Skip to content
YC Root AccessYC Root Access

How Diode Is 10x-ing Hardware Design

Davide Asnaghi and Lenny Khazan started Diode Computers with a question: why does hardware design still move so slowly? Drawing on their backgrounds in software and electrical engineering, they set out to reframe circuit board design as a software problem. One that AI could help solve. They went from building a tool to catch design mistakes to creating an end-to-end AI-powered system that generates production-ready boards. Along the way, they discovered that most companies didn’t want a better tool — they wanted the whole solution. That insight led to rapid growth, real customers, and over 100 boards designed in just a few months, all by a two-person team. Today, they announced $11.4M in Series A funding led by a16z. This is the story of two engineers trying to make hardware move at software speed, and the infrastructure making it possible. Learn more about Diode Computers at https://diode.computer. Apply to Y Combinator: https://ycombinator.com/apply Chapters: 00:22 - What is Diode? 00:31 - Customer Base and Early Growth 00:51 - The Origin Story 02:46 - Initial Challenges and Pivot 04:01 - Finding the Right Problem 05:02 - First Successful Deal 05:29 - Realization and Validation 06:40 - Reframing PCB Design as a Software Problem 11:28 - Technical Choices and Challenges 11:56 - Innovative Language Design 12:31 - Infrastructure and Security 13:33 - Future Prospects 14:21 - Recruitment and Team Building

Diana HuhostDavide AsnaghiguestLenny Khazanguest
Jul 22, 202515mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

Diode uses AI to automate PCB design as software-like workflow

  1. Diode builds an AI-enabled PCB design shop that generates complete board designs rather than just assisting engineers with individual steps.
  2. After initial YC customer discovery showed limited demand for “mistake detection,” the team pivoted to the higher-pain problem: producing the PCB design itself.
  3. The founders argue LLMs already “know” electrical engineering rules, but need a code-like intermediate representation instead of legacy visual tooling to express that knowledge.
  4. Diode combines automation with ownership of the deliverable by running an internal verification/acceptance pipeline so customers receive designs that work.
  5. To sell into regulated and conservative industries, Diode designed secure, deployable infrastructure including an offline core compiler (Rust) with browser-based visualization via Wasm.

IDEAS WORTH REMEMBERING

5 ideas

Customers wanted outcomes, not tooling add-ons.

Early prospects dismissed “mistake finding,” but responded strongly when Diode offered to deliver a finished custom board—because the real bottleneck is scarce PCB-design talent and limited internal bandwidth.

Customer discovery can invalidate impressive technical work—and that’s useful.

They built sophisticated data generation and training ideas for verification, but 100+ conversations revealed the problem statement was wrong, forcing a pivot to what buyers would immediately pay for.

A code-first intermediate representation unlocks LLM capability in hardware design.

LLMs are strong at producing and editing code, while PCB tooling is mostly visual; translating schematics into a language that’s readable for humans and “natural” for LLMs reduces errors and enables iteration.

“Services as a product” works when you own verification and acceptance criteria.

Instead of shipping an untrusted AI tool, Diode takes responsibility for correctness by running internal verification and aligning deliverables to customer acceptance checks, making AI-driven speed commercially viable.

Security and deployment constraints are key to landing conservative industries.

Aerospace/medical customers may require air-gapped or tightly controlled environments, so Diode built a compartmentalized system with an isolated offline core and local compilation/visualization.

WORDS WORTH SAVING

5 quotes

“Diode is using AI to automate circuit board design.”

— Davide Asnaghi

“People basically told us, ‘Uh, we don’t make mistakes.’”

— Davide Asnaghi

“They didn’t want a component of the solution. They wanted the solution.”

— Davide Asnaghi

“If you could reframe the PCB design problem as a software problem, you can take all of the great learnings… in the world of software.”

— Lenny Khazan

“Traditional tooling is very visual and graphical. Hasn’t really changed much since maybe the 80s or so.”

— Lenny Khazan

AI automation for PCB designPivot from verification to full board generationCustomer discovery and first deal validationPCB design as a software/compilers problemLanguage/intermediate representation for schematicsRust compiler, Wasm, offline/air-gapped deploymentSelling to Fortune 100, aerospace, medical; security constraintsServices-as-a-product with “software margins”Recruiting: researchy yet product-oriented engineers

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome