Humanoid robots don’t deserve their superhuman valuations

At least robots don’t need to be told to calm down. Investors are a more excitable bunch, as shown by the 460 per cent gain for humanoid robot maker Unitree on its market debut. The Chinese group’s $51bn market capitalisation makes even Tesla, another big promoter of robot potential, look cheap. It would take a ninefold increase in Unitree’s robot output at current prices just to bring its enterprise value, relative to sales, in line with Tesla’s.

Unitree raised about $900mn in Shanghai this week. Demand among retail investors was thousands of times the shares available, driven by a mix of mainstream familiarity with the Hangzhou company’s trick-performing humanoid and quadruped robots, and the soaring successes of other recent technology floats. Unitree’s net profits from last year selling 28,000 robots — mostly far cheaper robot dogs — will also have helped, along with a roster of illustrious backers from Tencent to DeepSeek.

Some of Unitree’s rocketing price can also be attributed to Beijing’s tight control of its IPO markets. Companies have raised $33bn going public in China so far this year according to LSEG: more than in all of 2025, but far short of recent boom years, narrowing investor opportunities.

Yet even at its $9bn IPO valuation and factoring in negligible net cash, on S&P Capital IQ numbers, Unitree priced at 27 times this year’s sales, assuming first-half figures are doubled. Its debut gains imply it is trading on a multiple of over 150. Tesla, known for giddy valuations and a retail investor fan base, is cutting car production to make space to build its Optimus general-purpose robots. It trades on 13 times 2026 forecast sales, according to S&P Capital IQ.

Elon Musk in April said he expected Optimus to be “not just Tesla’s biggest product ever, but probably the biggest product ever.” He and Chinese investors are not alone in their enthusiasm: total addressable market estimates for humanoid robots are huge. Morgan Stanley analysts last year estimated there would be nearly 1bn in use for simple tasks by 2050. A team at UBS reckons there will be 300mn by that point creating an addressable market worth up to $1.7tn.

The appeal of humanoids is their potential to do more than one task, unlike most existing robots. In theory one humanoid can, say, assemble multiple different components, manage transport around a warehouse or stock shelves with various items. At some point, the machines could be dexterous enough to support nursing care or manage household tasks.

Alas humanoids are not about to stack the dishwasher. They are currently far too pricey for industrial viability, let alone consumers, and they can usually only manage a couple of hours between charges. Unitree is clear that the 5,215 humanoid robots it sold last year were often used as platforms for research and development.

The company is not short of ambition, this week unveiling a new biped it has dubbed “Superman.” But just matching investor enthusiasm with reality is truly a superhuman task.

[email protected]