The AI-designed motor startup attacking China's drone supply chain

Three engineers kept hitting the same wall: every programme they led at Shield AI and Mach Industries ran late because of one unglamorous part. The motor. The actuator. The thing that actually makes an uncrewed platform move.

"No matter where we were, we were always hamstrung by our ability to procure critical components," Christian Mochen says. He, Tom Baron and Carlo Dela Rosa left to build the component supplier they could not find. On August 6, their company, Atlas Motion, came out of stealth with $11.5M in Series A funding led by Greycroft.

The component nobody thinks about until it is missing

Every uncrewed system runs on motors and actuators. Drones need them to spin propellers and move control surfaces. Ground robots need them for steering, suspension and manipulator arms. Maritime platforms use them for rudders and propulsion. A loitering munition steers in flight with a compact actuator. An autonomy stack is dead weight if the platform underneath cannot move.

Most of that hardware is made in China. The Pentagon has been moving to close that dependency, and from 2027 Chinese-linked components become unavailable to it. The three founders spent years inside companies that built platforms on top of this supply chain, and watched procurement delay their own programmes.

So they built the supplier.

Vector: design as a software problem

Atlas calls its design platform Vector. Mochen describes it in terms any operator understands: the software takes a motor specification, spins up a simulation environment, runs the checks an engineer would run by hand, and generates the manufacturing packages the factory needs to build it.

The headline figure is the speed. A design that normally takes weeks to months, with an electromagnetism specialist, a mechanical engineer and supply-chain people all in the loop, comes back in five to ten minutes.

Mochen is careful about the boundaries. "AI isn't replacing engineering," he said. "It's simply orchestrating a lot of critical decisions across the engineering toolpath very, very fast."

His accuracy claim is bold: 99% confidence between what the software outputs and what the hardware on the factory floor actually measures. That is a company figure, not an audited one. No customer has been named, and Forbes reported that Atlas's buyers sit in defence. Treat the design-speed and volume numbers as claims to track, not as settled fact.

The supply chain angle is the real story

The design speed matters less than what it enables. Switching a production line to a variant is usually expensive, which is why component suppliers push fixed catalogs. The company says standard raw-material inputs plus a software-driven line cut make that switching cost close to nothing. A buyer in India, Poland or Africa can bring a requirement to Atlas and get a motor designed to their specification faster than a Chinese supplier can quote a catalog part.

The geography is deliberate. Prototyping and engineering sit in Long Beach, California. Mass production runs out of a facility in the Philippines, a US treaty ally with agreements stretching back more than half a century. Mochen is blunt about why: standing up a cost-competitive motor plant in the US is inherently difficult, while the Philippines has the manufacturing depth and the talent pool.

As we wrote in August, Hadrian raised $1.37B in a Series D to scale automated factories for US defence production. Atlas Motion is a different bet: not a factory that makes weapons or munitions, but the critical-component layer underneath every autonomous platform, designed and produced with the same AI-driven logic Hadrian applies to machining.

The founders still need to be realistic about the path. The company currently uses Chinese magnets for the lowest tier of requirement, sourcing the rest from the US or Japan. Importing finished goods leaves it exposed to tariffs. And a company that incorporated in February, shipped its first motor in 60 days and now claims 10,000 units a month is moving fast by startup standards, but it is a rounding error against the Chinese supply base it wants to displace.

The numbers Atlas is putting up

The company says it has 14 customers, is in six-figure revenue and will hit seven figures next month. Production capacity of 10,000 motors a month is already committed; the target is 40,000 by December. A new design reaches scaled production in six weeks against an industry norm of up to six months.

None of this is independently verified. The company has not named a single customer, and its accuracy and volume figures are self-reported. What is verifiable is the shape of the market it is chasing: a Western drone and robotics industry that needs a non-Chinese motor supply base before the 2027 procurement window closes.

Baron's framing of the mission is worth sitting with: "It's about recognizing why the United States lost its manufacturing base, regaining that technical skill set, and then creating the conditions for us to be successful anywhere in the world."

What happens to the motor supply chain a year from now?

The investment angle of a component maker

The thesis here is not Atlas Motion specifically, and that distinction matters. An $11.5M Series A in a component maker is a small check in a sector where the platforms it supplies raise hundreds of millions. The interesting claim is structural: the component layer of the autonomous platform stack is becoming an investable category on its own, driven by procurement policy rather than consumer demand.

That shift favours a specific kind of company. The winners will be the ones that can demonstrate not just a working prototype but a production line that delivers volume, because the customer here is the Pentagon's industrial base, and it buys in quantity. Software-driven design is the lever a newcomer can pull to attack a mature Chinese-dominated category without a decade of process engineering.

Development scenarios

Sources