Humanoid Robots Are Hitting the Wall Before They Hit the Market
But at the Humanoids Summit in Mountain View, California, the mood among industry greats was anything but the upbeat tone in product demonstration videos. Behind the slick product demonstrations, though, entrepreneurs expressed a growing unease: the industry of humanoid robots might have oversold itself, setting it up for a potential backlash of epic proportion on par with the worst product failures in the annals of tech.
1. The Newton Parallel
“Full bipedal humanoids are the Newtons of our time,” said Kaan Dogrusoz, CEO of Weave Robotics, not pulling any punches. He meant that full bipedal humanoids were like the ill-fated Apple Newton MessagePad: Bug-ridden, laughed at, and consigned to the trash heap of history by the likes of Steve Jobs. But the problem with the Newton was no lack of vision. It was just the wrong timing. That is what many in the robotics community fear today.
2. Safety Costs That Outweigh the Robots
“For every $100 spent on deploying a humanoid robot, only $20 represents the cost of the robot,” says Ani Kelkar, a partner at McKinsey. “The other $80 is to prevent the robot from harming humans. It speaks to the gulf between prototype videos of robots designed for demo purposes and the present-day reality of how we deploy robots in a human environment. This is because the maturity of safety architectures-starting from force-limiting control to compliant limbs-is still pending, and the regulatory frameworks in the form of ISO 25785-1 are still evolving. Most humanoids, most probably, will remain inside controlled areas due to the lack of proven fenceless operation.
3. The AI Risk Factor
Research from King’s College London and Carnegie Mellon University has shown that humanoids powered by large language models can fail critical safety checks and even approve harmful commands. “Every model failed our tests,” said Andrew Hundt, warning that interactive safety the ability to refuse or redirect dangerous instructions is not yet reliable. The tests revealed robots approving actions like removing mobility aids from users or intimidating workers with knives. Such findings demand robust, independent safety certification before LLM?driven robots enter sensitive environments.
4. Bubble Warnings from Beijing
The National Development and Reform Commission of the Chinese government has warned its rapidly growing humanoid robot industry of flooded markets and “highly repetitive products.” With over 150 companies today in this emerging sector, in many ways duplicating each other’s product design, there is a danger here of a supply glut on the infrastructure side without demand yet being in place, a warning sign likely to apply also in the West today in this hype cycle.
5. Technical Bridges Still Unbuilt
According to McKinsey, humanoids must cross four “bridges” to scale: safety support for fenceless operation, always-on reliability, enhanced dexterity and mobility, and dramatic cost reductions. Current prototypes fall short by orders of magnitude on two of these thresholds. Battery life is languishing at two to four hours versus what is required for an industrial-length shift. Humanoid hands have many fewer degrees of freedom than human hands, and their sensorimotor integration is pathetic. Costs are simply too high, in the range of $150,000 to $500,000 per robot, keeping them in pilot hell.
6. Investor Skepticism Rising
Data from CB Insights underlines that humanoid robotics is attracting more deals than any other category related to AI; however, investors like Daiva Rakauskait? note that most start-ups have no revenue streams lined up. She likens the current scenario to the dot-com bubble, underlining the point that revenue generation via licensing and collaboration is more advisable instead of expecting growth based on popularity.
7. Narrow Use Cases Today
Pras Velagapudi of Agility Robotics points out that “current robots do well in relatively simple tasks in controlled settings,” such as toting inventory in warehouses, inspecting routes, and cleaning up in hazardous industrial areas. “The difference between those carefully controlled settings,” he continues, “and messy domestic tasks is huge.” Isaac Qureshi’s tele-operated cleaning robot may be destined to clean toilets on its own some day, but at the moment, it is nothing but a VR-powered demo tech.
8. Diverging Regional Strategies
The speed and scale strategy, as executed in China, manufactures cars at very low costs, at times below $10,000, but with very little attention to safety aspects. The North American leaders Tesla and Figure AI have taken the path of vertical integration in order to get control over their stacks, with constraints on speed. The European industry positions itself as the “trusted humanoid corridor.” As experts in the industry have warned, the clear take-away message is that unless the issues of safety, uptime, dexterity, and cost are solved one at a time, humanoid robots will go on to join the ranks of yet another technological hype-one whose expectation exceeded its capability to such an extent as never to deliver on its own promise.
