The supply closet that finally learned to speak up.
Pictured: a five-letter acronym - Data Analytic Real-World Visual Intelligence System - doing the job of a clipboard, a flashlight, and a very patient night-shift clerk.
Not a security guard. Not a nurse pulling a double. A camera on a swivel mount, quietly panning across rows of gloves, catheters, and saline bags. It reads a barcode here, notes a half-empty bin there, and updates a dashboard nobody is awake to look at yet. By morning, the order is already placed. DARVIS built that camera's brain.
DARVIS - short for Data Analytic Real-World Visual Intelligence System - is a health-tech company that turns ordinary supply rooms into self-reporting ones. Its software runs on smart PTZ cameras that continuously watch fill-levels, scan barcodes and QR codes without a human lifting a scanner, and translate all of it into live consumption data. The company likes to say it makes walls talk. The walls, for their part, have a lot to say about how much inventory just walked out the door.
Every health system runs on supplies, and almost none of them know, at any given minute, what they actually have. Inventory gets counted by hand, on clipboards, by people who would rather be doing nearly anything else. Items expire unseen at the back of a shelf. Critical stock runs out at the worst possible time. The result is waste measured in billions and a daily ritual known, without affection, as the escalation meeting.
The tooling that existed mostly asked humans to scan more, tag more, and count more. RFID tags on every item. Weekly audits. Spreadsheets that were accurate for roughly the length of time it took to print them. DARVIS looked at this and asked a quieter question: what if nobody had to count at all? What if a camera could simply see the shelf the way a person would, only without ever getting tired, distracted, or assigned to a different floor?
Manual counts are slow, late, and wrong by the time anyone reads them. DARVIS replaces the count with a continuous look.
Supplies quietly age out at the back of the shelf. The camera flags expiration before the bin becomes a write-off.
Stockouts surface during rounds, not before. Auto-replenishment moves the decision to the night before.
DARVIS did not begin in a hospital. It began in 2015 in California's gaming world as Hashplay, building immersive VR and real-world analytics - the kind of company whose old Twitter handle, @hashplaybeta, still lingers as a fossil from a previous life. Founders Jan-Philipp Mohr, Ingo Nadler, and Jan Schlueter had a stack of computer-vision technology and a hunch that its most valuable use was not entertainment.
So around 2019 they made the unglamorous bet: point the vision system at the least exciting room in the building and solve a problem nobody brags about at dinner parties. They reframed the cameras as inventory eyes, created a digital twin of the monitored area - a live model capturing detail a human eye cannot continuously track - and rebuilt the company as a B2B SaaS business aimed squarely at healthcare logistics. Trading social gaming for sterile supply rooms is not the pivot most VR founders dream about. It is, however, the one with recurring revenue.
The flagship is called DARVIS Digital Shelves, and the pitch is mercifully simple. Mount the smart PTZ cameras. Let them watch. The system reads fill-levels around the clock, scans barcodes and QR codes with zero human touch, and feeds an AI dashboard that turns raw observation into demand forecasts. When stock dips, it can trigger an automatic reorder straight through the hospital's ERP. The shelf, in effect, manages itself.
Underneath the convenience sits the harder engineering: object detection trained on the chaos of real supply rooms, expiration tracking, vendor-managed inventory, and integrations sturdy enough to survive a hospital IT department. It is also built to the standards that make healthcare buyers comfortable - GDPR, CCPA, SOC 2, and HIPAA compliance are part of the package, not an afterthought. A camera that watches medical supplies has to be trusted before it is admired.
Barcodes and QR codes read automatically - no handheld scanner, no missed scan.
Consumption data feeds forecasts so reorders happen before the shortage, not after.
Plugs into existing systems to trigger replenishment without a purchase-order chase.
A live model of the room captures what a roaming human eye never could continuously.
DARVIS's client and partner list reads like a tour of places where mistakes are expensive: MD Anderson Cancer Center, ProMedica, UHealth at the University of Miami. Beyond healthcare, names like Accenture, Airbus, HERMA, and Trident Industries show the vision platform travels. The 2021 partnership with AI firm Olive folded DARVIS capabilities - hygiene checks, bed logistics, medical inventory, sterile equipment completeness - into a platform reaching more than 900 hospitals across 41 states.
The money has followed the proof, if quietly. Investors include Anthem Venture Partners, Plug and Play, WS Investment Co., and the NVIDIA Inception Program, with later participation from Multiversum and VCO Ventures. Total funding sits at roughly $12M - modest by AI-hype standards, which is rather the charm of a company selling something hospitals can actually measure.
Approximate disclosed amounts. The Series A figure was not publicly broken out - shown as the balance toward the ~$12M total.
Chart: not a hockey stick. A company that grew at the speed hospitals actually buy - which is to say, carefully.
That is the whole thesis, and DARVIS keeps it refreshingly narrow. Not a robot in every hallway. Not an AI that replaces the staff. Just a set of eyes on the supply room, doing the one tedious job that humans have always done badly because it was beneath their attention and above their patience.
The leadership signal matters here. Hiring Michael Prokopis - who managed $3.2B+ in spend and drove $360M+ in savings at MD Anderson - as CEO is a bet that the next phase is less about inventing the technology and more about getting hospitals to adopt it. The inventor's instinct gives way to the operator's.
Most AI in medicine chases diagnosis and drug discovery - the headline acts. DARVIS picked the back office, where the work is invisible until it fails. As staffing pressures grow and margins tighten, the rooms that count themselves stop being a luxury. Every box that does not expire, every reorder that fires before a shortage, every hour a nurse does not spend with a clipboard adds up in a place where those things are scarce.
So picture that supply room at 3 a.m. again. The camera still pans across the shelves. Only now the half-empty bin has already been flagged, the order is already in, and the expired item never made it to the back of the shelf because something saw it coming. Nobody is awake to notice. That is the point. The best thing a watchful machine can do is make the next morning boring.