Gen AI that rides shotgun. The Dallas company making driver-safety intelligence cheap enough to run on the camera you already own.
A garbage truck, a school bus, a delivery van - the vehicles no one photographs. EdgeTensor points cutting-edge computer vision at exactly those, and calls the boring stuff where the real safety money lives.
EdgeTensor builds Gen AI-powered video telematics and fleet-management software. In plain terms: it takes the raw footage streaming off dashcams and in-cabin cameras, and turns it into structured answers - who is driving safely, who is distracted, which alerts matter and which are noise. The company runs those models at the edge, on the low-cost cameras and embedded chips inside the vehicle, and in a cloud platform that stitches everything together.
Founded in 2017 by computer-vision researchers Rajesh Narasimha and Soumitry Ray, EdgeTensor started from a stubborn engineering problem: modern AI is expensive to run, and fleets operate on thin margins across thousands of vehicles. The company's answer was to make the models small, fast and hardware-agnostic - AI affordable enough to scale to the edge.
The result is a platform aimed less at any single gadget and more at the messy reality of the telematics market, where a fleet might run three camera brands, two dashboards and a decade of mismatched hardware.
Fleets bought the cameras. Then they drowned in the footage. A typical video-telematics rollout generates a flood of alerts - hard braking, phone use, drowsiness - most of which are false or trivial. Managers stop watching. The safety value evaporates.
EdgeTensor's pitch is to fix the signal-to-noise ratio: process video at the edge, discard the junk, and surface only events that actually change risk. Fewer, better alerts - plus automated compliance and coaching so the work isn't manual.
Publicly referenced users include HDfleet and Fleet Operate.
* Figures are company-reported and approximate; bar widths are illustrative, not to a fixed scale.
AI fleet-management platform delivering real-time telematics, analytics and operational insight.
Safety and compliance - driver behavior, distraction detection, coaching and automated violation reporting.
One-stop layer that ingests, normalizes and filters data from any existing dashcam or telematics device.
Generative-AI analytics for video intelligence across fleet operations.
FNOL, fraud detection and behavior-based underwriting built from real driver-video data.
Driver and occupant monitoring plus ADAS runtime models for OEMs and Tier-1 suppliers.
Child-safety monitoring program using AI video for school-bus fleets.
AI-powered video analytics tuned for waste-collection fleet operations.
APIs, Docker, white-labeled portals and SSO so partners embed the AI in their own products.
Most of the well-known names in this market - Samsara, Motive, Netradyne, Nauto, Lytx, Geotab - sell integrated hardware and software as a bundle. Buy their camera, use their AI. It works, but it locks the customer in.
EdgeTensor took the opposite stance. It doesn't sell you a camera. It layers its AI onto the telematics devices a fleet already owns, across brands and generations of hardware. Refusing to sell the box is not a limitation here - it is the moat. The company positions itself as the AI brain that any device, dashboard or integrator can plug into.
That places EdgeTensor in the software-and-licensing layer of a fast-consolidating video-telematics market: less a rival to any single dashcam maker, more the connective tissue between them.
CEO & CO-FOUNDER
PhD in Electrical & Computer Engineering from Georgia Tech, with 16+ years in computer vision and machine learning. He worked at Texas Instruments and Samsung Research America - building deep-learning algorithms for phones and wearables - and helped start the U.S. R&D operations of metaio, an AR company later acquired by Apple.
CO-FOUNDER
Co-founder of EdgeTensor, part of the founding research team that set the company's edge-AI direction. EdgeTensor runs on two continents, with teams in Dallas, Texas and Bengaluru, India - a small, research-driven deep-tech shop rather than a sprawling org.
B2B SaaS and licensing. EdgeTensor sells its camera-agnostic Cloud AI platform and edge-deployable models to fleets, telematics providers, integrators, automotive OEMs/Tier-1s and insurers. Partners embed the technology through APIs, Docker, white-labeled portals and SSO. Third-party estimates put annual revenue around $1M (unverified).
In March 2021, 3Lines Venture Capital led a bridge-to-Series-A investment. Backers across the company's history include Plug and Play Tech Center, the U.S. National Science Foundation, Aristos Ventures, Capital Factory, SRI Capital and Plug and Play Japan.
Rajesh Narasimha and Soumitry Ray launch EdgeTensor in Dallas to make edge AI affordable for automotive, fleets and insurance.
3Lines Venture Capital leads a round with participation from Plug and Play, the NSF and others.
The offering consolidates around FleetGo, FleetShield and a Gen AI-powered Cloud AI platform.
Product demos and channel from the company.
AI video telematics: balancing safety and privacy (short).