Decades ago, the West outsourced a critical layer of the electronics stack to Asia: printed circuit boards (PCBs). Davide Nastagi, co-founder and CEO of Brooklyn-based startup Diode, is on a mission to bring this capability back to America. By leveraging artificial intelligence to co-design and manufacture circuit boards, Diode aims to become the “AWS for electronics,” bringing the rapid, iterative speed of software development to the physical hardware world.
The Problem: The Neglected Middle Child of Electronics
While silicon design is highly optimized due to the millions of dollars at stake, and software features rapid, low-cost iteration, PCBs have long been treated as a “middle child.” Prototyping cycles in the US are slow, and designs often suffer from low yield rates. Historically, the West abdicated this stack layer to Asia to cut costs, leaving a massive gap in domestic high-speed prototyping and manufacturing infrastructure.
Why Shenzhen Succeeded (It’s Not Just Cheap Labor)
A common misconception is that China dominates hardware manufacturing solely because of cheap labor. In reality, hubs like Shenzhen have built highly integrated, deeply automated ecosystems that divide and conquer complex manufacturing processes. To compete, US manufacturers must implement de novo automation. Diode’s core thesis is that automation succeeds when you co-design the circuit boards with the constraints of the assembly line in mind, utilizing AI to handle the variance that traditionally hurts manufacturing yields.
Automating the Stack with AI
Diode uses Large Language Models (LLMs) to automate the design pipeline. Through a partnership with Anthropic, Diode has trained models to become better electrical engineers. Today, Diode has automated roughly 90% of the schematic generation process. Physical layout and routing—traditionally highly labor-intensive—are currently 40% to 50% automated and rapidly improving. This automation allows enterprise clients, such as aerospace and robotics startups, to compress their product cycle timelines from nine months to just one month.
The Open-Source Philosophy and “AWS for Hardware”
Rather than gatekeeping their software tools, Diode operates with an open-source ethos. They actively contribute to KiCad, a free, open-source schematic capture and PCB layout tool, and have open-sourced their own compiler. Their ultimate goal is democratization: building a reliable, highly automated domestic manufacturing stack so that hardware startups can easily dispatch designs to physical production, just as software developers spin up servers on AWS.
Mentoring question
If hardware prototyping and physical production could move at the rapid speed of software, how would that change your team’s approach to product experimentation and risk-taking?
Source: https://youtube.com/watch?v=B8DbMJ1Zr4I&is=kFsO4EzlRU__NohQ