COSIMO.AI launches “Geometric Video” for AI, answering an industry problem: Legacy Video was never made for AI; it was made for human eyes.
The Physical AI industry has been reduced to an anecdote. A robotaxi launch date slips. A humanoid demo gets delayed. The headlines predictably lament: “AI is not ready.”
The industry’s answer is always the same. Bigger models. More data. More GPUs. More data centers. More power. Hundreds of billions of dollars in, launch dates and demos keep slipping.
The working assumption is that AI’s efficiency is fixed and that only brute-force scaling will enable AI to accurately perceive video, the primary visual input of the physical world.
COSIMO’s Physics Engine answers this problem. It encodes the geometry of shapes and motion directly into the video stream. The output is Geometric Video, which captures the naked geometry of the world, strips out the noise, and encodes objects and their motion in a deterministic, mathematically pure form.
COSIMO tested its Geometric Video on UCF-101. Five seeds. Forty epochs. NVIDIA L4 hardware. Against the Legacy Video baseline, Geometric Video delivers:
- +12.4 percentage points more accurate than legacy video.
- 78.5% fewer model parameters.
- 27× less GPU memory at inference.
- 1.17 milliseconds per frame on a five-year-old MacBook Pro. Under one watt.
- 3× tighter accuracy clustering across all five training runs. Stable enough to debug like source code.
Every performance result traces to a public test run that has been cryptographically verified. The validation pipeline is available at cosimo.ai/validation.
The economic implications are direct. A Tier-1 Physical AI footprint saves $8 to $10 billion a year across compute, storage, bandwidth, and power, and saves $2,700 per edge device. The biggest savings are in time, where COSIMO’s tighter accuracy clustering accelerates time-to-market schedules by an estimated 6 to 12 months. For a robotaxi or a humanoid program, that is a categorical advantage.
The open-source savings calculator is at cosimo.ai/savings.
The whitepaper is at: cosimo.ai/whitepaper.
About COSIMO. COSIMO’s Physics Engine encodes the geometry of shapes and motion directly into the video stream, generating “Geometric Video,” a new kind of video primitive, purpose-built for AI. www.cosimo.ai
View source version on businesswire.com: https://www.businesswire.com/news/home/20260513526696/en/
Media gallery
