What is the typical lifespan of a humanoid robot?

Why there's no good benchmark yet

The modern commercial humanoid robot market is too young to have an established lifespan benchmark. Most major deployments (Apptronik Apollo at Mercedes-Benz, Agility Robotics Digit at GXO, Figure AI 02 at BMW) are under three years old. There is no public field data answering "how long does an Apollo last in continuous warehouse use" because not enough Apollos have been in continuous warehouse use long enough.

By contrast, industrial cobots (collaborative robotic arms) have a 10–15 year track record of established lifespans in factories. Humanoids will eventually accumulate equivalent data; in 2026 we don't have it yet.

What we know at the component level

Even without a system-level number, the wear items are known:

Harmonic drives

The harmonic drives used in nearly every commercial humanoid's joint actuators are rated for roughly 20,000 to 30,000 hours of industrial-duty operation before requiring rebuild. At a single-shift duty cycle (8 hours/day, 250 days/year = 2,000 hours/year), that translates to 10–15 years per drive. At multi-shift continuous duty (16+ hours/day) the timeline compresses proportionally.

Battery packs

Lithium-ion packs typically tolerate 1,000–3,000 full charge cycles before significant capacity loss. At one charge cycle per shift, that's 3–10 years of practical battery life before replacement is needed.

Hot-swap battery systems (used by Apptronik Apollo and Agility Digit for continuous-duty operation) accelerate cycle accumulation per pack but make replacement straightforward.

Actuators and motors

Brushless DC motors used in humanoid joints last years under normal duty. Bearings and gear teeth are wear items; expected service intervals depend on load and use pattern.

Sensors

Cameras, lidar, IMUs, and tactile sensors are generally long-lived, but degraded over time by environmental exposure (heat, vibration, dust).

What kills a humanoid first, in practice

Field experience from related domains (industrial robotics, mobile robots) suggests the failure modes that matter:

1. Falls. A humanoid that falls hard can damage actuators, structural elements, and sensors. And a 35-kg humanoid falling is a serious mechanical event.
2. Cumulative actuator wear in high-torque joints. Typically hips, knees, shoulders.
3. Battery degradation. Pack capacity loss is gradual but eventually forces replacement.
4. Software obsolescence. A working hardware platform whose maker has dropped firmware support is functionally dead even if it physically operates.

Realistic operational life

Combining the component data, a reasonable working assumption for a commercial humanoid in regular duty in 2026 is:

• Hardware can plausibly serve 5–10 years with regular service and parts replacement.
• Battery replacement is needed every 3–5 years.
• Software support is the binding constraint: if the maker stops shipping updates, the practical life ends regardless of hardware condition.

This is an inference, not a benchmark. Treat any specific lifespan number you see in 2026 marketing material with skepticism: nobody has the data to back it.

Bottom line

There is no industry-published humanoid robot lifespan in 2026. Component data suggests 5-10 years of useful service under regular duty with proper maintenance. But the binding constraints are real-world fall damage, battery replacement, and the maker's software support window. See the risks of humanoid robots for related safety considerations. For methodology canonical references applicable to humanoid lifespan claims: verified-vs-claimed at within-entity granularity (component-level vs system-level claim depth distinction) + the 9-tier source-quality rubric.