[Korea 1/4] From Wolfsburg to Ulsan: Who Still Has the Power to Say No?
In Korea, the world’s first strike against humanoid robots
In Korea, Hyundai’s metalworkers’ union is about to launch daily walkouts, following a strike vote passed by an overwhelming majority on an issue that, for the first time, centers on the entry of humanoid robots into factories. At the same moment, in Wolfsburg, Volkswagen’s supervisory board finds itself blocked by its own employee representatives. Two blockages, on two continents, for two different reasons — which say the same thing about what remains, in 2026, of workers’ bargaining power in the face of the global automotive industry’s restructuring.
On July 9, in Wolfsburg, Volkswagen’s supervisory board met for nearly five hours to rule on CEO Oliver Blume’s “Zielbild 2030” plan — up to 100,000 jobs potentially cut worldwide, the possible closure of four to five German plants. We covered the details this same morning: the board approved the broad outlines of the plan (a smaller model range, reduced capacity), but postponed the most painful decisions — plant closures, the exact scale of job cuts. This blockage is no accident of timing: it stems from the very architecture of power at Volkswagen, where German codetermination gives employee representatives a strict parity of ten seats out of twenty on the supervisory board — a legal right to block, not merely a right to negotiate.
On the other side of the world, another standoff is playing out right now, for a very different reason — but with the same question in the background: who has the power to say no?
An unprecedented strike, passed by an overwhelming majority
The precise figures deserve to be laid out before anything else. The Hyundai Motor branch of the KMWU (Korean Metal Workers’ Union) counts 39,668 members; 94.15% of them took part in the vote, and 92.03% of those who voted — roughly 86.6% of the entire membership — came out in favor of a strike, after eleven rounds of wage negotiations without an agreement. The vote itself is nothing exceptional: according to the Korean business press, a strike authorization has never failed at Hyundai, in any year. What is less ordinary is what has just followed: according to The Korea Times, the union has announced that starting Monday, July 13, it will launch two-hour partial walkouts every day — the same week that, at GM Korea, the local union (94.96% in favor) is separately refusing early shifts, overtime, and weekend work. At Kia, five rounds of negotiations have so far failed to bridge the gap on a similar demand. If the walkouts go ahead, this will be the second consecutive year of concrete union action at Hyundai, following partial stoppages in the fall of 2025 — the first in seven years. The figure of 73,000 employees sometimes cited in the French press actually refers to the entire workforce of the Hyundai Motor Group, not to the KMWU members who voted.
The bonus being demanded — 30% of net profit — is itself nothing new: the union has pushed for it almost every year since 2004, except in 2020. It simply found a second wind after the agreement reached this spring by Samsung Electronics employees, who negotiated 10.5% of operating profit as a performance bonus — a reference the Hyundai union explicitly cites to justify its own demands. What has genuinely changed this year, according to The Korea Herald, are the job and working-condition guarantees tied to artificial intelligence and robotics: an issue that had never before surfaced in the group’s previous wage negotiations. The trade publication Automotive World goes further, calling this AI/robotics clause the “most structurally unprecedented” element of this year’s negotiations.
One last piece of legal context sheds light on this balance of power: the so-called “Yellow Envelope” law, which took effect in March 2026, broadened the legal definition of labor disputes to include business decisions affecting working conditions — giving the union a legal basis it did not have in previous years to argue that profit-sharing is negotiable. Labor lawyers dispute this reading, arguing that profit allocation remains a prerogative of management and shareholders; the debate therefore remains open, but the mere existence of this new legal lever is not trivial.
“We define this fight as a battle for our future right to survive, and we cannot and will not back down a single step,” warned Lee Jong-chul, head of the union’s Hyundai branch.
Veto power over robots, guarantees on batteries
The demands go well beyond a year-end bonus. The KMWU is demanding oversight — in effect, veto power — over any introduction of robots or artificial intelligence on the line: not a single robot will enter without an agreement between management and workers, it has warned. The union is also proposing to replace the current hourly wage with a fixed monthly salary, a way of pre-empting the drop in human working hours that automation will mechanically bring about. At Kia, the demand goes even further: that strategic electric-vehicle components — batteries, motors, reducers, fuel-cell stacks — remain mandatorily produced in Korea, precisely to prevent automation from being quietly accompanied by a shift of production abroad.
Twenty-five thousand robots, a lights-out factory by 2030
At stake are the Atlas robots, the humanoid built by Boston Dynamics, a Hyundai subsidiary since 2020. The timeline announced by the group is precise: parts sequencing starting in 2028, then full vehicle assembly starting in 2030 — with, ultimately, more than 25,000 units deployed across the group’s sites. The stated goal is a factory capable of running 24 hours a day, seven days a week, even with the lights off. This announcement comes as Hyundai’s operating profit declines under the combined effect of U.S. tariffs and increasingly intense Chinese competition — a context we will detail in a forthcoming article on the group’s industrial strategy.
Hyundai defends a very different reading of this same timeline. The company insists that the robots will be confined to the arduous or dangerous tasks that employees themselves are reluctant to take on, and that no job cuts are planned in the near term — the classic argument in any management defense of automation. The union rejects this outright, fearing instead what it calls an “employment shock.” At this stage, neither side has offered hard figures capable of settling which of the two scenarios will actually play out.
Blocking upstream, negotiating downstream, or having no leverage at all
This is where the parallel with Wolfsburg takes on its full meaning — not because the two disputes are about the same thing, but because they each reveal, in their own way, a different architecture of the power to say no in the face of the same pressure: the restructuring of the global auto industry under Chinese pressure.
In Germany, that power is institutional: it operates upstream, within the very body that makes the decision. The VW law and codetermination (Mitbestimmung) give workers a seat on the supervisory board, not merely a right to negotiate after the fact. That is what explains how a plan as heavy as Oliver Blume’s can remain blocked at the top, before it even reaches the shop floor.
In Korea, that power is union-based, in the classic sense: collective bargaining, mandatory mediation, strike votes. It is real, and it can bring 40,000 workers to a standstill — but it operates downstream, only once management has already announced its decision, never before.
In Georgia, in the United States, at the Metaplant site that Hyundai presents as the showcase of its robotics success, that power simply does not exist: there is no board seat, no union. One fact is enough to grasp what this means in practice: it was the very same Korean engineers, sent from home to build that showcase, who found themselves with no recourse whatsoever during the raid conducted by U.S. immigration authorities in September 2025 — protected, back home, by a body of labor law built up over decades, and left utterly unprotected the moment they stepped outside it.
Wolfsburg and Ulsan hit the same wall, the same month
Nothing requires Wolfsburg and Ulsan to echo each other in the same month — and yet everything points that way: both managements invoke the same competitive urgency (overcapacity for one, a technological lag for the other), both run into the same wall — their employees’ capacity to weigh on a decision already made at the top — and both are measuring, each at its own scale, the cost of trying to move faster than labor law allows.
Social dumping, union coordination, or a sovereigntist retreat?
The BYD-ization of the global auto industry, which we have so far mostly followed from the production side — factories, robotization, innovation, the pincer effect between budget and luxury segments — also has a social dimension. It is also playing out on another terrain: that, a deeply unequal one, of the bargaining power workers have inherited depending on the country where they live and work — and of the choice, rarely spoken aloud, made by management to accelerate robotization precisely where that power is weakest.
BYD-ization — A concept formalized and tested across an earlier series of articles. It describes the phenomenon by which Chinese companies (BYD, Huawei, CATL, DeepSeek) do not merely compete with established players, but redefine the very norms of competitiveness in their sectors. The concept rests on eight interdependent mechanisms: vertical integration, the pincer effect, accelerated cycles, standardized diffusion, technological substitution, the domestic market as a springboard, public coordination, and the offer of complete systems.
Three possible consequences of this observation deserve, at this stage, to be raised as questions rather than as predictions.
The first is a risk of social dumping, robot-style. If workers’ blocking power effectively becomes a criterion for industrial location on a par with labor costs, nothing would stop auto groups from choosing tomorrow’s sites for the most radical automation based on the weakness of local labor law — Metaplant rather than Ulsan, for reasons that would have little left to do with logistics or proximity to the U.S. market.
The second is the reverse possibility: a response through coordination. IG Metall and the KMWU have, to date, never coordinated on this specific terrain — even though each holds a piece of the problem the other might want to import back home: Korea’s veto right over the entry of robots would surely interest IG Metall just as much as Germany’s institutional seat on the supervisory board would interest the KMWU.
The third, less visible but historically at least as significant in both countries, is that of industrial sovereignty. In Germany, the VW law itself was born of this very concern: a mechanism designed so that no shareholder, be it a sovereign fund or a family, could alone decide the fate of a national champion — and the Lower Saxony state government, in opposing plant closures, is likely speaking as much in the name of that sovereignty as in the name of workers. In Korea, where Hyundai’s rise remains inseparable from the national narrative of industrial catch-up launched in the 1960s, and where national sentiment remains a particularly sensitive nerve, Kia’s demand for domestic production of batteries and motors may not be merely one more union clause: it could also be, potentially, a refusal to let the technological heart of the Korean car become a component outsourced to Chinese suppliers. One question remains that nothing, at this stage, allows us to settle: will this sovereigntist strand strengthen the workers’ camp in both countries, or could it instead break free of it — with a state or national capital deciding, one day, that radical automation, even at the cost of jobs, is the lesser evil against a growing technological dependence on China?
Sources: The Korea Times, The Korea Herald, Automotive World, Financial Times, Le Monde (July 10, 2026), Manager Magazin, NDR, WirtschaftsWoche.
All our publications are free and Copyleft. They may be republished, in whole or in part, under following conditions:
→ keeping the title and content
(text, images, editor’s or translator’s notes, boxes)
→ crediting the author(s) and/or translator(s)
→ inserting a clickable link to the source.
Any monetization or commercial use of these publications is prohibited and will be subject to legal action.