Flux raises $30M Series B led by APC to build AI hardware engineer for PCB design
Mar 2, 2026 with Matthias Wagner
Key Points
- Flux closes $30M Series B led by APC to build AI tools that generate PCB designs from text prompts, targeting hardware engineers locked out by inaccessible design tools.
- 75% of Flux customers had never designed a PCB before, mostly mechanical and firmware engineers who can now consolidate multiple off-the-shelf boards into single custom designs at material cost.
- Flux's AI models can optimize designs around specific factories' live inventory and production capacity, potentially pushing unit economics toward near-zero incremental cost as manufacturing integration deepens.
Summary
Flux has closed a $30 million Series B led by APC. Co-founder Matias Wagner describes the company as building the first AI hardware engineer, starting with PCB design.
The pitch rests on a gap in the tooling landscape. Software development has gotten dramatically easier over the past two to three decades. Hardware design has not. PCB manufacturing is now globally accessible—a fully assembled board can be ordered from China for around $20 and delivered in seven days—but the design tools remain inaccessible to most technical people.
Flux's product lets users describe what they want and generates a PCB design, similar to how Cursor or Claude generate code from a prompt. Wagner is direct about where the product sits today. Small to mid-complexity boards can be produced in a single shot. Anything more complex requires an iterative workflow. The long-term vision is prompt-to-iPhone-class device, but he frames that as a distant target.
Customer mix
Most customers are SMBs, and only about 25% had previously designed a PCB themselves. The majority are mechanical engineers, firmware engineers, and industrial designers who had the technical background but not the time or tooling. A vending machine manufacturer previously integrated four or five off-the-shelf boards per machine. With Flux, they can consolidate to a single custom board that is cheaper to assemble, easier to replace in the field, and priced at material cost rather than OEM margins.
Manufacturing integration
Wagner sees a compounding advantage as AI design gets closer to manufacturing. Pick-and-place assembly machines require that every component on a board already sits on their rails. Optimizing a design around a specific factory's live inventory is difficult for human engineers. Flux's models can, in principle, design toward what a given factory can actually produce in a given hour, potentially pushing unit economics toward near-zero incremental cost.
Wagner looked at opening a fab in Fremont but ran into the core constraint: PCB production is a chemical process, and permitting in the US is slow and difficult. He sees opportunity in US-based manufacturing but treats it as a hard problem rather than a near-term unlock.