Toyota's 1 trillion yen leaves 2.5 million yen a robot for 400,000 humanoids

Divide 1 trillion yen by 400,000 units and you get 2.5 million yen a robot ^[2], and that figure still has to cover the advanced production lines the money is nominally for ^[2]. So either the humanoids are cheap, or 1 trillion yen is the first tranche of a longer bill. Spread across the 60 plants the programme targets ^[5], the same money is roughly 16.7 billion yen per site ^[3], and the same unit count is about 6,700 machines per site ^[4].

Toyota takes 150,000 units into its own plants and puts 250,000 into affiliates ^[3], so 62.5 percent of the installations happen inside factories Toyota does not run ^[1]. The machines are learning frontline workers' movements on manufacturing lines now ^[4].

On the supply side there is one stated annual rate. Tesla began building a line in July for more than 50,000 Optimus robots a year at a U.S. plant and plans to extend deployment to gigafactories worldwide ^[6]. At 50,000 a year, Toyota's 400,000 is eight years of that single line ^[5]. Hyundai's own programme is 30,000 Atlas units in mass production in the United States by next year, then staged deployment at global sites ^[11], about a thirteenth of Toyota's count ^[6]. BYD is testing robots in transport, labeling and inspection work ^[7], and BMW has begun putting humanoids on its production lines ^[8].

The Hyundai union's position, as reported by the Seoul Economic Daily, is that "not a single robot can be deployed without agreement between labor and management" ^[9]. The paper attributes that leverage to the revised Trade Union Act, the so-called yellow envelope law, under which changes in working conditions are subject to bargaining and industrial action ^[10], and it says the union may yet obstruct the Atlas rollout ^[12]. The report cites no deployment the union has already blocked ^[15]. President Lee Jae-myung said "you cannot avoid the great cart" in stressing coexistence with robots ^[13]. The paper's own remedy is an overhaul of the law ^[14].

The Seoul Economic Daily writes that success or failure in advanced industries depends on speed, and that Korea will fall behind if union obstruction combines with policy misjudgment by the government ^[16]. I'd expect unit supply to set the pace of all this and labour consent to set the order. The counter-thesis: if every line change at Hyundai is a bargaining item ^[10], the union does not need to win a formal veto to cost the company years, and a plant that waits while a rival re-lays its lines pays for the wait in unit cost.

A Toyota supply contract at an annual build rate near its own requirement would break that reading, because it would move the scarce thing from robots back to consent. So would a Hyundai labour agreement that sets a price for consent per unit.