On 12 April 2026, a Unitree G1 humanoid robot named Edward Warchocki chased a group of wild boar across a Warsaw car park. The video, posted the same day on its own X account, passed 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 a single weekend, a 132 cm, 35 kg device made in China had become, in the words of the Polish press, the country's newest influencer.
One could stop there and debate what makes this robot so 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 threads are pulled together, is not the one you would expect. The world number one in humanoid volume is not exporting itself to the West as a workforce. It is exporting itself as a piece of entertainment for hire.
The week's hard number is a view count
Normally, the metric that matters in this sector is a production metric: units delivered, hours of operation, totes moved. This week, the figure pushed hardest is an audience counter. And it is anything but trivial: the boar chase of 12 April was not an urban wildlife management operation but a promotional stunt, as several international news outlets have established. The boar was not a problem to be solved; it was a prop.
The shift is total. The value produced by Edward Warchocki that weekend is not measured in work done but in attention captured. Its first commercial contract confirms this without ambiguity: promoting a luxury watch valued at around 80,000 zlotys — in other words, a full and outright entry into influencer marketing. The robot produces nothing; it endorses a brand.
The importer is not an integrator, it is an agency
The link that lands this robot in Europe is equally telling. Edward Warchocki was imported into Poland by the company MERA Robotics. And MERA is not announcing an industrial integration business: it plans to import around 100 humanoid robots from China by the end of July 2026, to deploy them in events and promotional work.
The spec sheet talks about the factory; the order book talks about the gala. Between the promise — logistics, security, retail — and what is actually being monetised lies the whole gap between an integrator and an events agency. What is genuinely deployed and invoiced is the robot itself as an attraction, and the advertising it carries. The export channel for Chinese volume into Europe, this week, is not a production line: it is Instagram.
The metric the folklore buries
You have to leave the Warsaw car park to see where the real market is being decided. On 9 April, TrendForce estimated that Unitree Robotics and AgiBot together will account for nearly 80% of global humanoid robot shipments in 2026, in a market that will exceed 50,000 units — an increase of more than 700% year on year. Unitree alone is targeting 20,000 units — against roughly 5,500 delivered in 2025 — with its CEO Wang Xingxing citing a range of 10,000 to 20,000, and a combined gross margin of 60% across its humanoid and quadruped segments. Chinese volume, then, is very real, and massive.
But that same week, in Boston, the State of Humanoids panel at the Robotics Summit delivered the reverse shot. Its speakers identified the cost of the perception stack — cameras and sensors — as the major obstacle to the economic viability of humanoids: this item alone comes close to 20,000 dollars, the equivalent of the target cost of the complete robot. Seeing and understanding its environment costs as much today as the entire device. And Mike Nielsen, of RealSense, adds that product development cycles in China run 30 to 40 times faster than elsewhere, citing precisely these robots capable of judo and dance as symptoms of that acceleration.
The week's great divide
Set end to end, these facts trace an irony that none of them, taken alone, spells out. While in Boston the industry admits that a humanoid that works is not yet economically tenable — its perception alone costing as much as the whole robot — a Unitree that does no work at all is racking up tens of millions of views chasing wild boar. Dance, judo, the chase: these are precisely the demonstrations China can 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 works the least. This is no accidental paradox: it is the form the export takes. Chinese volume does not arrive in Europe as a worker, because the robot-worker remains too expensive to exist at scale. It arrives as a promotional product, sold by a reseller dressing up an events agency as an industrial integrator.
The useful question, then, is not whether Edward Warchocki is endearing, nor whether China dominates the sector — it dominates it in volume, and the figures leave little room for doubt. The question is to look at what it is really exporting to Europe this week. And the answer fits in a single line: not a worker, but an influencer to hire for your galas — at the very moment when, in Boston, those building the robot that works admit it still costs too much to exist.