Mercedes ships axial flux at scale — EV motor topology finally shifts

What happened

Mercedes-Benz announced large-scale series production of an axial flux electric motor at its Berlin-Marienfelde plant, marking the first time a major automaker has industrialized this topology for road-going passenger vehicles. The motors are built by YASA, the Oxford-based startup Mercedes acquired in 2021, and will go into the upcoming AMG GT XX concept-derived performance EVs.

The specs are the headline: Mercedes claims the new axial flux unit weighs roughly one-third of an equivalent radial flux motor while delivering comparable or higher peak power — power density figures in the YASA literature hover around 10 kW/kg versus 3-4 kW/kg for typical premium radial flux units. The Berlin facility is purpose-built for the geometry, with new automated stator winding and rotor assembly lines that YASA has spent five years getting out of the hand-built prototype phase.

This is the long-promised payoff of the 2021 acquisition. YASA spent a decade as a hypercar supplier — Ferrari SF90, Lamborghini Revuelto, Koenigsegg Gemera — selling motors at five-figure unit prices because every one was effectively bespoke. The Mercedes deal was always a bet that the topology could be made manufacturable. Berlin is the proof.

Why it matters

For 130 years, electric motors have been overwhelmingly radial flux — the magnetic field flows perpendicular to the axis of rotation, between a cylindrical rotor and a cylindrical stator. It's a well-understood geometry that maps cleanly to automated copper winding and lamination stamping. Every Tesla, every BYD, every Hyundai EV ships radial flux because the manufacturing infrastructure is mature and cheap.

Axial flux rotates the geometry 90 degrees: the magnetic field flows parallel to the axis, between disc-shaped rotors and stators stacked like pancakes. The pancake geometry gives axial flux a structural advantage radial flux can never match — torque scales with the square of the diameter, not linearly, because the active magnetic surface area grows with radius rather than length. That's the physics behind the 3x power density. The catch has always been manufacturing: winding a flat stator with tight tolerances is harder than winding a cylindrical one, and the magnets sit in tension rather than compression, which complicates rotor assembly.

The community reaction on Hacker News is mostly engineers who've followed YASA for years finally seeing the topology cross the chasm. The top-voted comments dig into the trade-offs: axial flux has worse thermal characteristics in sustained high-load operation because the heat path through a pancake stator is shorter but has less surface area to a coolant jacket, which is why early YASA motors had short duty cycles before derating. Mercedes' Berlin design appears to use direct oil cooling sprayed onto the stator end-windings — a technique borrowed from the Formula E powertrains YASA has been supplying.

The interesting question is who follows. Renault's Ampere division demoed an axial flux concept in 2024. Stellantis has a partnership with Equipmake. Magnax in Belgium is supplying smaller players. The Mercedes announcement effectively de-risks axial flux for every other OEM's planning committee — the question moves from "can this be mass-produced" to "can we afford not to." For Chinese OEMs in particular, who compete primarily on cost per kWh of battery, a 70% reduction in motor mass means smaller batteries for the same range, which is a direct cost lever.

What this means for your stack

If you're not building cars, this still matters — and probably matters more, because the cost curve flips faster in lower-volume markets. The YASA Berlin line is rated for hundreds of thousands of units per year, which is the volume threshold where axial flux motor pricing collapses from "hypercar bill of materials" to "premium industrial component"— a 5-10x cost reduction over the next 24 months is the realistic projection from the supply chain analysts.

For robotics, this is the single biggest joint-actuator development since the cycloidal gearbox got cheap. Humanoid robots (Figure, 1X, Unitree, Tesla Optimus) are mass-constrained at every joint — every kilogram of motor mass at the hip costs three at the battery to move it around. A 3x power-density jump in the actuator translates almost linearly to runtime or payload. Watch for the first humanoid demos quietly switching topology in late 2026.

Drones and eVTOL are already heavy axial flux users at the high end — H3X, MagniX, and Wright Electric all bet the company on the topology — but they've been paying YASA prices. A commodity supply chain means heavier payload eVTOLs become economic. For terrestrial e-bikes and cargo bikes, the weight reduction at the wheel is felt more than anywhere else in the design.

Industrial actuators — CNC spindles, press brakes, conveyor drives — are slower to adopt because uptime trumps density, and the thermal duty cycle question is real. But for any startup building a new physical product where motor mass is on the bill of materials, the next 18 months are a decision point: design around current radial flux supply, or bet on axial flux pricing at scale.

Looking ahead

The 2010s were the decade of the lithium-ion cost curve — battery pack prices fell from $1,200/kWh to under $100/kWh and rewrote what was buildable. The 2020s might be the decade the motor follows. Mercedes shipping axial flux at six-figure annual volumes doesn't itself rewrite the industry, but it pulls the topology onto the cost curve. Once that curve starts, every product category with a spinning shaft gets a redesign window — and the winners will be the teams that started planning the redesign before the supply chain caught up.