What are inertial sensors and why a humanoid can't take a single step without them
Inertial sensors — more precisely inertial measurement units (IMUs) — are miniature devices that combine accelerometers, gyroscopes, and sometimes magnetometers into a single chip. Their task sounds simple: measure acceleration, rotation, and orientation in space in real time. For a humanoid robot moving on two legs in an unpredictable environment, this data is absolutely critical — without it, it couldn't maintain balance, let alone dynamically respond to uneven terrain or unexpected obstacles.
Unlike ordinary IMUs found in smartphones or drones, sensors for humanoid robots must meet much stricter requirements: extremely low latency (response in microseconds), high precision even under vibrations and temperature fluctuations, and above all certification according to functional safety standards such as ISO 13849 or IEC 61508. In other words — when a gyroscope in a phone fails, it's an inconvenience. When it fails in a robot working next to a human, it can mean serious injury.
The market will grow by 11% annually. Who's driving it?
According to a June report by analytics firm IndexBox, the market for inertial sensors for anthropomorphic robots will reach an index value of 285 points by 2035 (from a baseline of 100 in 2025), corresponding to a compound annual growth rate (CAGR) of approximately 11%. That is significantly higher than the broader IMU market, which will grow at roughly 7.2% annually.
Growth is driven by five main segments:
- Humanoid robots (35% of the market) — Tesla Optimus, Boston Dynamics Atlas, Figure AI, or Agility Robotics Digit. With the transition from labs to real-world deployment, the number of sensors per robot is increasing (up to several IMUs per limb and torso).
- Collaborative robots — cobots (25%) — Universal Robots, Fanuc, ABB. Safety regulations increasingly require certified inertial sensors for collision detection and speed monitoring.
- Medical prostheses and orthotics (15%) — Companies like Ottobock and Össur integrate IMUs into bionic limbs for more natural gait.
- Exoskeletons (15%) — Ekso Bionics, Sarcos. Inertial sensors detect user intent and provide assistive force.
- Research and development (10%) — Universities and research labs testing new sensor fusion algorithms.
Asia dominates, Europe holds the quality edge
Regionally, Asia-Pacific leads with a 45% share — primarily thanks to China's aggressive humanoid robotics development program. In the first nine months of 2025 alone, Chinese robotics startups received over 50 billion yuan (roughly 7 billion dollars) in funding. As a result, Chinese companies like AgiBot or Unitree are already delivering humanoids in thousand-unit batches.
North America holds 25% of the market — with Tesla, Boston Dynamics, and Figure AI forming the innovation core — and Europe accounts for 18%. Europe is notably home to key manufacturers of the sensors themselves. Germany's Bosch Sensortec is a global leader in MEMS inertial sensors, Switzerland's STMicroelectronics ranks among semiconductor industry giants, and France's SBG Systems supplies high-precision inertial navigation systems for mobile robotics.
For the European — and by extension Czech — industry, this means one crucial thing: value isn't created only in the final robots, but also in their components. Czech companies in precision engineering, automation, and sensor electronics manufacturing (for example in the cluster around CTU or BUT) can benefit from demand for calibration and testing services, which are essential for certifying safety IMUs.
Why the market isn't taking off even faster
Despite the optimistic outlook, the market faces several obstacles. The main one is lengthy qualification cycles — a safety sensor must undergo months to years of testing before a robot manufacturer approves it for mass deployment. This slows the entry of new suppliers and keeps barriers to entry high.
The second constraint is competition from visual odometry — cameras and LiDARs can, in many cases, estimate a robot's position without the need for expensive IMUs. However, cameras fail in darkness, fog, or during rapid movements, so inertial sensors remain an irreplaceable backup. The third obstacle is the concentration of MEMS chip manufacturing — most of the world's capacity is in the hands of a few foundries, creating the risk of supply chain bottlenecks.
Who the key players are
In its analysis, IndexBox identified twenty key market participants. At the top, alongside the aforementioned companies, are also TDK/InvenSense (Japan), Analog Devices and Honeywell (USA), Norway's Sensonor (part of TDK), and Japan's Murata and Epson Toyocom. An interesting role is also played by Xsens (Netherlands, now part of Movella), whose motion capture systems are used in robotics research worldwide, including at European university labs.
A separate chapter is the software layer — companies like CEVA (SenslinQ) supply sensor data fusion algorithms that allow even cheaper IMUs to achieve performance comparable to more expensive units. AI-driven sensor fusion is, according to IndexBox, one of the strongest trends that will reduce costs and open the market to new applications in the coming years.
What this means for Czech companies and research
The Czech Republic may not be among the manufacturers of humanoid robots, but it has a strong base in industrial automation and robotics. Companies like ABB (with a development center in Brno), Škoda Auto (which uses robots extensively in production), or research institutions such as CIIRC CTU and CEITEC BUT actively monitor trends in robotic sensors. As European regulation (including the EU AI Act and upcoming standards for physical AI) raises safety requirements, demand for certified components will grow — and this is where even smaller European suppliers can find opportunities.
Moreover, with the approaching European AI chip factory (including the planned node in Ostrava as part of the Czech AI Factory), Central Europe's significance is growing as a region that can host testing and calibration capacities for robotic sensors. From IndexBox's perspective, the market is gradually modularizing and standardizing — communication interfaces like SPI, I2C, or CAN are becoming universal, lowering barriers to sensor integration from different manufacturers.
Numbers worth paying attention to
To complete the picture: according to Precedence Research, the humanoid robot market will reach 8.78 billion dollars by 2035 (compared to 2.16 billion in 2026), at a CAGR of 16.9%. Each such robot requires several inertial sensors. For the most advanced platforms, such as Tesla Optimus or Boston Dynamics Atlas, we're talking about several IMUs per limb, torso, and head — partly for redundancy, partly for specialization (different sensor for the foot, different one for the arm).
Meanwhile, global shipments of humanoid robots have grown from 53 thousand units in 2025 to a projected 810 thousand in 2030, according to SAG estimates. Even with a conservative estimate of three specialized IMUs per robot, that means demand for millions of sensor units annually.
What's the difference between a regular phone IMU and a sensor for a humanoid robot?
The fundamental difference lies in precision, latency, and safety certification. A phone IMU needs only approximate motion detection with latency in milliseconds. A robotic sensor must respond in microseconds, withstand vibrations and temperature fluctuations, and meet functional safety standards (ISO 13849, IEC 61508). Failure of a phone gyroscope is a cosmetic flaw; failure of a robotic one can cause the machine to fall and injure a person.
Are European sensor manufacturers competitive against Asia?
Yes, and in some segments they even lead. Bosch Sensortec (Germany) is the global leader in MEMS IMUs, STMicroelectronics (Switzerland) is the world's second-largest semiconductor sensor supplier. European companies benefit from a long tradition of precision engineering and a strong emphasis on safety certification — exactly the qualities that robotics customers demand.
When will we see humanoid robots in regular operation in Czechia?
The first deployments in Czechia can be expected in industrial manufacturing and logistics, likely within a 2–4 year horizon. Škoda Auto is already testing robotic automation across many processes, and Czech integrators like ABB Robotics have ready-made platforms for collaborative robots. Consumer humanoids for households, however, face a longer journey — estimates point to broader adoption only after 2035.