Wiring / Thermal

Wiring is the most supply-elastic structural link in this whole report, a humanoid’s wiring-harness set is a thin slice atop a vast, reconfigurable auto base, and on raw volume it does not bind. A robot carries one harness set, and the world already builds about ninety-two thousand thousand of them a year, one hand-built harness set for each of the roughly ninety-two and a half million vehicles made in 2024², against a worldwide robot need that reaches only about forty-seven thousand thousand sets in total by 2040. That is a fraction of a single year’s harness output, so the four-series gap never opens: cumulative global production runs roughly thirty-five times the cumulative robot need at 2040, a coverage of about three thousand six hundred percent, and the global crossover year is null⁷. Building from one bloc barely changes the answer, because harness assembly is labor-led and relocatable in months. China’s slice, the single largest national node, but only about a fifth of global harness production geography⁶, still runs about seven times the cumulative need, so China-only is null too. Only the thin United States slice, roughly two percent of harness output because the work is offshored to low-wage assembly hubs⁵, ever slips behind, and only in the late 2030s, around 2039. And even that is not a real wall: it reflects where cheap assembly labor sits today, not a capacity ceiling, a new harness line stands up in about a year, so a United-States shortfall is a relocation choice, not a volume constraint.

A wiring harness is the bundled set of insulated copper wires, connector terminals, fuses and clips that carries power and signals between a machine’s battery, motors, sensors and controllers, the nervous system that ties every other component together. It is built almost entirely by hand: harness assembly is roughly seventy to eighty percent manual labor, cutting, crimping, routing and taping wires onto a form board⁵, which is exactly why it is the most labor-led and least capital-intensive part in the robot. The model counts one harness set per robot, a modeling convention that mirrors the one-harness-per-vehicle basis of the automotive industry⁴. The line is named “Wiring / Thermal,” and the robot does also carry a thermal-management loop, pumps, valves and cold-plates that move heat away from the actuators, but that loop rides the same kind of abundant, low-barrier supply base as the harness and is not the binding constraint, so the analysis here centers on the harness set. The makers are the global harness houses, Sumitomo Electric, Yazaki, Aptiv and Leoni lead, the first two together a large share of world output³, joined by Furukawa, Lear and others¹.

Because a wiring harness is built into a robot and stays there for its service life, the harness sets the fleet needs are a cumulative stock, the total number of robots ever built worldwide, one set each, not a per-year flow. Integrating the consensus shipment ramp⁷, the worldwide installed fleet reaches about five hundred and eighty-five thousand robots by 2030, ten million by 2035 and forty-seven million by 2040, requiring roughly five hundred and eighty-five thousand, ten million, and forty-seven million harness sets respectively. In the chart below the black line is that worldwide cumulative need, the grey band is cumulative global harness production, and the gap never opens, because a single year of global harness output already exceeds the entire fleet’s lifetime need many times over. That absence of a wedge is the finding, not an error. The other embodiments only pile more onto that same elastic base: a drone too carries one harness apiece, so even a cumulative drone fleet of roughly four hundred and seventy-eight million by 2040 (units built, since about half are single-use military craft) is still a fraction of a single year’s harness output and never moves the picture.

Production is the only thing that differs between the Global, China and US views. The worldwide humanoid fleet and the harness sets it needs stay identical in all three; only the supply line changes, asking whether one bloc’s factories alone could equip the whole global fleet. Global harness output runs from about ninety-two thousand thousand sets a year today, one per vehicle across roughly ninety-two and a half million vehicles², toward a modeled one hundred and twenty-four thousand thousand by 2040, growing the slow few percent a year that the auto base adds (the automotive wiring-harness market compounds at about three and a half percent⁴, and the broader wire-harness market runs from roughly one hundred billion dollars in 2024 toward one hundred and forty-seven billion by 2034¹), plus a small humanoid tail. China is the single largest national node but, contrary to the common impression, only about a fifth of global harness production geography, roughly eighteen thousand thousand sets a year⁶, because harness assembly is sited near final vehicle assembly: Mexico for North America, Eastern Europe, Morocco and Tunisia for Europe, and China, Vietnam and the Philippines for Asia⁵. The frequently-cited larger China figure is consumption, not production⁴. US-located output is the thin line: only about two percent, roughly one thousand eight hundred thousand sets a year, because North America is served almost entirely from Mexican plants at three-and-a-half to four-and-a-half dollars an hour, and US harness work is a niche aerospace-and-defense tier⁵. These post-2030 figures are a modeled extrapolation of the auto base, not a reported forecast, and crucially, none of these shares is a ceiling: they are snapshots of where cheap assembly labor sits.

For wiring, the crossover, the year cumulative production drops below cumulative need⁷, never occurs globally, and never occurs for China either. Cumulative global output reaches something like one million seven hundred thousand thousand harness sets by 2040 against a fleet need of only about forty-seven thousand thousand; the robot slice is simply too small a share of a huge, established market to overrun it, so both the global and China crossover years are legitimately null. The only view where cumulative production slips behind is the US-only slice, and even then it comes late, around 2039, and means something different than it does for the grinding-gated parts. The US slice is thin only because harness assembly has been offshored to low-wage hubs⁵, not because the United States lacks the capacity to make wire and crimp connectors. So when cumulative US-located output finally trails the worldwide fleet’s cumulative need in the late 2030s, that is a relocation choice, not a volume wall: a new harness line stands up in about a year, brownfield in roughly half that, so any bloc that decided to supply itself could close the gap inside the lead time of a single product cycle. The honest contrast with the precision binders is the whole point, roller screws or cross-roller bearings bind early and hard because their lines take years to build, whereas a harness line is the easiest factory in the entire bill of materials to clone.

Wiring is the link that essentially does not bind, and it is worth being explicit about why. Walk down the supply chain and there is no choke point: the inputs are commodity copper wire and connector terminals, both made at enormous scale, and the assembly step is low-cost manual labor, cutting, crimping, routing and taping, about seventy to eighty percent done by hand⁵. There is no precision-grinding gate, no rare-earth feedstock, no multi-year tool qualification; wiring is not part of the shared precision-grinding pool that throttles the harmonic and roller-screw lines, so its effective output equals its standalone output with no pool drag. The one input that matters, low-wage assembly labor, is geographically mobile, which is exactly why the industry already migrates between Mexico, Eastern Europe, Morocco, Tunisia, China, Vietnam and the Philippines as labor costs shift⁵. A greenfield harness plant takes about a year to build and a brownfield line about half that, against a global base that already turns out tens of thousands of thousands of sets a year and a market worth roughly one hundred billion dollars and growing at a steady few percent¹. So the four-series picture is unambiguous: a humanoid harness set is a thin, late-arriving slice atop a vast, reconfigurable automotive base, the cumulative robot need stays a small fraction of even one year’s output through the whole horizon², and the only thing that ever “crosses” is a single bloc’s self-imposed share, which it can reverse in a year if it chooses to. Of every structural link in this report, wiring is the one the fleet will never have to wait on.