Should you build or buy Computer-Aided Design (CAD)?
Computer-Aided Design (CAD) software creates 2D drawings and 3D models for mechanical, product, and industrial design — geometry, assemblies, tolerances, drawings, and file exchange with manufacturing partners. Buyers are engineering teams designing physical products who need a proven kernel and a supplier ecosystem that already speaks the same file formats.
Buy the CAD platform. Nobody outside Dassault-scale history has built a production general CAD system for internal use, Boeing tried and stopped, and the moat is not just the kernel (Parasolid licenses to 200+ vendors) but drawings, exchange, supplier familiarity and validated workflows. Spend the engineering budget on CadQuery or build123d automation, OCCT-based internal tools and PLM integration where your workflow is distinctive, and keep the seats.
Build it, buy it, or bridge?
When building makes sense
General-purpose CAD is one of the clearest buy cases in the index, and the historical build attempts prove why. Dassault Aviation's internal CATI system eventually became CATIA and spun out as a separate company rather than staying an internal tool; Ford's PDGS ran for decades and was eventually replaced by commercial software; Boeing tried building internal CAD and abandoned it after it failed to scale to real functionality needs. What is genuinely buildable is narrow, differentiated automation on top of a licensed platform: CadQuery and build123d let engineering-led teams script parametric geometry as code, with a real deployment at FX-Bricks. Open Cascade Technology (OCCT) is embedded at BMW and other manufacturers as underlying infrastructure, not a replacement CAD system, and FreeCAD has partial integration reports elsewhere. None of this adds up to a company running its own general CAD platform instead of AutoCAD, SolidWorks, or Fusion 360 — it adds up to companies extending a licensed platform with code.
When buying makes sense
Buying the CAD platform is the clear call, and confidence is high. The moat isn't just the geometry kernel: Siemens' Parasolid alone licenses into a large roster of vendor products. The rest of the moat is drawings, file exchange, supplier familiarity, and validated workflows built up over decades. Industry documentation on neutral file exchange describes real information loss, dropped history, and lost constraints when models move between systems, the exact interoperability tax a from-scratch system would have to solve from zero. Seat pricing runs from budget tiers into several thousand dollars annually depending on platform and configuration, real cost, but nowhere near what building and maintaining a kernel-grade platform would take. Spend your engineering budget on CadQuery or build123d automation, OCCT-based internal tooling, and PLM integration where your workflow is genuinely distinctive, and keep the seats for the geometry engine itself.
The desk read
Build-versus-buy analysis for Computer-Aided Design (CAD) is being written. In the meantime, the framework that drives every B4 call is on the B4 Index page.
Frequently asked
Has any company successfully built its own CAD system?
Not as an internal tool that stayed internal. Dassault Aviation's CATI became the commercial product CATIA and spun out as a separate company; Ford's PDGS eventually gave way to commercial software; Boeing tried and abandoned an internal CAD system after it couldn't scale to real functionality needs.
Can you script CAD instead of using a GUI?
Yes, for parametric automation on top of a licensed platform. CadQuery and build123d let engineering teams define geometry as code, with a real deployment at FX-Bricks. This complements a commercial CAD seat; it doesn't replace the geometry kernel or drawing tools underneath it.
What are the main CAD vendors?
Representative vendors include AutoCAD, Autodesk Fusion 360. B4 Pro includes the category score and the full vendor list.