MOVRS

Here is a problem that sounds simple and is not: how, exactly, does a human body move? Not roughly - a golfer swings, a pitcher throws - but precisely, in numbers, frame by frame, joint by joint, and how does that motion change over time as the athlete tires or improves? For decades the answer required a small ceremony. You put the person in a special suit. You dotted them with reflective markers. You surrounded them with an array of expensive cameras in a controlled room. And then you asked them to move naturally, which is a strange thing to ask of someone wearing a dotted bodysuit in a room full of cameras.

"The inability to understand quantitatively how a human body was moving, AND changing over time."

MOVRS, founded in 2020 and run out of San Francisco and Los Angeles, looked at that ceremony and asked a deceptively cheap question: what if you skipped all of it? What if the measurement device was just the camera you already had - the broadcast feed, the phone, the single lens pointed at the field - and the hard work happened in software? The company describes what it builds as "vision models for understanding how people move and interact in the real world." Translated out of startup, that means: video goes in, and structured 3D data about human movement comes out.

This is the sort of thing that is either trivial or extraordinarily hard depending on how seriously you take the word "precise." A model that guesses roughly where an elbow is has existed for a while. A model that produces biomechanical measurements a coach or a broadcaster would actually trust - and does it in real time, from ordinary footage, without markers - is a genuinely difficult research problem. MOVRS holds a patent on a real-time AI system for 3D human movement, and its research staff has published on neural networks, NeRFs, and 3D scene reconstruction. The difficulty is the point. If it were easy, it would already be a feature inside somebody's larger product rather than a company.

Start where movement is watched most closely. Then aim at everywhere movement matters.

They started with sports, which is sensible, because sports is the one arena where human movement is watched obsessively, replayed endlessly, and monetized directly. But the founders noticed the thing that turns a sports-tech startup into something larger: teaching a machine to see how a body moves is not really a sports problem. It is a movement problem. And movement shows up in workplace safety, in physical therapy, in insurance claims, in defense, and - increasingly interesting - in training the robots that will eventually need to understand what humans do. MOVRS treats sports as the front door, not the whole house.

The commercial evidence that this is more than a slide deck arrived in early 2024, when NBC Sports announced a pilot with MOVRS to develop next-generation content and visualizations. As part of that work MOVRS captured data and generated content across golf, soccer, tennis, track and field, football, and gymnastics. The tell is what was absent: no sensors on the athletes. When the measurement disappears into the footage, the output stops looking like a lab experiment and starts looking like something a broadcaster can put on air between plays.

None of this makes MOVRS a large company. It is small - roughly fifteen people - and its disclosed outside funding is a modest $50,000 round in March 2022 tied to the Comcast NBCUniversal SportsTech accelerator, the program that plausibly opened the NBC Sports door in the first place. That is worth stating plainly, because the anti-slop version of this profile does not pretend a fifteen-person startup with a pilot deal has already won. What it has is a specific, hard, well-scoped idea and early proof that a major broadcaster found it useful enough to test.

The team reads like a deliberate three-legged stool. Dorian Pieracci, co-founder and CEO, is a former high-level athlete with a master's in Sports Industry Management from Georgetown and a background in commercializing 3D human data - the person who understands both the locker room and the balance sheet. Abdullah Chand, co-founder on engineering, brings more than a decade in scalable vision systems and built the company's patented real-time pipeline. Syed Safwan Khalid, lead AI research engineer, handles the frontier stuff - 3D, NeRFs, generative scene reconstruction. Athlete, builder, researcher, all pointed at the same question.

What can you actually do with it? If you run a broadcast, you can turn a swing or a sprint into on-screen data and story. If you coach, MOVRS and its partner NewtForce offer real-time pitching biomechanics - cloud-based, remote, no lab - to colleges, pro teams, and training facilities, with AI coaching agents in development. And if you build AI models, MOVRS will generate the labeled 3D human-movement datasets that such models are starving for. Three customers, one underlying capability.

The honest risk is the one every "understanding X is a platform" company faces: platforms are harder to sell than features, and a good-enough markerless competitor could commoditize the base layer. But the underlying wager is clean and, in its way, obvious once stated. Human movement is everywhere and structured almost nowhere. MOVRS wants to be the company that writes it down.