Three clues do not make proof as in the movies, but they may signal a trend. And over the past year, as already highlighted by the latest edition of the DEVO Lab’s HIT Radar, the signs that humanoid robotics is becoming a topic of widespread relevance, not just confined to research labs, have multiplied in ways that are difficult to ignore.
Signals not to ignore
Media relevance. On February 17, 2026, millions of viewers around the world watched the Chinese New Year Gala on CCTV. The unexpected stars of the show: humanoid robots. A dozen Unitree models performed complex kung fu sequences, including backflips, jumps, and acrobatics requiring real-time fault recovery. The comparison with the previous year, when robots from the same Unitree were limited to twirling handkerchiefs, is emblematic of how quickly technology is advancing. The performance generated more than 23 billion views on Chinese platforms in just three days. On the commercial side, within two hours of the broadcast, robot orders on JD.com had increased by 150 percent, with online searches for “robot” up 300 percent. China deliberately used its most visible stage to send a message to the world: humanoid robotics is no longer a promise for the future.
The rise of robotics at industrial scale. In October 2025, the Telegraph published an article with an emblematic headline: Western executives who visit China are coming back terrified. The piece gathered testimonies from several Western top managers returning from China with a new and unsettling awareness of the country’s technological progress in robotics, both collaborative and humanoid. Jim Farley, CEO of Ford, after visiting a series of Chinese plants, said: “It’s the most humbling thing I’ve ever seen,” referring not only to electric vehicles but to pervasive automation in factories. Greg Jackson, CEO of the British energy supplier Octopus Energy, described the “dark factories”: facilities completely unlit because robots do not need light, where entire assembly lines operate without a single worker. The data confirm this perception: in 2024 alone, China installed 295,000 new industrial robots, compared with 34,000 in the United States and 27,000 in Germany. With 567 robots per 10,000 manufacturing workers, China has surpassed even Germany (449) in robot density.
Market figures suggest this is not just hype. Within the broader robotics landscape, the global humanoid market was worth about $4.16 billion in 2023 (Global Market Insights). Projections converge on extraordinary growth: the same source estimates a value of $74 billion by 2032 with a CAGR of 37 percent; BCC Research projects $11 billion by 2030 with a CAGR of 42.8 percent.
Oversonic Robotics’ RoBee at SDA Bocconi
Four drivers of adoption and three application areas
On the production front, signals are even more concrete: China shipped about 13,000 humanoid robots in 2025, 90 percent of them domestically produced. Chinese shipments are expected to reach 62,500 units in 2026, a 270 percent increase in a single year. Goldman Sachs forecasts a drop in production costs from $150,000 to about $30,000 per unit, while Bank of America estimates annual production could reach 1 million robots as early as 2030. Geographically, the Asia-Pacific region already holds more than 53 percent of the global market and is expected to maintain its dominant position throughout the forecast period.
But why are robots leaving the lab? Four major forces are driving demand:
- Global demographic aging.
- Growing industrial automation.
- Advances in AI and machine learning.
- Consumer demand for personal assistants.
The most advanced experiments are concentrated mainly along three application areas, where humanoid robotics can deliver the most direct advantage over more traditional automated systems: manufacturing and logistics; healthcare and personal assistance; operations in high-risk environments for humans.
Manufacturing and logistics. Automotive plants represent the most active context for experimentation. BMW has launched a pilot with Figure AI robots in its production chain in Germany, and Hyundai is testing Boston Dynamics’ New Atlas in its facilities. Amazon and Spanx use Agility Robotics’ Digit robots in warehouse operations, while GXO Logistics has started collaborations with Agility for parcel handling. These deployments are still largely in the feedback phase toward robotics companies, but they signal a clear direction: the humanoid robot can operate in environments designed for humans without requiring costly infrastructure modifications.
The automotive sector appears to be the most promising candidate for short- to medium-term adoption (before 2030): major manufacturers have supply chains that significantly overlap with those of humanoid robotics, and production volumes allow for highly favorable cost negotiations. Tesla aims to produce 5,000 units of its Optimus in 2025, with plans to scale to 12,000. BYD aims to deploy 1,500 humanoid robots in 2025, with a target of 20,000 in 2026.
Healthcare and assistance. The healthcare sector is identified by several market analyses as the main application area for humanoid robots by 2025, with an estimated 28 percent share of the global market. Applications range from elderly care to supporting nursing staff, from hospital logistics to rehabilitation support. The critical issue in this context is not only technological but regulatory: certification as a medical device is a high barrier, but also a fundamental guarantee of trust for patients and healthcare professionals.
Dangerous environments and high-risk operations. A third application stream concerns repetitive, dangerous, or physically exhausting tasks: industrial painting, handling in extreme temperature environments, plant inspection, logistics in confined spaces. In these contexts, the economic advantage of humanoid robots over traditional solutions may be more immediate, and worker acceptance thresholds are higher because the robot operates “on tasks that human beings deserve not to have to do,” as Oversonic Robotics’ manifesto states.
To be or not to be humanoid?
In a landscape dominated by US and Chinese players, Italy is home to an internationally relevant company. Oversonic Robotics, founded in 2020 in Besana Brianza, is now recognized by CB Insights among the 13 leading global players in humanoid robotics for industry. At CES 2026 in Las Vegas, the company presented its humanoid robot RoBee in the pavilions of Intel and STMicroelectronics, alongside global technology giants.
RoBee is the first Italian cognitive humanoid robot certified for industrial use, and the only one in Europe already operational both in manufacturing contexts and in clinical healthcare trials.
Oversonic’s distinguishing element lies not only in its technology but in its approach: instead of competing in the arena of “spectacular” robots that are not yet reaching the market, the company has focused on systems that can be certified, are reliable, and can be integrated today. “We did not start from what the robot might be able to do in the future, but from what can be certified and put to work now,” explained Leonardo Ancesci, Chief Delivery and Sales Officer at Oversonic, speaking at a recent DEVO Lab meeting.
The meeting offered an opportunity for an honest assessment of the sector: it must be acknowledged that the gap between performance demonstrated in a lab or on a television show and operational reliability in real industrial environments remains significant. The main obstacles are:
- Consistency and repeatability. Current humanoid robots struggle to guarantee the same performance over long production cycles and under unpredictable environmental variations. Existing deployments are still largely closed experiments whose results have not yet been made public.
- Costs. Although declining, prices remain high for mass adoption. Unitree’s G1 model starts at 85,000 yuan (about $12,000) in China, but robots for industrial use in Western markets typically range between €100,000 and €200,000.
- Regulation and safety. There are still no clear standards for certification and safe operation of humanoid robots in mixed work environments. Shared regulatory frameworks are not expected before 2027.
- Limited availability. Most models announced by leading players are not yet commercially available at scale. Media hype often exceeds the product’s actual state of maturity.
Alternative forms of robotics (wheeled robots, fixed industrial arms, non-humanoid collaborative systems) remain better suited for many specific applications and offer a more favorable cost-benefit ratio in the short term. The humanoid form is not inherently superior: it is optimal where there is a need to operate in environments designed for the human body without modifying the infrastructure.
