Paid dashcams on cabs and trucks that catch Indian road edge cases
TL;DR Dashcamfleet puts dashcams in cabs, trucks and delivery vans and pays drivers in rupees for the kilometres they share. It then sells short, blurred clips of tricky Indian road moments to teams building driver-assistance systems, and buyers pay per scenario set, not per terabyte.
Driver-assistance systems trained on Western roads have a tough time in Indian traffic. Think autos cutting across, cattle, unmarked speed breakers, bikes going the wrong way and handcarts at night. Public Indian data is thin. The India Driving Dataset from IIIT Hyderabad, which people still point to, has 10,000 images from 182 drives in two cities[1].
Indian teams are building for these roads now. One Indian startup, for example, pitches autonomy for emerging markets like India with features for local conditions[2]. Each team shoots its own footage with a few test cars, which covers very little ground and almost none of the rare events that matter most.
What it does, in regulatory terms: video data collection and licensing. Footage shows people and number plates, and that counts as personal data under the DPDP Act 2023[3]. Dashcamfleet blurs faces and plates before any clip leaves its servers and gives drivers clear notice and consent terms.
Drivers on cab, truck and delivery fleets get a dashcam plus a monthly rupee payout tied to the kilometres they upload and any clip a buyer later licenses. On the server, models flag the odd moments - hard braking, near misses, animals, vehicles going the wrong way. Reviewers then tag them by scenario. Buyers search a catalogue and license sets like "two-wheeler cut-ins at night, Pune" or "cattle on highway, Rajasthan".
The original version of this idea paid drivers in tokens on a decentralised network. I dropped that. In India, gains on virtual digital assets are taxed at 30% with 1% TDS and no loss set-off[4], so tokens are a poor way to pay a driver. Rupees over UPI are simpler and easier to explain.
Drivers supply the footage and get paid for it. The buyers are ADAS and autonomy teams at carmakers, tier-1 suppliers and startups, plus simulation and research labs. Today they pay for their own test fleets and annotation vendors. What does an hour of Indian driving footage sell for? Nobody says publicly. One crowd-data network reports selling 3,000 hours to one lab but keeps the price to itself[5].
Paying the crowd for driving data is already a business abroad. One global network reports 2.89 million registered drivers in 33 countries and over 165,000 camera hours a month[5]. Closer home, an Indian-founded fleet-camera company became a unicorn in January 2025[6], which tells you fleets will live with cameras. And Survey of India's 2021 guidelines opened up geospatial data collection for Indian firms[7].
There's no public figure for an Indian driving-data market, and I found none for what's spent on ADAS training data in India either. The narrowest signal is the gap itself - a common public Indian dataset holds 10,000 images[1]. On the adjacent, global side, that network's scale[5] shows there are buyers for crowd footage, though not how much they pay.
Two crypto-token networks pay drivers in tokens for camera data, and both lean towards maps and Western markets[5][8]. Fleet-camera firms run cameras on fleets for driver safety, and that footage stays with their fleet customers[6]. IIIT Hyderabad's dataset is academic and small[1]. I found no Indian firm paying drivers in rupees for edge-case clips.
| Difficulty | Hard (56/100) |
| Starting capital | ₹1.85–3.75 Cr |
| Time to launch | 6–9 months |
Fitting cameras is the easy bit, and fleets already accept them. The hard part is digging rare events cheaply out of thousands of hours of dull footage, and then you also have to blur well enough to meet data-protection duties and keep drivers uploading when each month's payout is small.
No specific licence for selling road footage was found. Privacy and geospatial rules are what matter here.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't rely on automatic blurring without a human spot-check on every clip you sell. One missed face or plate is a privacy breach, and it's on you.
Want to work out the AI layer for Dashcamfleet? Talk to Yash →
This is a good fit for a founder with computer-vision and fleet-partnership experience who can sell to engineering teams at carmakers. If you're hoping for a consumer app then look elsewhere. Before anything else, prove two things - that two ADAS teams will pay for a scenario set, and that 70% of drivers keep uploading after two months.
Plan for roughly ₹1.85–3.75 Cr, covering dashcams, driver payouts, storage, the review team and the mining software.
No specific licence was found. You need DPDP compliance with blurring and notice, a check of the geospatial guidelines for foreign buyers, and proper TDS on driver payouts.
Yes, carmakers want local edge cases in every market. Privacy law on filming public roads varies a lot, so read yours before you install a single camera.
The free Dashcamfleet validation kit has the line-by-line capital breakdown, the effort scoring, the data-protection and payout checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for ADAS teams and drivers, and 5 assumptions to test. Download the kit →
Thinking about building Dashcamfleet or something close to it? The make-or-break work is landing the first paying ADAS team and mining rare events cheaply, and that's the kind of thing I help founders work through. If you need a hand, you know where to find me. Talk to Yash →
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.