The Leap Barra and Fain Won’t Make

Source: “GM Cut 1,000 Workers at Its EV Plant, Then Added Robots,” Autoblog, June 20, 2026.

Two people looked at the same fifty robots at GM’s Factory Zero this month and described two different universes. Mary Barra has said AI “empowers our workforce to focus on craftsmanship,” and that by “merging technology with human ingenuity” GM unlocks new levels of innovation. Shawn Fain told the UAW convention that artificial intelligence and humanoid robots are a profound threat — that workers are in a fight for humanity, and that if the technology keeps being used this way, it has to be stopped. Same cobots. Same assembly line. Two universes.

They are going to spend the next two years proving how far apart those universes are, all the way to the 2028 bargaining table. And here is what almost no one writing about this week’s headline has said out loud: the fight that is coming is unnecessary, and the two of them cannot stop themselves from having it.

Start by clearing the ground, because the headline itself is false. Fifty robots did not replace a thousand workers; the arithmetic alone kills that. The thousand went over the past year because Washington pulled the floor out from under electric vehicles — the $7,500 consumer credit repealed, the federal emissions standards rescinded, EV demand in the United States actually falling in 2025 while it rose everywhere else in the world. Factory Zero builds nothing but EVs. The cuts were a policy story. The robots are a separate, smaller thing — fifty cobots bolting panels alongside the people who remain. That is augmentation, not replacement. So the robot fight isn’t really about these robots. It’s a rehearsal for the real one in 2028.

Let me make a prediction I would stake my name on. The UAW knows it cannot fully resist automation, because no one ever has. So it will not try to. It will extract a price: sharp wage increases and job-security guarantees bolted to the machines — some modern descendant of the Jobs Bank, the 1984 provision that paid laid-off members close to full wages to sit in rooms doing nothing. The Jobs Bank felt like a labor victory for two decades. It was a slow death. It helped price the Detroit Three out of their own market and was finally killed in the 2009 bankruptcy. I do not need to read about that one. I watched Buick City close in 1999 with that cost structure on its back. The union extracts a price for the robot; the price gets paid until it cannot be; the plant closes anyway. A wage rent does not compound. It just runs until the money stops.

And the maddening part — the part that should keep a thoughtful person up at night — is that the better answer is not theoretical. It has been built, twice, on two continents, and it is sitting in plain sight.

Go to Amberg, in Bavaria. Siemens runs an electronics plant there that is roughly three-quarters automated, turns out around sixteen million units a year across more than a thousand product variants, and holds a defect rate near eleven per million — manufacturing five to ten times faster than a conventional line. It is the most automated plant of its kind in Europe. Now ask the question Fain would ask: what happened to the people? They are still there. Siemens did not thin the floor; it moved the floor up — into monitoring, programming, optimization, the judgment the machines cannot supply — and it invested in the retraining to get them there. And here is the part that should stop Fain mid-sentence: those workers are not non-union. They are IG Metall — the most powerful industrial union on earth — and under German law, half the seats on Siemens’ supervisory board, the body that signs off on the machines, are elected by labor. The union sits in the room where the automation is decided. That is not less union power than the UAW has; it is far more. Codetermination was written into German law in 1976 with an explicit purpose — to make investment in human capital profitable — and it does exactly that: automation went up, and the workforce went up with it, in skill and in pay, because the people facing the machine had a seat at the table where it arrived. That is the formation route. It does not refute the union. It refutes the belief that a union must fight the machine.

Then come home to Cleveland, because the second proof is American, non-union, and almost a century old. Lincoln Electric has paid its production workers by what they make — piecework — since 1914, layered a merit bonus on top since 1934 that in good years has approached a worker’s annual base pay, and has not laid off a single worker in its U.S. operations since 1948. Its people are roughly twice as productive as the industry and some of them clear six figures on a factory floor. And here is the detail that should stop Shawn Fain cold: James Lincoln understood that a worker only welcomes a new machine when his job is secure. Security is what dissolves the resistance to the tool. Lincoln solved in 1948 — in Ohio, with no union, no co-determination law, no Germany — the exact problem the UAW is preparing to fight to perdition over in 2028. Give a worker security and a real share of what he produces, and he will pull the robot onto the line himself.

Look at what the two factories share, because that is the whole argument. One of them is union to its core; the other has no union at all. That is the tell: the variable that decides whether the leap is possible was never union or no union. Amberg makes the leap on co-governance — a seat at the table where the machine is chosen. Lincoln makes it on trust — security deep enough, and a share real enough, that the worker pulls the robot onto the line himself. The UAW has neither the seat nor the trust, which is why it is left holding the only lever that remains: the strike, and the price. In both factories, pay rises because capability rises. Amberg’s people earn more because they are worth more; Lincoln’s ea