On May 20, IEEE Spectrum published an opinion piece by two industry veterans — Jonathan Hurst, co-founder of Agility Robotics and professor of robotics at Oregon State University, and Hans Peter Brøndmo, former VP at Google X, where he founded and led the Everyday Robots project. Both have years of experience deploying AI-powered robots in real-world settings, and it is this experience that shapes the article’s central argument.
The authors argue that a single breakthrough moment — analogous to ChatGPT’s emergence in late 2022 — is unlikely to happen in robotics. Instead, progress will come through the well-engineered coordination of multiple AI tools: perception models, language models, planning systems, and motor control working in concert.
One of the article’s key points is a warning: “Never trust a YouTube robot video.” The gap between carefully scripted demonstrations and real robot work in unstructured environments remains significant. The authors cite the Unitree humanoid performance at the Chinese 2026 Spring Festival Gala as an example — spectacular, but far removed from autonomous factory work.
Meanwhile, investment in the sector is hitting record levels. In 2025, total funding for robotics companies reached $40.7 billion — 9 percent of all global venture capital. The question is no longer whether AI-powered robots will become part of everyday life, but which path will get them there.
In the authors’ view, the fastest path is not a bet on a single form factor such as humanoids, but the development of diverse robot types performing increasingly sophisticated tasks with a range of AI tools. Over the coming decades, billions of autonomous robots will work alongside people in factories, warehouses, disaster zones, and eventually homes. Some will look like us; most won’t.
Source: IEEE Spectrum — “Will Robotics Have a ChatGPT Moment?”