ChronoSpace AI

ChronoSpace AI is an early stage company working on capturing the 4D physical world. Its public presence is a single page carrying the line Digitizing the 4D Physical World and a partnership contact address, and its own description is that it is building technology to capture 4D spatial reality. There is no product, no documentation, no demo, no team page and no announced funding.

The company is associated with the Brown University group behind PackUV, a CVPR 2026 paper on packed gaussian UV maps for 4D volumetric video. PackUV maps every gaussian attribute into a sequence of structured multi scale UV atlases, so that a gaussian splat sequence is stored as images and a run of those images over time is itself a video. The authors describe the result as the first unified volumetric video representation compatible with standard video codecs without losing quality, which addresses the delivery half of volumetric video rather than the capture half.

That work comes from Aashish Rai, Angela Xing, Anushka Agarwal, Xiaoyan Cong, Zekun Li, Tao Lu, Aayush Prakash and Srinath Sridhar, across Brown University, UMass Amherst and Meta. Sridhar leads Brown's Interactive 3D Vision and Learning Lab. The accompanying PackUV-2B dataset was captured with more than 50 synchronized cameras across 100 sequences and 2 billion frames, and the authors report scaling to sequences of up to 30 minutes with consistent quality.

Note that ChronoSpace AI is distinct from Chronosplat, an unrelated open source browser player for 4D gaussian splat sequences, and from a similarly named virtual reality business in Poland. The names are close and the subject matter overlaps.