Aurorin CAD Is What Happens When You Rebuild SolidWorks from Scratch with AI at the Core

The Macro: CAD Software Is 30 Years Old and It Shows

SolidWorks launched in 1995. NX traces its lineage to the 1970s. Creo (formerly Pro/Engineer) started in 1987. The dominant CAD tools used by mechanical engineers today are built on architectures that predate the modern internet. They are powerful, but they are also slow, crash-prone, and hostile to new users.

The performance problems are real. Opening a complex assembly in SolidWorks can take minutes. Large parts with hundreds of features slow to a crawl. Rebuilds after parameter changes take ages. Engineers working on complex designs spend significant time waiting for their CAD software to catch up.

The learning curve is equally painful. Becoming proficient in SolidWorks or NX takes months of dedicated training. The interfaces are dense with toolbars, menus, and feature trees that overwhelm new users. This creates a skills bottleneck: companies cannot hire fast enough because qualified CAD operators are scarce.

AI could transform both problems. Natural language interaction could flatten the learning curve. Modern architecture could eliminate the performance issues. But bolting AI onto a 30-year-old CAD kernel is like adding a turbocharger to a horse-drawn carriage.

Aurorin CAD, backed by Y Combinator, started from scratch. They built a new parametric CAD kernel designed to be fast, stable, and AI-native from day one.

The Micro: A SpaceX Intern Who Built His Own CAD Kernel

Michael Baron founded Aurorin CAD with a background that screams “I know what CAD should be.” Three SpaceX internships (Raptor combustion simulation, Dragon guidance and control, Starshield flight software) and an Apple internship on GPU driver performance. He has used CAD extensively and understands both the engineering and performance sides.

Building a new CAD kernel is one of the hardest software engineering challenges that exists. Geometric modeling, constraint solving, parametric history trees, and surface mathematics are deeply complex. Most CAD companies are built on licensed kernels from Parasolid (Siemens) or ACIS (Spatial). Building your own is a multi-year endeavor that most startups would not attempt.

But owning the kernel gives Aurorin a decisive advantage: they can integrate AI at the deepest level rather than wrapping it around a black-box geometry engine. When a user asks “make this flange 2mm thicker and add a fillet,” the AI can directly modify the parametric model rather than navigating a legacy API.

The performance claims are strong: “Parts that take hours to load elsewhere open instantly.” This is achievable with a modern kernel that takes advantage of contemporary hardware (multi-core CPUs, GPU compute) rather than the single-threaded architectures of legacy systems.

The product is available as a free download for Mac and Windows, which is an aggressive go-to-market for a CAD tool. SolidWorks starts at $3,995. The free tier builds user adoption that can monetize later through team features and advanced capabilities.

Competitors include SolidWorks (Dassault Systemes), Fusion 360 (Autodesk), Onshape (PTC), and newer entrants like Zoo.dev. Aurorin’s advantage is the clean-sheet kernel with native AI integration.

The Verdict

Aurorin CAD is attempting one of the hardest and most valuable things in engineering software: replacing legacy CAD tools with something fundamentally better. The kernel ownership and AI integration create a genuine technical moat.

At 30 days: how many engineers are actively using Aurorin for real design work, not just testing?

At 60 days: can the kernel handle the complexity of real-world assemblies with hundreds of parts?

At 90 days: are engineering teams considering Aurorin as a SolidWorks alternative, or is it still a novelty tool?

I think Aurorin has the most ambitious vision in the CAD space right now. Building a new kernel is incredibly hard, but if they succeed, the result is a CAD tool that is genuinely better in every dimension: faster, easier to learn, and AI-native. The founder’s SpaceX background means he understands what engineers need from a CAD tool. If the kernel delivers on performance and the AI integration makes design faster, Aurorin could reshape mechanical engineering workflows.