Circuit Mind raises $25M to automate circuit board design with AI agents

Automating the hardware lifecycle

Circuit Mind, a British startup, recently secured $25 million in new funding to expand its artificial intelligence platform. The company aims to automate the manual, labor-intensive processes involved in designing printed circuit boards. While software teams have seen rapid acceleration from AI tools, electronics design has remained relatively stagnant for decades.

The startup's platform helps engineers move from a high-level hardware block diagram to a finished schematic and bill of materials in minutes. Instead of manually reading thousands of pages of datasheets, engineers can receive several optimized options from the software. These options are tailored to specific priorities, such as minimizing power consumption or reducing overall costs.

The new funding will support the growth of the company's sales, engineering, and customer success teams. Additionally, Circuit Mind plans to use the capital to establish a headquarters in the United States. This move is intended to provide better support for their growing customer base in North America.

Physics-based AI agents

Unlike general-purpose large language models, Circuit Mind's platform uses a patented, physics-based engine. This engine tests every design decision against specific electrical limits to ensure the hardware functions correctly. The system relies on machine-readable models of electronic components rather than just text-based data.

The company utilizes AI agents that work with these physics models and the specific engineering rules of each customer. These rules can include a company's list of approved manufacturers and specific design constraints. The agents are designed to handle the heavy lifting of verification, which is often where mistakes stay hidden until a physical prototype is built.

The company recently launched the Agentic Verification Engine to specifically target the design review phase. This tool allows engineers to hand off complex tasks to a swarm of AI agents that trace every connection and bus. The agents check for mismatches in pin tables or operating voltages by cross-referencing the customer's own review checklists and component datasheets.

Real-world performance gains

Early adopters of the technology are already reporting significant improvements in speed and accuracy. At Los Alamos National Laboratory, engineers finished a medium-difficulty board in just over four hours using the platform. This was a major improvement over the estimated 60 to 80 hours required to complete the task by hand.

Schneider Electric SE has also seen notable results during early evaluations of the tool. The company reported a cut of more than 90% in the effort spent exploring different electronic architectures. These gains suggest that the platform can handle complex industrial applications while maintaining high standards of precision.

Beyond speed, the platform is helping companies manage their manufacturing costs more effectively. Some customers have reported that the tool helped bring down their bill of materials costs by 25%. By providing optimized options for cost and power, the software helps engineers make better procurement decisions during the initial design phase.

Expanding the design scope

While the platform currently excels at turning diagrams into schematics, Circuit Mind plans to push deeper into the hardware lifecycle. The company intends to expand its capabilities to include hardware testing and firmware development. This expansion aims to provide a more comprehensive suite of tools for hardware engineers from start to finish.

The platform is designed to integrate seamlessly with the software tools that engineering teams already use. Once a design is finished, it can be exported directly into existing electronic computer-aided design software. This ensures that the AI's output is immediately actionable for the human engineers overseeing the project.

The company's leadership views these tools as the hardware equivalent of popular coding assistants like Cursor or Claude Code. By providing a way to