For Humanoid Robot to spread in earnest, the price has to come down to smartphone levels. If the price drops to $20,000–$30,000 (about 30 million–45 million won) within the next five years or so, we expect adoption to speed up, led by industrial and manufacturing sites. ST is not stopping at supplying individual semiconductors; we are focusing on bundling sensors, microcontrollers (MCUs), motor control, and power semiconductors at the system level to shorten customers' development time and reduce cost.
Allan Lagasca, strategy program applications director for the Robotics segment in the smart industry division at STMicroelectronics (ST), said this in an interview with ChosunBiz on the 15th during a visit to Korea to attend the annual tech conference (K-TEC 2026).
Physical artificial intelligence (AI) has emerged as a major topic in the global tech industry this year because multiple corporations have unveiled Humanoid Robot models that are close to commercialization. However, there are many hurdles to clear before humanoids become mainstream. Issues include the maturity of the AI software that drives robots, lack of training data, battery life, precise hand and joint control, safety regulations, cybersecurity, and supply chain problems.
ST is an integrated device manufacturer (IDM) that supplies semiconductors to a wide range of electronic applications, including automobiles, industrial equipment, data centers, and home appliances. Last year's revenue was $11.8 billion (about 17.7 trillion won), and it collaborates with more than 200,000 customers worldwide. In the robot market, rather than targeting the AI chip that serves as the central brain of a humanoid, the company has set a strategy to bundle and provide key peripheral semiconductors needed for a robot to see, move, and control itself in the real world at the system level.
◇ Humanoid price expected to be in the $20,000–$30,000 range in five years
Lagasca cited price as the biggest factor hindering the spread of humanoids. He said, Humanoids are still in the early stages, but the number of deployments is growing rapidly, and added, The next two to three years will be a period when the market is validated, and after that, higher levels of adoption may begin. He continued, In about five years, the price of a humanoid could come down to the $20,000–$30,000 range, and said, It is still expensive compared with smartphones, but if the price drops significantly from now, it could be a tipping point for market expansion.
To the question of when Humanoid Robot can reach smartphone-level pricing, he answered, It is hard to predict precisely, but looking at the trend, it is not that far in the future. He said, In 2024, about 2,400 humanoids were deployed worldwide, and last year that increased to about 18,000–20,000, and added, This year, we see the number already surpassing 50,000. When humanoids reach the level of surpassing 1 million units, that could be the point when mass production scale becomes possible. He said, If rising production volumes and falling prices coincide, market diffusion could accelerate further.
◇ System integration, not individual chips… "The industrial and manufacturing market will open first"
Lagasca identified industrial and manufacturing sites as the first markets where Humanoid Robot will gain a foothold. He said factories and logistics centers have relatively standardized workflows and work environments, making it easier to apply robots there than in homes or general service spaces.
He said, Manufacturing and logistics environments are relatively structured, but for humanoids to enter environments like homes, where floor plans differ from house to house, they need situational perception and understanding capabilities and higher-level AI, and added, There are still clear limits in motion speed, precision, and operating time.
Lagasca emphasized that ST's semiconductor technology can help humanoid customers with technology development and cost reduction. ST approaches Humanoid Robot by dividing it into the head, body processing unit, communications, power management, joints, hands, and sensing. It supplies MCUs, sensors, motor drivers, power management semiconductors, and wired and wireless communication chips that can be used in a customized way for each area.
He said, Semiconductor integration can play an important role in lowering robot expense, and added, If you integrate MCUs, analog, and power semiconductors into a single chip or module, you can reduce development time and manufacturing cost. He continued, We have a portfolio of more than 500 parts for Humanoid Robot, and added, We see a bill of materials (BOM) opportunity of about $600 per robot.
Although ST supplies a wide range of semiconductors needed for humanoids, it does not directly develop the central AI brain. Lagasca said, ST is not a graphics processing unit (GPU) supplier like Nvidia, nor are we trying to be the main brain of a humanoid, and added, However, some AI functions are essential for processing data in each module, such as sensors and joints. He continued, For inference at the level of a camera recognizing an image and identifying whether it is a phone or a water bottle, a small AI accelerator is sufficient instead of a large GPU.
ST is collaborating with Nvidia to expand the physical AI ecosystem. It is connecting ST's sensing and control technologies with Nvidia's robot development platforms to help robot manufacturers develop products faster. Most recently, together with Leopard Imaging, it unveiled a multimodal vision module that integrates with Nvidia's Jetson and Isaac platforms.
Regarding the Korean market, he said it could be an alternative to the robot supply chain outside China. Lagasca said, In Korea, activity in the Robotics field is very vibrant, including humanoid companies and robot parts companies, and added, China has many robot parts suppliers, but for Western robot manufacturers seeking a supply chain outside China, Korea can be a good alternative.
Lagasca joined ST's Taiwan office in 2003 as a senior applications engineer. He later worked in the field application engineering (FAE) organization and the systems engineering and strategy program areas for China, Southeast Asia, and India, and he now leads the expansion of ST solutions and markets in industrial automation and Robotics.