The company that turns a flat MRI into a rotatable, 3D map of every muscle you have - and then counts them.
Above: Springbok's calling card - the same scan your knee surgeon orders, re-imagined as a color-coded census of the human muscular system. The machine didn't change. The reading did.
Somewhere right now, a strength coach for a pro basketball team is rotating a 3D model of a player's left hamstring on a screen, comparing it muscle-by-muscle to the right. Six months ago that comparison was a hunch. Today it is a number.
That number comes from Springbok Analytics, a 38-person company in Charlottesville, Virginia that has quietly become the muscle-reading layer for elite sport and serious medicine. Its software takes a standard MRI - the kind hospitals have run for forty years - and converts the flat, gray, 2D slices into an interactive 3D model that quantifies up to roughly 140 individual muscles. Volume. Fat infiltration. Left-right asymmetry. Scar tissue. The kind of detail a radiologist could only describe in adjectives, Springbok hands over as data.
More than 80 professional sports teams and human-performance programs now run on it. So do pharmaceutical researchers measuring whether a drug actually rebuilds muscle. The NBA didn't just sign up as a customer - it invested. And in early 2026, GE HealthCare, which makes the MRI machines themselves, signed a deal to build Springbok's analysis directly into the scanner.
Muscle matters. We quantify it.
Here is the strange gap Springbok set out to close. We can measure bone density to the milligram. We can quantify blood across dozens of markers. We can map the brain. But muscle - the largest organ system in the body, the thing that fails first in aging, the thing that tears in athletes and wastes away in cancer patients - has historically been assessed by squeezing it, eyeballing a scan, or asking the patient to push against a hand.
The clinical tools that existed were blunt. A DEXA scan or a bioimpedance reading gives you a single lump number for "lean mass," as if all muscle were one undifferentiated thing. It isn't. A sprinter with a perfectly healthy quad can have a quietly atrophying glute that no scale will ever catch - until it pulls a hamstring on national television.
The data was already sitting inside every MRI scan. Nobody had taught a computer to read it muscle by muscle. The information existed; the translation did not.
The most overlooked organ in medicine is the one you use to stand up.
In 2009, a group of University of Virginia professors weren't thinking about basketball at all. Biomedical engineer Silvia Blemker, imaging specialist Craig Meyer, and kinesiologist Joe Hart submitted a proposal to build an MRI tool that could help surgeons treat children with cerebral palsy. To know what was wrong with a muscle, they first had to define what a healthy one looked like - precisely, in three dimensions.
That problem turned out to be the whole company. Blemker's lab had spent years building computational models of muscle. The bet was that artificial intelligence could automate what had been painstaking manual work - segmenting each muscle out of an MRI - and do it accurately enough to trust clinically.
It took more than fifteen years of research before the science was ready to leave the lab. Springbok was incorporated, licensed its core technology from UVA, and brought in Scott Magargee as CEO to carry it from journal papers to the field. The founders were later named UVA's 2024 Innovators of the Year - a polite, academic way of saying the bet paid off.
CEO & Co-Founder. The operator who took 15 years of muscle science out of the lab and into pro locker rooms and clinical trials.
Co-Founder & Chief Scientific Officer. The UVA biomedical engineer whose lab modeled muscle long before AI could read it.
Co-Founder. UVA imaging expert who helped translate raw MRI signal into something a machine could segment.
Co-Founder. Kinesiologist (now at UNC) who grounded the science in real questions of injury and rehabilitation.
The workflow is almost rude in its simplicity. A person lies in an ordinary MRI for somewhere between ten and forty-five minutes - lower body, core, or full body. The scan goes to Springbok. The AI segments it, muscle by muscle, and returns an interactive web report: a 3D model you can rotate, with hard numbers attached to each structure.
Not just "you have muscle." Rather: this specific muscle is this many cubic centimeters, this much of it has been replaced by fat, it is this much smaller than its mirror on the other side, and here is how all of that has changed since the last scan. Springbok calls the running record a "digital muscle twin" - a personalized model you can track across seasons, treatments, or decades.
The first technology to accurately quantify over 100 individual muscles - now closer to 140.
It is easy to build a demo that impresses. It is harder to get the National Basketball Association to write a check, the Utah Jazz and DC United to run their athletes through it, and GE HealthCare - a company that could in theory build this itself - to instead partner with you. Springbok has all three.
The customer list reads like two different worlds stitched together: elite sport on one side, serious medicine on the other. Pharmaceutical sponsors use Springbok's metrics as endpoints in clinical trials, where "did the muscle actually grow back" needs to be a number a regulator will accept, not a vibe. Researchers studying muscular dystrophy and age-related muscle loss use it to see change too subtle for older tools.
And the funding tells its own story - oversubscribed at every round, which is the polite market way of saying more people wanted in than there was room for.
Total includes earlier disclosed capital beyond the two priced rounds shown. Both priced rounds were oversubscribed - investors include the NBA, the Chicago Blackhawks, TitletownTech and Transition Equity Partners. Bars scaled within the priced-round figures; the antelope keeps jumping.
The machine that makes the scan now wants the software that reads it.
Springbok's stated goal is unglamorous and enormous at the same time: make precise, individual-muscle analysis a routine part of research, clinical care, sport, and healthy aging. The same engine that tells a basketball team which glute is lagging can tell an oncologist whether a cancer patient is wasting away, or tell a 70-year-old whether their strength program is actually working where it counts.
That breadth is the point. Muscle loss is one of the most reliable predictors of how aging goes - whether you keep climbing stairs or stop. If Springbok can turn that into something measured early and tracked over years, the use case stops being "elite athletes" and becomes "most people, eventually."
What if you tracked your muscles the way you track your heart rate - for decades, not just after an injury?
The near future for Springbok runs through the scanner. The GE HealthCare partnership aims to fold muscle analysis into the MRI workflow itself, which is how a niche capability becomes a default. The NBA-backed research with UW-Madison is trying to turn muscle data into fewer injuries, on a measurable timeline through 2026. Each of these moves the company from "interesting analysis you can order" toward "the standard way muscle is read."
So return to that strength coach, rotating the 3D hamstring on the screen. A few years ago the question "is this muscle in trouble?" was answered with experience, palpation, and a guess. Now it's answered with a number that didn't exist in clinical practice before - pulled from the same machine, read by software that learned to see what humans couldn't quantify.
The MRI didn't change. The muscle didn't change. What changed is that someone finally decided to count it. Springbok Analytics is that someone.