In one of the more unusual endurance experiments of 2026, a humanoid robot named Qestr lined up alongside human athletes at the Antarctic Ice Ultra and promptly demonstrated a hard limit of current robotics technology.
On November 8, near Ultima Camp in the Schirmacher Oasis of Queen Maud Land, the machine covered only about three kilometers before its battery failed in temperatures around −24°C (−11°F). The humans, by contrast, kept going—some for more than 160 kilometers.
The episode widely reported as a failure, served as a rare, unfiltered stress test of humanoid hardware in one of Earth’s most extreme environments.
The Race at the Bottom of the World
The Antarctic Ice Ultra is a multi-distance event staged on a roughly 4.2-kilometer groomed snow loop around Ultima Camp. Competitors choose distances ranging from 50 kilometers up to 161 km, which requires dozens of laps.
Participants typically assemble in Cape Town, South Africa, and fly in on a heavy transport aircraft that lands on a blue-ice runway. The 2026 edition began in the early hours of November 8 under the midnight sun, with the start time advanced to avoid incoming weather.
About 25 human runners took part. Chinese ultrarunner Yusheng Ni won the 100-mile race in 17 hours, 2 minutes and 24 seconds, shattering his own previous Antarctic mark by more than five hours. Canadian Stephanie Case finished second overall in 17:31:39 and became the first woman to complete 100 miles on the continent.
Into this field stepped Qestr the first humanoid robot ever entered in the event, according to race organizer Oliver Wang. The machine wore a race bib number 29 and a navy thermal puffer jacket intended to shield its electronics. Footage shows it moving with a functional running gait across the snow and ice in the early stages.
Why the Robot Stopped
The limiting factor was not balance, traction, or mechanical failure of the legs. It was power. Lithium-ion batteries, the standard energy source for most current humanoids, suffer sharply reduced performance in deep cold. Electrochemical reactions slow, internal resistance rises, and usable capacity drops. Even with the protective jacket, Qestr’s battery depleted after roughly three kilometers.
Humans generate metabolic heat and can refuel with food and drink at aid stations. A robot carries only the charge it starts with. Once that is gone, the motors, sensors, and computers shut down.
Reports indicate the robot did not fall; it simply ran out of electricity and was withdrawn. Some accounts note that after the attempt it performed a brief dance between the flags which is an incongruous but telling image of a machine designed more for controlled indoor or mild outdoor settings than polar extremes.
Details about Qestr’s manufacturer remain limited in public reporting. It has been described as silver-white or pale with a dark visor face, jointed lower body, and relatively simple hands. One French trail-running site reported it as Chinese-built, approximately 80 cm tall with 32 joints, though these figures have not been widely corroborated.
Broader Lessons for Humanoid Robotics
The short run is less a punchline than a data point. Humanoid robots have advanced rapidly in locomotion, balance, and task performance in factories, warehouses, and demonstration halls. Deploying them outdoors in unstructured, temperature-extreme environments is a different challenge. Cold is especially unforgiving for energy systems. Similar battery degradation issues appear in electric vehicles and consumer electronics in winter conditions; at Antarctic temperatures the effect is simply more severe.
A Useful Failure
Calling the attempt a pure failure misses the point. No previous humanoid had raced in conditions this cold or traveled this far south to compete. The mechanical systems handled snow and ice well enough for a short distance. The humans who finished 100 miles demonstrated, once again, the remarkable endurance of biological systems that can convert food into sustained effort while generating their own heat.
Qestr’s early exit simply confirmed what battery chemists and roboticists already suspected: power remains the bottleneck for extreme outdoor autonomy. Future attempts will almost certainly arrive with better thermal protection, cold-optimized cells, or different energy architectures. When they do, the Antarctic Ice Ultra or similar polar tests will remain one of the clearest ways to measure progress.
