
Common Sense Machines (CSM)
Common Sense Machines is building artificial intelligence that learns to translate the world into a 3D simulation.
Location
Cambridge, MA
Size
17 People
Total raised
$10M
Hiring?
Yes
Platform
Cloud / web-based
GPU
None — cloud processing
Pricing
Free tier available
License
Proprietary
Best for
Generative 3D from images & text
Updated
July 2026
Common Sense Machines is training artificial intelligence that learns to translate the visual world into a 3D simulation for people and machines. 3D simulations have great value; we build 3D models to create photorealistic images of products and spaces at a fraction of the cost of photography, develop diverse assets for game and augmented realities, use CAD to prototype and print new parts, or create synthetic worlds to test and train autonomous systems. But 3D models are still too slow, hard, and expensive to create.
Common Sense Machines is building a Real-to-Sim platform for creating build human-like 3D models with machine-like speed and scale. Common Sense Machines was founded by a group of PhDs and faculty from MIT. We have worked on large AI systems at DeepMind and have collectively founded three companies with successful exits. The company is based in Cambridge, MA. We are venture-backed and looking to expand our team of scientists and engineers.
Frequently Asked Questions
What is Common Sense Machines?
CSM is a "Real-to-Sim" AI company building tools that translate images, sketches, and text into 3D models and simulation-ready assets. It was founded by a team of PhDs and faculty from MIT and is based in Cambridge, MA.
What is CSM used for?
Creating photorealistic product and space visualizations, game and AR assets, CAD-ready prototypes, and synthetic worlds for testing and training autonomous systems.
Is CSM a capture tool?
No — CSM is generative. Rather than reconstructing scenes from multi-view captures, it generates 3D from images, sketches, or text. CSM was named to CB Insights' AI 100 list in 2024 alongside Luma AI.



