Quilter, a startup building AI to automate the painstaking process of printed circuit board (PCB) design, has raised a $25 million Series B led by Index Ventures. The funding, announced by CEO Sergiy Nesterenko, brings Index’s Nina Achadjian to the board and will fuel Quilter’s mission to solve a problem that has plagued hardware engineers for over 60 years.
Every modern electronic device is built on a PCB, yet the process of laying one out remains a stubbornly manual, weeks-long slog in CAD software. As Nesterenko puts it, “Imagine if software engineers still wrote machine code by hand.” That’s the bottleneck Quilter claims it can eliminate.
From Heuristics to Physics
For decades, the industry’s best attempt at automation came from “auto-routers”—tools that tried to “auto-complete” an engineer’s work. But according to Nesterenko, these tools never truly respected the underlying physics of a circuit, leading to a deep-seated distrust among designers.
Quilter is taking a fundamentally different approach. Instead of learning from human-designed boards, its platform uses reinforcement learning trained on the first principles of physics. The system generates millions of potential layouts, evaluates them against the laws of electromagnetics and thermodynamics, and optimizes for performance, manufacturability, and cost. “No one solves Maxwell’s equations in their head,” Nesterenko noted, highlighting the computational advantage of their approach.
The company is building its entire stack from the ground up, including its own CAD kernels and physics engines, to support this new paradigm. The goal is to create a “compiler for hardware,” transforming a specialized, expert-driven process into something exponentially faster and more accessible.
With customers already in the aerospace, automotive, and consumer electronics sectors, Quilter says its platform can reduce a multi-week design process to just hours. By removing the layout bottleneck, the company hopes to enable hardware teams to innovate at the speed of software, freeing up engineers to focus on system-level problems instead of manually routing thousands of tiny copper traces.
