become genuinely useful, and at what cost?
Take Enactic, a Tokyo-based company led by 24-year-old CEO Hiro Yamamoto. Their humanoid robots are designed to handle mundane tasks in short-staffed Japanese care homes, freeing up human workers to focus on more meaningful interactions. Yamamoto emphasizes safety, equipping the robots with soft exteriors to prevent injuries. ‘It has to be safe,’ he says, ‘because these robots will live alongside people in chaotic, ever-changing environments.’
In China, XPeng, an electric vehicle maker, recently unveiled a humanoid robot with a glowing visor and athletic build, drawing cheers and whispers. XPeng’s CEO, He Xiaopeng, boldly predicts that robot sales could one day surpass car sales. But while their robots can walk and dance autonomously, handling objects—a far more complex task—remains a challenge. ‘Their dexterous fingers and flexible skin won’t replace factory workers anytime soon,’ He admits. The cost of a single robot hand, for instance, could pay a Chinese worker’s salary for years.
Yet, XPeng co-president Brian Gu envisions a future where AI humanoid robots could perform ‘almost any human role,’ from nannying to gardening. But how realistic is this? Physical AI faces unique challenges compared to text-based AI like ChatGPT. While the latter is trained on vast amounts of text, physical AI must master vision, spatial awareness, and real-world interactions. Enactic’s approach involves teleoperated robots learning tasks through remote demonstrations—just 30 to 50 repetitions are needed to fine-tune their ‘vision-language-action’ models. This on-the-job training is crucial for robots to eventually act autonomously.
US-Norwegian startup 1X is taking a similar path with NEO, a humanoid home helper set to launch next year. Priced at $20,000, NEO’s performance is still shaky, as seen in a viral video where it struggles to close a dishwasher door—even with human assistance. These moments highlight the ‘big gap’ between AI systems and physical capabilities, as noted by Sara Adela Abad Guaman, a robotics professor at University College London. ‘Nature has shown us that adapting to the environment requires the right body,’ she explains, using the example of a mountain goat navigating icy terrain.
Despite these challenges, investments in physical AI are booming. Japan’s SoftBank recently acquired ABB Robotics for $5.4 billion, calling physical AI its ‘next frontier.’ But as automation advances, concerns about job displacement grow. Abad, however, remains optimistic. ‘Our sense of touch is incomparable,’ she says, suggesting that humans will always have unique skills machines can’t replicate.
So, what do you think? Will physical AI revolutionize our lives, or are we moving too fast without considering the consequences? Are we ready to share our spaces with robots that can think, act, and learn? Let’s spark a conversation—agree or disagree, the future of AI is ours to shape.