New Research: Three Robotics Breakthroughs Redefine Energy Limits
New findings from Figure AI, ESA, and German researchers show robots pushing past energy and endurance limits that previously defined the field.
5 min read
0:00
0:00
New findings from Figure AI, ESA, and German researchers show robots pushing past energy and endurance limits that previously defined the field.
Figure AI's humanoid robots completed 24 consecutive hours of autonomous operation, a first for the company and a meaningful marker for industrial deployment readiness.
According to Interesting Engineering, Figure AI announced that its humanoid robots have crossed the 24-hour continuous autonomous work threshold. The company described this as entering 'uncharted territory.' From a builder's perspective, that framing matters: continuous runtime is not just a performance metric, it is a reliability signal. Any machine running nonstop for a full day is accumulating thermal stress, joint wear, and software edge cases simultaneously. What the data suggests is that Figure AI is stress-testing its systems under conditions that resemble actual industrial deployment, not controlled lab runs.
Industrial operators do not care how fast a robot moves in a demo. They care whether it can hold a shift. A robot that performs brilliantly for four hours and then needs extensive downtime is operationally expensive. The 24-hour benchmark, if reproducible, starts to close the gap between humanoid robots and fixed automation systems that run continuously. That is the comparison operators are making.
The announcement, as reported by Interesting Engineering, does not detail how many charge cycles were involved, what tasks were performed during the run, or what failure modes appeared and were corrected. Those details matter enormously for anyone trying to assess whether this is a controlled showcase or a replicable operational baseline. Honest assessment: milestone, yes. Proof of readiness, not yet confirmed by independent data.
University of Gothenburg researchers built a soft robot using artificial muscles rated to 10 MeV radiation, capable of navigating unstructured terrain without rigid actuators.
As reported by Interesting Engineering, a team led by researchers at the University of Gothenburg developed an inchworm-inspired soft robot designed for planetary exploration. The key engineering detail is the radiation tolerance: the artificial muscles can withstand 10 MeV radiation levels, which are relevant for Mars surface conditions. The robot uses soft actuators rather than conventional rigid motor-and-gearbox systems, which changes the energy profile significantly. Soft actuators can store and release energy passively, which reduces the power draw per movement cycle.
Rigid actuator systems, like the harmonic drives and brushless motors used in most humanoid robots, require continuous power input to hold position and generate force. Soft actuators can exploit material compliance to absorb and redirect energy more naturally. For planetary exploration, where power budgets are extremely tight, this is not a nice-to-have. It is a design requirement. The Gothenburg team's approach connects directly to ongoing debates in humanoid robotics about whether quasi-direct drive or series elastic systems can deliver similar passive energy benefits.
According to the Interesting Engineering report, the inchworm robot navigates Mars-like terrain using its soft muscle architecture across multiple degrees of freedom. Without rigid joints defining movement limits, the robot can conform to irregular surfaces in ways that traditional legged or wheeled systems cannot. The limitation is obvious: soft robots sacrifice speed and payload capacity for adaptability. For exploration tasks in unstructured environments, that trade-off may be acceptable. For warehouse logistics, it is not.
German researchers are developing a robot-assisted system to recover and repurpose EV battery cells, addressing a supply chain bottleneck that directly affects robot energy storage capacity.
As reported by Interesting Engineering, German researchers are building a robotic system to recover usable cells from end-of-life electric vehicle batteries and prepare them for reuse. The connection to robotics energy efficiency is direct: humanoid robots depend on high-density battery cells, and the supply chain for those cells is under pressure. A system that can reliably sort, test, and repurpose EV battery cells at scale could reduce the cost and improve the availability of battery packs used in mobile robotics platforms.
EV battery packs are not designed for disassembly. Cell formats, adhesive types, and structural configurations vary significantly across manufacturers and model years. A robotic system that can handle this variability needs sophisticated sensing, adaptive gripping, and decision-making at the cell level. The German research effort is tackling exactly that problem, though the Interesting Engineering report does not specify current throughput rates or the range of battery formats the system can handle. Those gaps matter for assessing commercial viability.
All three efforts share a core constraint: energy. Runtime, radiation-tolerant actuation, and battery recovery are all responses to the same underlying limit on how long and how efficiently robots can operate.
From a builder's perspective, these three stories are not coincidentally published in the same week. They reflect a field-wide reckoning with energy as the binding constraint in robotics. Figure AI is pushing runtime. The University of Gothenburg team is building actuators that survive on minimal power in hostile conditions. German engineers are working on recovering the energy storage components that make mobile robots possible. The specs tell a different story than the individual headlines: the robotics industry is not just a compute problem or a software problem. It is fundamentally an energy management problem.
All three studies are early-stage or announcement-level results. Independent replication, throughput data, and real-world deployment metrics are largely absent from current reporting.
Honest assessment of where each project stands: Figure AI's 24-hour milestone lacks publicly available methodology details on task complexity, charge management, and failure events during the run. The University of Gothenburg's inchworm robot is a research prototype; the gap between a lab-capable soft robot and a deployable planetary rover involves years of engineering and testing. The German EV battery recycling system is described as under development, with no confirmed throughput or commercial timeline in the Interesting Engineering report. These are real advances worth tracking. They are not finished products.
Key follow-on signals include Figure AI's deployment partner announcements, ESA's next-phase funding for the soft robot project, and the German recycling system's pilot production results.
For Figure AI, the 24-hour runtime milestone becomes meaningful when paired with a commercial deployment announcement. Which industry partner is running these robots, under what task conditions, and at what uptime guarantee? For the inchworm robot from the University of Gothenburg, the next signal is whether ESA or another space agency moves the project into a funded mission development phase. For the German EV battery recycling system, watch for pilot scale data: how many cells processed per hour, at what recovery rate, and whether the system can handle multiple battery formats. Each of those follow-on data points will separate credible progress from well-publicized prototypes.
According to Interesting Engineering, Figure AI has now demonstrated 24 hours of continuous autonomous operation, which the company describes as an industry first. Most humanoid robots in commercial deployment operate in much shorter cycles, with runtime depending heavily on task intensity and battery capacity.
Radiation-tolerant soft actuators are flexible, muscle-like components that can operate in high-radiation environments where conventional motors would fail. As reported by Interesting Engineering, University of Gothenburg researchers built such actuators for a Mars-exploration robot, rated to withstand 10 MeV radiation levels.
Humanoid robots rely on high-density lithium-ion battery cells. As the EV sector generates large volumes of retired battery packs, robotic systems that can recover and repurpose usable cells could ease supply pressure and reduce battery costs for mobile robotics platforms, including humanoids.
The core challenge is energy density and runtime. Current battery technology limits continuous operation, and sustained actuation generates heat that degrades both motors and battery cells. All three research projects covered this week address different facets of this shared constraint.
Based on current reporting, all three projects are at prototype or early-development stage. Figure AI's runtime milestone is the closest to operational relevance, but independent methodology data is not yet publicly available. The University of Gothenburg soft robot and the German battery recycling system are research prototypes with no confirmed commercial timelines.