Diode Computers
- Diode makes software to design printed circuit boards, which are used in most electronics.The startup raised $11.4 million in Series A funding led by A16z.Diode aims to work with robotics, medical device, and aerospace companies.
Green, copper skeletons — sometimes no thicker than a piece of printer paper — are the literal core of today's electronics. The printed circuit boards, or PCBs, that power most devices on our planet, from smartphones to drones, can be slow and costly to produce, according to a startup hoping to automate the design and manufacturing process.
Diode Computers, a startup that uses artificial intelligence to create custom PCB designs in code rather than mapping out designs in pictures, claims it can slash the time and manpower required to create boards for specific products — a timeframe the startup says used to take weeks to months, but may now happen in a matter of days.
"The people that have the ability to generate circuit boards are retiring, and so now you have a very small set of people that can generate correct designs from experience — and a ton of people that would love to learn, but there's not enough manpower to actually teach them," Diode cofounder and CEO Davide Asnaghi told Business Insider.
The startup raised $11.4 million in a Series A funding round led by Andreessen Horowitz, known as a16z, with participation from Caffeinated Capital, Box Group, and startup mentorship program Y Combinator. The round brings Diode's total funding to just over $14 million.
Diode's cofounders, Davide Asnaghi and Lenny Khazan, were in the summer 2024 Y Combinator cohort. There, they came up with the idea to help companies generate circuit board designs from scratch using AI.
Instead of designing PCBs using pictures or diagrams, Diode translates PCB layouts into code using open-sourced AI software the company created. The startup says the software makes it easier to spot mistakes on the circuit boards, quickly reconfigure them for specific products, and scale production. Diode uses OpenAI and Anthropic large language models, but it's also using reinforcement learning to train smaller models to spot circuit board design mistakes, Asnaghi said.
Diode charges its customers for custom designing and manufacturing circuit boards. The startup outsources production to partners in California and on the East Coast, but Asnaghi's long-term goal is to become a "one-stop shop" that will design, manufacture, and deliver PCBs to clients:
"Our eventual vision is that our customers should be able to push code to us in the same way that you deploy on AWS, and we will take that code and generate a physical product that gets shipped back to their assembler or their factory," he said.
The startup is focused on working with robotics, medical device, and aerospace companies because "these three industries are very interested in manufacturing in the US," Asnaghi said. AI robotics startup Physical Intelligence and autonomous maritime startup Saronic are both Diode customers, the company said.
Diode hopes to tap into Silicon Valley's renewed focus on scaling American manufacturing, particularly for aerospace and defense. Venture capitalists and founders building hard tech discussed this trend at the Reindustrialize Summit, a conference hosted in Detroit in July.
"In terms of conflict and warfare, you need to be able to iterate on your hardware and electronic systems as quickly as you're iterating on your design," Erin Price-Wright, a partner with A16z's American Dynamism practice — which invests in aerospace, defense tech, public safety, and other national security startups — told BI.
"Instead of designing one PCB board that's going to be in an airplane for the 50-year lifetime of an airplane, you're thinking about having to redesign a new electronic stack every couple of months of a drone being deployed onto a battlefield."
The PCB design market is pretty crowded. Cadence Design Systems and Altium both make PCB design software. And Quilter, another startup that makes AI-powered PCB design software, raised a $10 million Series A led by venture firm Benchmark in February 2024.
Before launching Diode, Asnaghi designed custom silicon chips at Chromatic, a stealth hard-tech startup. There, he experienced what he now sees as a recurring holdup in electrical engineering: PCB design errors that stalled production.
"It was insane to me that you had this wonderful complex device, which is the custom silicon, and then the simpler node created unforeseen delays for no reason," he said.
Asnaghi also previously worked as an embedded software engineer at Apple and Butterfly Network, a medical imaging company. He earned his master's in engineering at Berkeley and studied engineering in college in Italy and at the Hong Kong University of Science and Technology, where he was an exchange student. His interest in electrical engineering, he told BI, stems from meeting the founder of the open source electronics platform Arduino.
"Electronics is the most magical of the engineering disciplines," he said. "You can take software, which is somewhat ethereal, and make it do things in the real world."
Asnaghi met Diode cofounder and chief technology officer Lenny Khazan at Butterfly Network, where Khazan was an intern. He previously worked as a software engineer at Chromatic and Instabase, an AI data company.
"We are seeing a rise in the importance of physical products," Asnaghi said. "There's more need for infrastructure — data centers, power generation, all of these things rely on prepared circuit boards on some level. Virtually any product that has some electronics in it will have a circuit board."
