Currently, 'text to 3D mesh' models are emerging like Shap-E AI from OpenAI. They could maybe provide material for game engines but lack the features and accuracy to be used in industrial grade mecanical projets.
Will we see a 'text to 3D CAD' or 'image to 3D CAD' system powered by AI able to translate any reasonably complex part description into industrial grade CAD models before 2027 ?
The system would need to feature at least:
Part modeled with correct geometry (size, shapes), and parametric features (holes, fillets, chamfers, bores, etc...) according to the input description.
Usable in at least one CAD software
Part must be ready for CAD assembly
1,000At 86%, the market is pricing in a level of maturity that current text-to-CAD systems haven’t demonstrated. I put the probability at 55%: they can produce simple drafts, but tolerances, DFM, GD&T, and reliable assembly-ready industrial parts remain conspicuously absent.
@ConnorTabarrok Very promising, if this product release and is able to generate a good range of part with various geometries and correct features this could be it !