Hourly delivery forecasts and banking plans for factories buying green power
TL;DR Offtakely is software plus a small analyst desk for factories and offices that buy wind and solar power through open access. It forecasts how much green power will arrive each hour, matches that to the plant's load and tariff slots, and checks the DISCOM's banking credits every month. The energy or finance head pays a yearly fee.
Since June 2022, a consumer with just 100 kW of demand can buy green power through open access, down from 1 MW, and the DISCOM must bank any surplus[1]. That pulled in buyers who have no power-trading staff. Their supply swings with the wind and the clouds, while their bills now run on time-of-day slots - solar hours 10–20% cheaper, peak hours 10–20% dearer for commercial and industrial users from April 2024[2]. A plant manager sees one monthly bill and can't tell whether a bad month came from weather, poor banking use or a DISCOM accounting slip. Generators face tighter deviation bands from April 2026[3], and some of that cost flows into buyer contracts.
The seed pitched federated AI forecasting across many renewable plants. Generator-side forecasting is already a crowded, specialist service, and rival generators rarely share data. So I moved the product to the other end of the wire - the buyer.
Offtakely takes three inputs - the contracted wind and solar plants' locations and capacity, the buyer's meter data, and the state's tariff and banking rules. It produces a day-ahead and week-ahead forecast of green power arriving in each time slot, against expected load. The plant team gets plain advice on which shifts to move into solar hours, when banked energy is worth drawing, and when to buy grid power instead.
Once a month it reconciles the DISCOM bill against its own forecast and the generator's meter readings, flagging missing banking credits or wrong slot allocation. Buyers with several sites in different states see one view. The analyst desk handles disputes and contract renewals, where the forecast history becomes bargaining evidence.
The user is the plant's electrical or utilities head. The payer is usually the CFO or sustainability head, who signed the open-access deal to cut costs and meet green targets. My assumption is ₹3–8 lakh a year per site, priced below what one billing error typically costs. I couldn't find a public price list for comparable buyer-side services. Developers who sell open-access power are a second channel, since they could bundle Offtakely to keep customers happy.
The 100 kW threshold and 15-day deemed approval date from June 2022[1]. Time-of-day tariffs became mandatory for commercial and industrial users in April 2024[2], which made the hour of delivery matter as much as the volume. CERC's tighter solar and wind deviation bands started in April 2026, though court petitions are pending[3].
I found no public figure for the number of open-access green power buyers or what they spend. Adjacent - India had 299,246.53 MW of renewable capacity on the latest MNRE table, including 171 GW of solar and 59 GW of wind[4]. Only a slice of that serves open-access buyers. Count the buyers in two target states from state portal data before raising money.
A Bengaluru energy-analytics firm sells AI and sensing tools for renewable asset and grid management, mostly to generators and utilities[5]. Forecasting and scheduling agencies serve plant owners under state rules. Open-access developers give buyers generation reports, but those come from the seller. The gap is an independent, buyer-side view that ties weather, banking and tariff slots to the buyer's own bill.
| Difficulty | Moderate (40/100) |
| Starting capital | ₹60–110 L |
| Time to launch | 6–9 months |
The forecasting itself is solved science. The work is in the rules, because every state handles banking, slots and charges differently, and bills arrive in formats only a DISCOM could love. Getting meter data from generators and DISCOMs takes persistence. Each new state is close to a new product.
No special licence, as long as Offtakely only advises and reconciles. It never buys or sells power, so the main job is following the rules your buyers live under.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't trust AI to make the final call on a billing dispute or a contract term. An analyst checks every flagged error before it goes to the DISCOM or the seller.
Want to work out the AI layer for Offtakely? Talk to Yash →
I think it's worth a look for a founder who has worked in power trading, open-access sales or a DISCOM billing team. It is a poor fit for a pure data scientist, because the value sits in the state rules. Prove one thing first - that a month's reconciliation for three real buyers finds errors worth more than the fee.
Plan for roughly ₹60–110 L, mostly on analysts, data work and a state-by-state rules library.
No sector licence, as long as you only advise and reconcile. Scheduling or trading power would need a CERC trading licence, so get legal advice before you go near that.
Yes, wherever large buyers contract wind and solar directly. The rules on banking and time slots differ by market, so check with your local regulator first.
The free Offtakely validation kit has the line-by-line capital breakdown, the effort scoring, the open-access and data checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for energy heads and developers, and 5 assumptions to test. Download the kit →
Thinking about building Offtakely or something close to it? The make-or-break work is getting meter and bill data out of generators and DISCOMs in your first state, and that's the kind of grind I help founders plan 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.