Three scenes, the same week, and only one possible reading. In Tokyo, the Humanoids Summit declares that walking is a solved problem and that the real question is now scaling up. In Shanghai, the stock exchange approves Unitree Robotics' listing, the first embodied AI player admitted to the A-share market. And in between, almost inaudible beneath the choreography of demonstrations, the US NIST acknowledges that after billions invested, no one yet knows how to measure what these machines can truly do. The industry has listed the scale before building the metre.
Tokyo shifts the question: from 'can it walk?' to 'can it scale?'
The sector’s doctrine has changed its footing. The central question of the Humanoids Summit Tokyo 2026 is 'what must be true for all this to truly scale' — a clear shift from previous editions, where in 2024 we asked 'can it walk?' and in 2025 'can it work?'[1]. Locomotion — walking, climbing stairs, balance recovery — is now treated as a solved problem and a basic expectation for major platforms; the frontier has moved to dexterity and fine motor control[2].
The shift is elegant on paper. It nevertheless assumes a given that remains to be proven: declaring locomotion 'solved' already assumes we have an instrument to observe it. Yet nothing in the Summit’s folklore provides this instrument. The most telling proof lies in a showcase: GMO AI & Robotics displays its G1 robot, designed for airport freight handling and equipped with a camera vision system, whose internal mechanics are actually supplied by Chinese firm Unitree[3]. The 'working' robot at the airport is thus a Japanese-branded assembly with Chinese innards — an image of scalability, not a measure of work.
Shanghai capitalises the narrative: a first listing for embodied AI
On 1 June 2026, the Shanghai Stock Exchange’s listing review committee approved Unitree Robotics' listing, making the company the first embodied AI firm authorised to enter the Chinese A-share market[4]. The operation, conducted on the STAR Market, seals this status as the sector’s first listed company[5]. It aims to raise 4.202 billion yuan — approximately 616 million dollars — by issuing at least 40.45 million new shares[6].
On the accounting front, the trajectory is spectacular. Unitree’s revenue rose from 159 million yuan in 2023 to 393 million in 2024, then to 1.699 billion in 2025 — a growth of 335% in the last year alone[7]. Volumes follow suit: humanoid robot deliveries exceeded 5,500 units in 2025, placing the manufacturer at the top of the global rankings by volume delivered[8]. This is what the public market is invited to capitalise on: a narrative of scaling, backed by rising curves.
What the prospectus admits: units sold to labs, a collapsing price
Yet the same filing says something else, in a quieter voice. More than 70% of Unitree humanoids sold in 2025 are destined for research and education — not for a workstation[9]. In other words, what is 'scaling up' goes mostly to laboratories, not to production lines. And the unit price is not rising: it is collapsing. The average price of Unitree humanoids fell from 593,400 yuan (approximately 85,000 dollars) in 2023 to 167,600 yuan (approximately 25,000 dollars) in 2025, a drop of 72% in two years[10].
The company has also delivered more than 30,000 quadruped robots between 2022 and September 2025[11] — a real industrial know-how, but which says nothing about the ability of humanoids to perform measurable work. What is being capitalised publicly, in short, is units sold to academic buyers and a free-falling price. Not proven work.
The missing metre: for the first time since 2015, a neutral arbiter
This is where the fact that the whole dance covers up comes in. According to Aaron Prather, director of the Robotics and Autonomous Systems programme at ASTM International, the NIST benchmark project constitutes the first standardised performance reference for humanoid robots since the DARPA Robotics Challenge of 2015 — and he emphasises that there is still no consensus method to measure the capabilities of current platforms, despite billions invested[12]. The finding is stated in black and white: despite the sums sunk into platforms like Tesla Optimus, Figure, Agility Robotics, Apptronik and Unitree, until the NIST project of April 2026, there was no consensus method to compare what these machines can do[13].
The significance of the event lies in the identity of the one speaking. Until now, the standard was written by the manufacturers themselves, judges and parties to their own demonstrations. NIST is neither a manufacturer, nor a shareholder, nor a distributor: it is a neutral third party. And it enters the scene precisely the week a public market seals the value of an industry it declares, in the same breath, not yet measurable.
The order of operations, reversed
The simultaneity is no calendar coincidence. The body counter — Unitree, first by volume, first on the A-share — gets its public listing on the very day an independent arbiter notes that the work counter does not yet exist. Tokyo decrees the scale, Shanghai capitalises it, NIST reminds us that we still do not know how to measure it. Demonstrations are worth a narrative; they are not worth a metric. And the needle threaded on display, like the G1 that 'works' at the airport, proves nothing until a shared instrument comes to attest to it. The industry has listed the scale before inventing the metre — it remains to hope that the two eventually coincide.