DESIGN INTELLIGENCEJuly 1, 2026·Mary · DEPIX Design Intelligence

The spot weld is dying. It decided what cars could be.

There are roughly four to five thousand of them in a conventional steel car body, fired in under six seconds each by four hundred robots on a body-in-white line. The resistance spot weld is the most-repeated manufacturing decision in the industry, and it is quietly being retired. Nobody in a design studio voted on this. They should have, because the way a car is joined has always decided what a car can be.

Here is the problem the weld ran into. A resistance spot weld works by passing current through two overlapping sheets until they fuse. That trick only works cleanly on similar metals. The moment a body goes multi-material — aluminium castings bonded to high-strength steel, extruded sections, composite panels, the whole lightweighting palette every EV programme now reaches for — the current meets a metallurgical wall. Weld aluminium to steel and you grow brittle iron-aluminium intermetallic compounds at the joint, a seam engineered to crack. So the industry has spent a decade quietly building a second toolbox: self-piercing rivets, flow-drill screws, structural adhesives, friction element welding, and the newer solid-state methods — friction stir and refill friction stir spot welding, which stir two aluminium sheets together below their melting point and leave no hole behind.

Each of these is not a like-for-like swap. Each rewrites the geometry the designer is allowed to draw. A self-piercing rivet needs a wide flange and access from both sides. A flow-drill screw is single-sided, which is the only reason it can close an aluminium extrusion or a sealed box section a rivet gun can never reach. Structural adhesive wants a continuous, clean, generously-sized bond line and cure time, and it will not tolerate the tight radius a spot gun shrugged at. Refill friction stir demands flat, accessible aluminium and, for now, pays for its beautiful hole-free joint in cycle time and tool wear. The Audi TT famously held its body together with eight different joining methods at once. The aluminium-intensive F-150 bonds, rivets and flow-drills the same seams. This is not one method replacing another. It is a committee.

And that committee sits down at the concept phase, whether design attends or not. The join method dictates which material combinations are even possible, which dictates section thickness, flange width, edge access and panel geometry — which dictates form. A designer who wants a knife-edge shut line, a section that vanishes into a wheel arch, a flush multi-material transition, is really asking a question about joining that gets answered months later by a manufacturing engineer optimising for weld count and cycle time. By then the surface is frozen and the answer is no.

The honest boundary: none of this is bad engineering. Cold and solid-state joining is why mixed-material bodies are lighter, stiffer and safer than the all-steel monocoque ever was; friction stir is greener and stronger on aluminium than the weld it replaces. The point is ownership. For a century the spot weld was a constraint so universal it was invisible — it silently pre-edited every body, saying no to the unbuildable seam before anyone costed it. Retire that single universal rule and you inherit a dozen narrower ones, each with its own geometry tax, and no single person in the building is tracking how they add up into a shape.

That accounting has to move upstream. The material palette a car can wear is set by what the factory can join, and that decision is made — invisibly, in a different building — at the exact moment the form is still soft enough to change. This is precisely where concept-phase intelligence earns its keep: pressure-test the boldest version of a body against the joins it will actually demand, across real material transitions and real access, before the flange widths and bond lines are tooled into permanence. The question the studio used to ask was "can the line weld this," and the line always answered. The new question is "what can this collection of joins let the car become" — and it is only answerable while the surface can still move.

The weld is dying. It was never neutral. It was the oldest, quietest co-author in the studio, and its replacement writes in a language design does not yet speak.