Aurorin CAD Wants to Do for Mechanical Engineering What Cursor Did for Code

For a generation of mechanical engineers, the everyday experience of CAD has been defined by a handful of incumbents – Solidworks, NX, Creo, Catia, Fusion 360, Onshape – and a strangely consistent set of complaints. The software is slow. Files take minutes (sometimes hours) to open. Design workflows are tedious, multi-step, and hostile to iteration. A new Y Combinator company called Aurorin CAD is wagering that the only real fix is to throw out the foundations and start over.

Aurorin CAD, a Winter 2026 Y Combinator batch startup, describes itself as “Claude code for mechanical engineers.” More technically, the company is building AI CAD software from the ground up on a modern technology stack, with AI as a first-class primitive rather than a bolted-on feature. The premise: almost every existing CAD program is built on top of one of a few underlying CAD kernels designed in the 1980s. Those kernels are why even basic operations on real assemblies can feel painful, and why opening a single file in a major package can take hours. Aurorin is replacing that ancient substrate with a custom parametric and B-Rep kernel built for the speed and structure that modern AI tooling demands.

The result, the company claims, is a CAD system that is dramatically faster on its own terms – and that can fully leverage AI in ways that legacy products cannot. Because Aurorin owns the kernel, it can expose model state to AI agents in ways that retrofitted plugins simply cannot match. That opens the door to a new class of mechanical engineering workflows: agents that propose design changes, optimize geometries against real constraints, and handle the drudgery of feature trees and constraint management without the user ever leaving the tool.

Aurorin CAD was founded in 2026 by Michael Baron, a notable solo founder with deep aerospace and systems engineering credentials. Baron is a three-time SpaceX intern and a former Apple intern, and his work has spanned some of the most demanding software environments in the industry. At SpaceX, he worked on Raptor combustion simulation, Dragon guidance, navigation, and control, and Starshield flight software. At Apple, he focused on GPU driver performance – exactly the kind of low-level engineering chops that a CAD-kernel-from-scratch project demands.

The market opportunity is substantial. Mechanical CAD is a multi-billion-dollar software category dominated by a handful of incumbents whose products are widely beloved for capability but widely loathed for performance and ergonomics. The arrival of AI-native tools in adjacent verticals – most visibly in software engineering, where a wave of AI-augmented IDEs has rapidly reshaped how code is written – has only sharpened the appetite for similar leaps in mechanical design. Aurorin’s bet is that AI CAD software for mechanical engineers – built kernel-up rather than retrofitted – will unlock the same generational shift already underway in software development.

For Aurorin CAD, the long-term vision is to become the default tool for the next generation of mechanical engineers – the people designing rockets, robotics, electric vehicles, medical devices, and the physical infrastructure of modern industry. If the founders are right that the kernel is where the real moat lives, Aurorin has a rare opportunity to displace a category of software that has not seen serious foundational reinvention in decades.