Porsche Design and Puma: How a 3D Printed Shoe Sole Works

• 1
  What Porsche Design and Puma Actually Changed
• 2
  How a Printed Lattice Carries Load
• 3
  Powder-Bed Printing: What Sets the Real Limits
• 4
  Material Choices and What They Cost You
• 5
  Post-Processing Decides the Feel
• 6
  Printed Lattice vs Foam vs Machined Solid

In March 2023, Porsche Design and Puma released a running shoe with a 3D printed sole. It was the first time the two brands used printing for a sports shoe sole. The interesting part is not the brand pairing. It is that a printed lattice replaced a foam midsole, which means the cushioning behavior is set by geometry instead of by a chemical blowing agent.

A foam midsole is tuned by density and cell size. A printed lattice is tuned by strut diameter, cell angle, and the number of cells per unit area. Change the strut from 0.8 mm to 1.2 mm and you change the load curve. That is a design parameter you can hold in a CAD file, so two production runs behave the same way. Foam tooling cannot do that.

The sole still has to survive real loading. A runner puts roughly two to three times body weight through one foot at midstance. The lattice has to compress, return, and not take a permanent set. That is a fatigue problem as much as a stiffness problem.

So the collaboration matters less as a fashion story and more as proof that a printed lattice can pass wear testing. That is the claim engineers care about.

A lattice is a repeating network of struts. Load enters through the top nodes and leaves through the bottom nodes. Each strut sees a mix of compression, bending, and shear. The strut angle decides which one dominates. Steep struts carry compression well. Shallow struts bend and give you softer travel.

Cell size sets how many struts share the load. A 4 mm cell puts more struts under the foot than an 8 mm cell, so the same strut diameter feels stiffer in the finer lattice. This is why you tune cell size and strut diameter together, not one at a time.

Density gradient is the second lever. A heel needs more support than a forefoot. You can thicken the struts under the heel and thin them toward the toe inside one build. No extra tooling. No extra assembly step.

There is a limit. Below about 0.6 mm strut diameter, most powder-bed processes cannot hold the shape reliably. The strut may print but it will not survive handling. Above roughly 2 mm, the lattice behaves like a solid block and you lose the cushioning you wanted.

Most printed midsoles come from a powder-bed process, either laser sintering or a binder jet variant. A roller spreads powder, a heat source fuses a cross-section, the bed drops, and the cycle repeats. Layer thickness usually sits between 60 μm and 120 μm.

Layer thickness drives three things at once. Thinner layers give a smoother surface and better strut resolution, but the build takes longer. Thicker layers run faster and cost less per part, but small features blur. For a lattice with 0.8 mm struts, 80 μm is a reasonable middle.

Powder reuse is the hidden cost driver. Fused powder can be sieved and blended back, but the refresh ratio matters. If you must add 50 percent virgin powder each cycle, your material cost per part doubles compared with a 20 percent refresh. Ask any supplier what their refresh ratio is before you compare quotes.

Thermal distortion is the other boundary. Long, thin, unsupported sections curl as they cool. Lattices are mostly self-supporting, which is why they print well. A tall solid fin on the same build plate is a different story. Orientation is not optional.

Printed midsoles are usually a thermoplastic elastomer or a TPU-family powder. The material has to flex millions of cycles without cracking. That rules out most rigid plastics, no matter how cheap they are.

TPU powders hold up well in flex fatigue and come in a range of shore hardness values. PA12 and PA11 are stiffer and cheaper, and they suit structural parts rather than cushioning. If you want a shoe sole that feels like a sole, you pay for elastomer powder.

For non-cushioning prototypes, resin printing is faster and gives a smoother surface. The trade-off is brittleness. A resin lattice will look right in a photo and crack on the third flex. Use it for form studies, not for wear testing.

If your part has to take real load and also look finished, machining may beat printing. We machine 6061, 7075, 316L, and Ti-6Al-4V to ±0.005 mm on 5-axis centers, which is a different order of tolerance than any powder-bed process can hold.

A printed part comes out of the build with loose powder in every gap. Bead blasting or media tumbling removes it and knocks down the surface texture. Skip this step and the lattice sheds powder inside the shoe.

Surface texture also sets friction. A rough lattice grips; a polished one slides. For a sole, some texture is useful. For a sliding wear pad, it is a liability. Decide which one you need before you specify a finish.

Color is the last step and the easiest to get wrong. Dyeing a porous printed part gives uneven results because absorption varies with local density. Painting adds a layer that can crack at high flex. If color matters, plan for it at the design stage.

We run bead blasting, tumbling, brushing, and polishing in house, plus anodizing, plating, powder coating, and laser marking down to 1.5 mm character height. Those finishes apply to machined parts. Printed elastomer parts get mechanical finishing only.