Cartwheel targets animation bridge

Cartwheel is building animation tools that turn flat video and text prompts into editable three-dimensional motion instead of one-shot generated clips. (cnet.com) CNET reported on April 11 that the startup was founded by Andrew Carr, a former OpenAI scientist, and Jonathan Jarvis, a former Google Creative Lab designer. Carr and Jarvis said the company is focused on “2D vision and 3D execution,” with motion from a backyard dance video translated into a three-dimensional skeleton animators can adjust. (cnet.com) That pitch departs from many generative video systems that output finished pixels and leave users with little control beyond trying another prompt. Cartwheel said its system is meant to produce motion data artists can edit, reuse and export into existing production software. (cnet.com, getcartwheel.com) The technical problem is basic and expensive: three-dimensional animation usually needs keyframing, motion-capture shoots or large motion libraries, and those steps can take hours or days. TechCrunch reported in June 2024 that Cartwheel was trying to generate a first draft of movement from text so animators could start with a usable motion instead of a blank screen. (techcrunch.com) Cartwheel has argued that motion data is much scarcer than the text, image and video data used to train many other artificial intelligence systems. Jarvis told CNET that building those datasets proved “harder than we thought by probably a factor of 10 or 100.” (cnet.com) The company emerged from stealth in May 2025 with a browser-based platform and a $10 million funding round led by Craft Ventures, bringing total funding to $15.6 million, Forbes reported. Forbes also reported that the product supports exports to Maya and Unreal and was designed for games, films, advertising and social content. (forbes.com) Animation World Network reported later that month that Cartwheel opened its tool suite to the public after a closed beta and said the waitlist had topped 60,000 creatives. The company’s site now pitches Cartwheel Canvas and an application programming interface for motion generation, blending, editing, retargeting and character creation. (awn.com, getcartwheel.com) Cartwheel has staffed the effort with veterans from both machine learning and studio animation. Forbes and Animation World Network reported that Catherine Hicks, whose credits include *Coco* and *Inside Out*, joined as head of animation innovation, and Neil Helm, who worked on crowd animation for *Inside Out 2* and *Turning Red*, became head of interactive animation. (forbes.com, awn.com) Jarvis’s own site says Cartwheel’s models are being aimed not only at games, movies and advertising but also at robotics, which uses the same underlying problem of describing movement in space and time. That leaves Cartwheel trying to sell a narrower idea than text-to-video: not finished scenes, but controllable motion that other tools can build on. (jonathanjarvis.com, cnet.com)