Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
The current wave of AI-based generative animation often feels like a magic trick that only works once. You write a prompt, a video comes up, and if you don’t like the result — maybe all the feet are wonky, a regular problem with AI generations — your only real option is to try a different prompt. This “black box” approach is exactly what Cartwheel, a new 3D animation company, is trying to dismantle.
Andrew Carr and Jonathan Jarvis, veterans of OpenAI and Google, respectively, founded the company, which is building a future where AI handles the technical drudgery of animation while leaving the creative spirit to the artist.
I spoke with Carr and Jarvis about launching their company, defining “taste” with AI, and the technical and creative difficulties of animation in 2026.
What makes Cartwheel special?
According to the founders, one of the biggest hurdles in this field is that 3D motion data is remarkably scarce compared to the endless oceans of text and images available online on which AI models are trained.
“If you look at all the big tech companies, they’ve built their models on written language, audio, images and video, because there are so many of them, so those patterns are much easier to find,” Jarvis said. “We knew it was going to be difficult, but it turned out to be harder than we thought by a factor of 10 or 100 to get that data.”
Read more: Generative AI in games exists, but faces opposition from players and developers
While other tech giants focus on generating the final pixels, Cartwheel has spent years mapping how humans actually move. Their models are designed to understand the nuances of performance so that a simple 2D video of someone dancing in their backyard can be translated into an accurate and realistic 3D skeleton.
This shift from flat images to 3D assets is what gives animators the control they lacked in the age of AI.
Cartwheel has spent years tackling the difficult task of mapping how humans actually move.
Preventing “similarity” of artificial intelligence
Cartwheel executives said they view AI “sameness” as a byproduct of a lack of control. If everyone uses the same generator to produce a video, the results may end up starting to look very similar.
“Our system’s output is designed for people to edit. It’s designed for people to be able to touch and manipulate, and we don’t want someone writing something and then turning it into a final animation. That’s not the point. That’s boring. Who’s going to watch it?” Carr said.
“The fact that it’s so easy for people to get into it and edit it completely eliminates the similarity problem,” he said. “You put it on different characters, you put it in different environments, you change its appearance, you push the performance, you pull the performance, and in that sense (similarity) turns into a problem rather than a problem.”
The solution, Carr and Jarvis said, is to provide a “control layer” where the AI output is just the starting point. By generating 3D data instead of flat video, a creator can change lighting, move the camera, or adjust a character’s pose after the AI has done its initial work — making the technology an advanced power tool rather than a replacement for the artist.
Founder Andrew Carr said one of his basic scientific hypotheses is that motion and motion are a fundamental type of data.
The future of animation with artificial intelligence
In addition to making animation faster and lowering the barrier to entry, the company is looking toward a concept they call “open storytelling” or “open world building.” In modern gaming and social media, the demand for content has reached a level that hand animation cannot match.
Cartwheel envisions characters that are not just programmed through a handful of movements but are powered by motion models that allow them to interact and perform in real time. It’s less about designing each frame individually and more about “rehearsing” with a digital actor who understands the intent of the scene.
Ultimately, the goal is to bridge the gap between 2D vision and 3D execution, the founders said.
“One of the basic hypotheses that we hope will be true in the next three years for Cartwheel is that everyone will be working in 3D even if it was written in 2D, even if the final output is just 2D video,” Carr said.
By focusing on “the layer underneath the pixels,” Carr and Jarvis said they hope that as animation becomes more robotic, it also becomes more personal. The machine handles the biomechanics and file export, but the human retains the final say on taste, timing, and the substance of the story.