I want to talk about a tool that rarely gets mentioned in electronics design textbooks — or rather, one that is often discussed with a hint of condescension in certain circles. KiCad. The open-source EDA software that everyone knows, that some people use on weekends for personal projects, and that we, at Codium, use every single day to design boards that end up in industrial products. Real ones. The kind that run in factories.

Let me explain why we find the scepticism amusing.

KiCad's legacy at Codium

My former boss chose this tool originally when he started his company. It was a natural choice when we started ours.

I'll be honest: in the early versions, you really had to want to use it. The user experience was... let's say, particular. Non-intuitive keyboard shortcuts, a somewhat dated interface, no meaningful 3D visualisation. We worked with it. We had our habits, and it got the job done.

Since then, KiCad has undergone a fairly spectacular transformation. Version 5 brought serious 3D. Version 6 rebuilt the interface practically from scratch. Recent versions feature an interactive push-and-shove router, differential routing with length matching, and proper library management. Today it is a powerhouse — not a weekend maker's toy.

To be clear: we do not choose to keep KiCad out of nostalgia or open-source dogmatism. We keep it because it does the job, because it is part of our industrial flow, and because we have invested in it. The rest of this article explains those three points.

"Oh, you use KiCad..."

I know that tone well. You encounter it at conferences, with clients, talking to other engineers. There is a tacit hierarchy in the EDA world: at the top of the pantheon, Altium Designer, Cadence Allegro, OrCAD. Below that, everything else. And right at the bottom, alongside hobbyists and university projects — KiCad.

I'll be straight: I don't really know Altium, OrCAD or Cadence. I don't use them, so I'm not going to pretend to compare them technically. That would be dishonest. What I do know, however, is their commercial model. And on that, I have things to say.

The per-seat licence problem

These tools run on rigid licensing models, often per seat, with significant entry costs. An Altium licence costs thousands of euros per year per user. Cadence is a whole other planet of pricing complexity. OrCAD, the same.

The result: companies that buy them make a heavy investment and amortise it over years — often 10 to 15 years. They don't update. They use the same version until Windows no longer supports it, or until the maintenance contract expires. Then they get hit by 15 years of evolution all at once. Training, new interfaces, new file formats, new workflows. It's a leap from the past that nobody enjoys making.

We receive the source files, open the archive — "Oh, these are OrCAD files from 2012!?" — "Don't worry, we have the import module in KiCad :)"

A situation we've experienced more than once

And the import module works. The schematics open, the netlists are recovered, we get back to work. No drama. Meanwhile, the owners of the original files are struggling to find a compatible version of their proprietary software.

The software doesn't route the board for you

Another thing we see regularly: boards designed with high-end tools... that don't look like much. The EDA tool, however expensive, does not compensate for poor design decisions. That's not Altium's fault, nor KiCad's — it's just reality.

We get a project to take over. We open the file. We look at the routing. "What a mess... it was done in Altium... the software doesn't route the board for you."

Workshop feedback, multiple times

What matters is what the engineer does with the tool. And KiCad, in the hands of someone who knows how to use it, produces exactly the same board quality as any other software.

What KiCad genuinely brings

All the features you need

KiCad's functional scope in 2025–2026 covers everything required to design professional electronic boards:

  • Full schematic editor, hierarchical, with BOM and netlist management.
  • PCB editor with push-and-shove router, differential routing, length matching, customisable DRC rules.
  • Integrated 3D visualisation that is honestly impressive — useful for mechanical checks before ordering PCBs.
  • Structured library management with a massive official community library.
  • Manufacturing file exports: Gerber, Excellon, Pick & Place, BOM — perfectly compatible with assembly house requirements. We have tested it thoroughly ourselves and with others. Zero negative feedback related to file formats.

Open source — and free

Open source does not automatically mean free. KiCad is both. This point is far more strategic than it might seem. An intern who joins — they download KiCad, install it, and they're operational on our projects. No licence request, no dedicated machine, no administrative delay. A new team member — same thing. A client who wants to access their board's source files to archive or bring in-house — same thing. They download, they open, done.

That kind of friction seems trivial until the day you need it not to exist. It is a quiet but real comfort.

Open source = moddable: our digitised industrial flow

This is probably the most important reason why we will not be changing tools any time soon. KiCad exposes a Python API and a clean plugin architecture. We have built on top of it.

Today, our production flow is entirely digitised from KiCad. And I mean entirely — not "we export a CSV that goes into Excel and gets glued together somewhere else by hand". Zero Excel files in the loop.

  • KiCad is connected to our ERP.
  • The BOM generated from KiCad feeds directly into purchase orders and inventory.
  • Pick & Place files are sent to the placement machine automatically, in the right format, with the correct naming.
  • The complete manufacturing package is generated in one click — Gerber, drilling, assembly BOM, P&P file.
  • Production tracking, inspection, test bench, shipping — all linked.

We built this flow because we had a tool we could open and modify. Some of the connector modules we developed are available to the community. Because that's how open source works, and we think it's the right approach.

And today with AI: need a small feature in the flow? A custom check script, an export in a specific format for a particular assembly house? We add it. In a few hours. That was already true before — it's even more true now. The barrier to extending the tool is practically zero.

A bit of history — because it helps understand where we are

KiCad did not come from nowhere. And its evolution into a credible industrial tool was not a coincidence — it had a precise catalyst, called CERN.

KiCad history

  • 1992 Created by Jean-Pierre Charras, then a teacher at the IUT of Grenoble. A personal project that became a complete EDA tool distributed freely.
  • 2011 Decision at CERN to actively join KiCad development. The BE-CO-HT section (Beam Controls, Hardware & Timing) needed a robust open-source EDA tool for its own designs. KiCad was chosen.
  • 2013 Major technical contributions from CERN: deep refactoring of the internal architecture, introduction of the interactive push-and-shove router, differential routing and length matching. These are top-tier professional features — precisely the ones that were most lacking.
  • 2015 — KiCad 4.0.0 First major release integrating the CERN contributions. KiCad enters a new category. The first serious companies start to look at it differently.
  • 2018 — KiCad 5.0.0 Improved 3D visualisation, better library management, general stabilisation. The tool starts to look like something genuinely solid.
  • 2021 — KiCad 6.0 Complete UI overhaul. New rendering engine. Significantly improved design rule checking (DRC). The version that convinced many still-sceptical professionals.
  • 2025 — 2026 De facto open-source standard in the electronics industry. Teaching tool in engineering schools. A credible professional alternative — and increasingly, the alternative.

What CERN contributed is concrete and measurable: the push-and-shove router, differential routing with length matching, a deep refactoring of the internal architecture, and ongoing financial support. Without that, KiCad would probably have remained a nice tool for hobbyists. With it, it became something else.

"An EDA tool designed and used for CERN's internal needs — it is not targeting weekend projects."

— Thomas, Codium

So no, it is not for tinkering

Between my previous company and Codium, hundreds of professional electronic boards have been designed under KiCad. Boards that went to assembly houses in France and across Europe. Hundreds of thousands of boards produced — by us, by others — without the file format ever causing a single problem.

Industrial control boards. Communication boards. Embedded boards. Boards with RF constraints, thermal constraints, isolation constraints. Not blinking LEDs on a breadboard.

KiCad is the tool on which Codium was built. Not just for the design office — for the entire industrial production infrastructure around it. The ERP, the machines, the test benches: everything runs around the data that comes out of KiCad. It is the backbone of the flow.

And it does not depend on a large company that could decide tomorrow to change its licensing model, get acquired, or drop support for a version. It depends on a global community, on contributors with real needs — including CERN, which is a difficult argument to dismiss.

So when people imply that KiCad is for hobbyists — we laugh. And in the meantime, we design boards, and we manufacture them. With KiCad.

The project is there if you want to take a look: kicad.org.

Your next project designed with KiCad?

We design your board from A to Z — schematic, routing, manufacturing. In-house, with our own tools.

Contact us Design Office