A new player in the AI animation space, Cartwheel, is working to transform how stories are created by bridging the gap between 2D ideas and 3D execution. Founded by former OpenAI scientist Andrew Carr and ex-Google creative director Jonathan Jarvis, the startup is positioning itself as a solution to one of generative AI’s biggest creative limitations—lack of control.

Unlike many current AI animation tools that rely on prompt-based outputs, Cartwheel focuses on creating editable 3D motion data. This allows creators to refine animations after generation, offering flexibility in adjusting camera angles, lighting, and character movement—something traditional AI video tools often lack.

The company highlights a key industry challenge: the scarcity of high-quality 3D motion data compared to abundant text and image datasets. To address this, Cartwheel has spent years developing models that understand human motion, enabling the conversion of simple 2D videos into accurate 3D skeletal animations.

This approach also tackles the growing concern of “AI sameness,” where content generated from similar tools begins to look identical. By giving creators a “control layer,” Cartwheel ensures that AI-generated outputs serve as a starting point rather than a finished product, preserving originality and creative intent.

Looking ahead, the startup is focused on advancing what it calls “open-ended storytelling.” This concept envisions dynamic digital characters capable of reacting and performing in real time, reducing the need for frame-by-frame animation while expanding creative possibilities.

As demand for scalable content grows across gaming and digital platforms, Cartwheel’s model could signal a shift in how animation is produced—where AI handles technical complexity, and creators retain artistic direction.