Machina Labs CEO Edward Mehr on AI-powered sheet metal forming, defense contracts, and the Golden Dome opportunity

Summary

Machina Labs is building what CEO Edward Mehr calls a "robotic craftsman" — an AI-powered system that can form, trim, and inspect complex metal parts without purpose-built tooling. The company's first application is sheet metal forming, a $280 billion industry where conventional production requires expensive dies, four-story stamping presses, and development timelines of seven to nine years. Machina's two-robot system uses large mechanical fingers to push and pull flat metal sheets into complex shapes — no dies required — compressing the time from design to first part down to hours.

The die elimination is the core commercial argument. A single die for a car door can cost up to $1 million. By removing that capital requirement entirely, Machina can offer faster turnaround at comparable cost and still capture higher margins. Mehr is candid that those margins will compress over time as the process matures, which is why the company is framing the robot as a platform — adding bending, hemming, and forging operations over time to stay ahead of margin erosion.

The AI bet

The harder problem, Mehr argues, isn't kinematics — industrial robots have long had the physical capability to match human range of motion, with more precision and force. The missing piece was intelligence: the ability to reason through complex, multi-step manufacturing sequences. LLMs and transformer architectures have changed that, but only where training data exists. For sheet metal forming, Machina could generate its own data at scale from day one, which is precisely why the company started there. Shoe manufacturing, by contrast, lacks the structured data needed to train a reliable model — and that constraint applies broadly across manufacturing.

Mehr is bullish on humanoid robots, but skeptical they'll find their first commercial home in factories. Industrial robots already solve the physical problem. Humanoids make more sense, he argues, in homes and unstructured environments where the form factor actually matters.

Nvidia, tariffs, and margins

Nvidia is an investor in Machina Labs. A recent collaboration with Nvidia and an OpenAI artist-in-residence demonstrated a fully autonomous pipeline: a sculptor spoke his intent, and the system generated code, ran the robots, and produced a physical metal sculpture without any manual intervention.

On tariffs, Mehr says roughly 70% of Machina's bill of materials is off-the-shelf hardware, sourced primarily from the US or allied nations — robots come from Japan. That limits tariff exposure and, combined with software-like margins, gives the company room to absorb cost increases that would squeeze a conventional manufacturer.

Defense and Golden Dome

Machina already manufactures missile body components for defense primes, which puts it directly in the supply chain for hypersonic interceptors — a key enabler of any national missile defense architecture. The technical edge here is material processing: Machina can form titanium sheets at room temperature without tearing, something conventional methods can't reliably do. High-temperature alloys like nickel and titanium are essential for hypersonic vehicles.

On the Golden Dome missile defense program, Mehr says the plan is for it to reach roughly 7% of the DoD budget once fully programmatic, with approximately $20 billion allocated to missile defense in the near term. No contractors have publicly commented on program structure yet, but the addressable opportunity for a supplier that can process exotic materials faster and without dedicated tooling is significant.