On 12 April 2026, a Unitree G1 humanoid robot named Edward Warchocki chased a group of wild boars across a car park in Warsaw. The video, published the same day on its own X account, surpassed 3.8 million views on the platform and more than 11 million on Instagram within three days, before being picked up by CNN, ABC News, USA Today, Euronews, Le Parisien and Fox Business. Over the course of a weekend, a 132 cm, 35 kg device made in China became, according to the Polish press, the country’s new influencer.
One could stop there and debate what makes this robot endearing. One could also ask a less flattering question: what exactly is China selling to Europe this week? Because the answer, once the week’s three key threads are cross-referenced, is not what one might expect. The world’s ‘number one’ in humanoid volume is not exported to the West as a workforce. It is exported as a rental entertainment object.
The hard figure of the week is a view count
Usually, the metric that matters in this sector is a production metric: number of units delivered, operating hours, bins moved. This week, the most prominent figure is an audience counter. And it is no coincidence: the wild boar hunt on 12 April was not an urban wildlife management operation, but a promotional stunt, as established by several international press sources. The wild boar was not a problem to solve; it was set dressing.
The shift is total. The value produced by Edward Warchocki that weekend is not measured in work accomplished but in attention captured. Its first commercial contract confirms this without ambiguity: promoting a luxury watch estimated at around 80,000 zlotys, which is a full and proper entry into influencer marketing. The robot produces nothing; it embodies a brand.
The importer is not an integrator, it is an agency
The link that brings this robot to Europe is also revealing. Edward Warchocki was imported to Poland by the company MERA Robotics. And MERA is not announcing an industrial integration activity: it plans to import around 100 humanoid robots from China by the end of July 2026, to deploy them in event and promotional activities.
The product sheet speaks of a factory; the order book speaks of galas. Between the promise — logistics, security, commerce — and what is actually monetised, there is the entire gap between an integrator and an events agency. What is truly deployed and billed is the robot itself as an attraction, and the advertising it carries. The export channel for Chinese volume to Europe, this week, is not a production line: it is Instagram.
The metric that folklore buries
You have to leave the Warsaw car park to see where the real market lies. On 9 April, TrendForce estimates that Unitree Robotics and AgiBot will together account for nearly 80% of global humanoid robot deliveries in 2026, in a market that will exceed 50,000 units, representing an increase of more than 700% year-on-year. Unitree alone aims for 20,000 units — compared with around 5,500 delivered in 2025 — its CEO Wang Xingxing citing a range of 10,000 to 20,000, with a combined gross margin of 60% on its humanoid and quadruped segments. Chinese volume is therefore real, and massive.
But the same week, in Boston, the State of Humanoids panel at the Robotics Summit delivers the counter-shot. Its speakers identify the cost of the perception stack — cameras and sensors — as the major obstacle to the economic viability of humanoids: this item alone approaches $20,000, equivalent to the target cost of the entire robot. Seeing and understanding one’s environment currently costs as much as the entire device. And Mike Nielsen, from RealSense, adds that product development cycles in China are 30 to 40 times faster than elsewhere, citing precisely these robots capable of judo and dance as symptoms of this acceleration.
The week’s great disconnect
Laid end to end, these facts outline an irony that none articulate in isolation. While in Boston industry figures admit that a humanoid that works is not yet economically viable — its perception alone costs as much as the whole robot — a Unitree that does not work at all racks up tens of millions of views hunting wild boars. Dance, judo, chase scenes: these are precisely the demonstrations that China knows how to produce 30 to 40 times faster, and these are the ones that cross the border.
The most-watched robot of the week is, literally, the one that does the least work. This is not an accidental paradox: it is the form that export takes. Chinese volume does not arrive in Europe as a worker, because the worker-robot remains too costly to exist at scale. It arrives as a promotional product, sold by a reseller that disguises an events agency as an industrial integrator.
The useful question is therefore not whether Edward Warchocki is endearing, nor whether China dominates the sector — it dominates in volume, the figures leave little doubt. The question is to look at what it is truly exporting to Europe this week. And the answer comes in one line: not a worker, but an influencer for hire at your galas — at the very moment when, in Boston, those building the working robot acknowledge that it still costs too much to exist.