Satellite and smart-meter estimates of hidden rooftop solar, feeder by feeder
TL;DR Hiddensolar tells an electricity distribution company how much rooftop solar is generating on each feeder right now and over the next few hours. It reads cloud movement from Indian weather satellites and checks it against smart-meter data. DISCOMs pay a yearly subscription to plan power purchases with fewer surprises.
India had 33.41 GW of grid-connected rooftop solar by 30 September 2026[1]. Almost all of it sits behind a customer's meter, so the DISCOM only sees what a home or factory draws from the grid, not what its panels make.
On a clear day that hidden generation pulls daytime demand down. When a cloud bank crosses a city it vanishes within minutes and demand jumps right back.
Load dispatchers plan purchases a day ahead and adjust in 15-minute blocks. If they can't see a city's solar then they buy too much on sunny days and scramble on cloudy ones. As time-of-day tariffs and smart meters spread[2], getting the solar hours right matters more and more to DISCOM finances.
The seed proposed collecting solar power in orbit and beaming it down. NASA's 2024 study found such a system would cost more than renewable options on the ground, even if it ran from 2050[3]. So I kept the space angle and the clean-energy goal, and pointed both at a problem India has today.
What it does, in regulatory terms: forecasting and analytics software for utilities. It does not dispatch power, trade electricity or control any customer's system.
Hiddensolar pulls cloud imagery from INSAT-series satellites through ISRO's MOSDAC archive[4], turns it into estimated sunlight for each patch of a city, and multiplies that by the rooftop capacity registered on each feeder. Smart-meter readings from solar homes correct the estimate every day. The DISCOM gets a dashboard and a data feed - hidden solar per feeder and substation, now and for the next six hours, with the forecast error shown alongside.
The user is a DISCOM's load dispatch or power purchase team, and the payer is the DISCOM, through a tender or a pilot work order. I found no public price for this kind of service. Utility-scale plants already buy commercial forecasts from global forecasting firms[5], so the budget line exists. My assumption is a pilot priced per circle or city, then a yearly subscription per DISCOM.
Rooftop capacity has passed 33 GW[1]. Time-of-day tariffs reached commercial and industrial users from April 2024 and were due for most others by April 2025, with solar-hour prices 10–20% below normal[2]. The same rules say smart meters must be read remotely at least daily[2], which gives you the correction data this product needs.
Nobody publishes what Indian DISCOMs spend on load or solar forecasting. The narrowest number is the 33.41 GW of rooftop solar they can't see directly[1].
The count of buyers matters more than size, though. India has a few dozen DISCOMs, so a handful of contracts makes or breaks this.
Global forecasting firms sell irradiance and power forecasts worldwide at up to 90 m resolution, including India coverage[5]. MOSDAC publishes a three-day, 15-minute solar and wind forecast[4]. Both forecast sunlight or plant output.
I found no Indian product that turns sunlight into an estimate of hidden rooftop generation per feeder, tied to a DISCOM's own customer records.
| Difficulty | Moderate (45/100) |
| Starting capital | ₹60 L–1.2 Cr |
| Time to launch | 6–9 months |
A small team that knows satellite data and power systems can crack the modelling. The buyer is the hard part, because DISCOMs move slowly, procurement runs through tenders and you need their feeder maps and rooftop registers before your numbers mean anything at all.
No sector licence, because Hiddensolar forecasts and doesn't trade or dispatch power. The checks are about data.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't let the forecast drive grid switching on its own. Operators must make dispatch calls, and the forecast error should always sit next to the number.
Want to work out the AI layer for Hiddensolar? Talk to Yash →
Here's what can go south -
I like this for founders who've worked inside a DISCOM, an SLDC or a forecasting firm and can get a meeting with a chief engineer. If you're a pure data-science team with no utility contacts then it's a tough road.
Prove first that one DISCOM will share a year of feeder and smart-meter data for a free back-test.
Plan for roughly ₹60 L–1.2 Cr over 6–9 months, mostly on data science, utility pilots and long sales cycles.
No sector licence, since you don't trade power. You need MOSDAC registration, a DPDP processing agreement with each DISCOM and the usual company and GST registration.
Yes, wherever utilities can't see behind-the-meter solar. Swap in local weather-satellite data and check your data protection law.
The free Hiddensolar validation kit has the line-by-line capital breakdown, the effort scoring, the data-access and procurement checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for DISCOM teams, and 5 assumptions to test. Download the kit →
Thinking about building Hiddensolar or something close to it? The make-or-break work is getting one DISCOM to share its feeder data and sign a pilot, and that's a conversation I can help you prepare for. 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.