A flood-history report for a plot or building before you buy, build or lend
TL;DR Floodfile tells a lender, valuer or land buyer how often one specific plot or building has gone under water over the past few decades. It uses free satellite archives and checks them against local reports. Lenders and developers pay per report or per portfolio scan, and it sells what actually happened, not long-range climate forecasts.
Floods damage about 12 lakh houses in India in an average year, and in 2015 that figure was close to 40 lakh, going by Central Water Commission data from 1953 to 2020[1]. After the Chennai floods that year, a press report put builders' losses near ₹30,000 crore and said buyers were staying away from rain-hit localities[2].
Cities keep building right over the spots where water used to sit. In Nagpur, built-up area grew from 72.85 to 185.4 km² between 2000 and 2023 while water bodies shrank from 12.2 to 2.29 km², and 145 mm of rain on one day in September 2023 flooded the city[3].
So what does a buyer or lender know about that exact street? Usually nothing at all.
Floodfile makes a short report for any address or plot boundary. It counts how many times the site and its access roads showed surface water in satellite images, when that happened and how long the water stayed, and how the land around it has changed since. A model trained on known flood dates flags likely waterlogging the images miss, and a reviewer signs off every report.
The data is free to start with. The EU's Global Surface Water dataset covers about 41 years of Landsat images[4], and Sentinel-1 radar sees through monsoon cloud at down to 5 m[5]. Local news, municipal complaints and residents' photos fill in the street-level detail that 30 m pixels cannot.
I narrowed the seed here, which pitched AI forecasts of climate risk for real estate and infrastructure. Plot-level forecasts for 2050 are hard to defend with Indian data and a lender can't act on a number it can't check. A dated record of past water is evidence anyone can verify.
Home-loan lenders and their empanelled valuers are the first buyer, because a flood record is one more collateral check they can file with the valuation. Developers and land aggregators buying plots come second, since a flooded site costs more to raise and drain. Homebuyers could get a cheaper single report later. No Indian price is published for this, so per-report and per-portfolio prices are assumptions you'll have to test.
Sentinel-1C launched on 5 December 2024, bringing back two-satellite radar coverage[5]. The Global Surface Water record was updated for 2024[4]. And built-up land keeps spreading over lakes and drains, as the Nagpur study showed in 2025[3], so the gap between a plot's paperwork and its flood record only gets wider.
I found no public figure for an Indian property flood-risk report market. The closest anchors are about losses - CWC data puts average annual flood damage at about ₹6,428.67 crore across crops, houses and public utilities[1], and Chennai builders alone faced a reported ₹30,000 crore hit in 2015[2].
Those numbers show how big the problem is, not what anyone spends on reports.
An Australian-born climate-risk platform offers a hub that screens single assets or portfolios and reports to TCFD and ISSB standards[6]. A US climate-analytics firm sells asset-level multi-peril projections to banks, asset managers and REITs[7]. Neither mentions India coverage on its homepage, and both sell forward-looking models to big institutions. I found no Indian service selling verified flood history per plot, though web search was down when I checked.
| Difficulty | Moderate (43/100) |
| Starting capital | ₹60 L–1.2 Cr |
| Time to launch | 6–9 months |
Processing satellite archives is well documented, so that bit is fine. The hard parts are accuracy at street level, where a basement can flood without a single pixel showing water, and selling to lenders whose credit and legal teams move slowly and will ask who is liable if a report misses a flood.
No sector licence applies to flood-history reports, but data and liability rules matter.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't let AI decide alone that a plot is safe. A clean result can still miss a basement flood, so a human reviewer signs every report.
Want to work out the AI layer for Floodfile? Talk to Yash →
Here's what can go south -
I think this is a solid niche for a founder with remote-sensing skills and contacts in housing finance or valuation. If you're hoping to sell glossy climate scores to homebuyers in year one then this isn't it.
Prove two things first - that your reports match known flood events in two cities, street by street, and that one lender will pay for a portfolio scan.
Plan for roughly ₹60 L–1.2 Cr, mostly for remote-sensing talent, report review and a long sales cycle with lenders. The core satellite data is free.
No sector licence. You need to follow the geospatial data guidelines and satellite data terms, comply with the DPDP Act for homebuyer data, and have clear liability terms.
Yes, the satellite archives are global. Check local mapping rules, data protection law and whether public flood maps already do the job.
The free Floodfile validation kit has the line-by-line capital breakdown, the effort scoring, the data, liability and privacy checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for lenders and valuers, and 5 assumptions to test. Download the kit →
Thinking about building Floodfile or something close to it? The make-or-break work is proving street-level accuracy and getting a lender to put your report into its loan process, and that's the kind of thing I like working through with founders. 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.