A factory with rooftop solar, a diesel genset and a new battery has three ways to get power and no one deciding between them. That's a clean energy business waiting to happen.
I'll work through it using one idea, Tariftimer, and show you four others nearby so you can compare.
The problem
Mid-sized factories have added solar fast. Many are now adding batteries. But the decisions are still made by habit. The battery charges when it charges, the genset starts when the power drops, and heavy machines run whenever the shift wants them to.
Meanwhile Time of Day tariffs mean power costs more at some hours and less at others (source: SolarQuarter). If nobody times the loads, the factory pays peak rates it didn't need to.
How the business works
Tariftimer is software that decides each day when to charge the battery, when to discharge it, when to start the genset and when to run heavy machines. The goal is simple - buy less power at peak rates.
It doesn't own any hardware. It reads the meters and inverters the factory already has and sends a schedule, either to the equipment or to the plant manager.
Who pays
The factory owner or plant head pays a fee, ideally tied to savings. Your best first buyer is a plant that already has solar and a genset and has just bought a battery, because they've spent the money and want to see it work.
What it costs
Our estimate for Tariftimer is ₹65 L–1.2 Cr over 6–9 months. Most of it goes on integrations with different inverters and battery systems, and on field visits.
Compare it with the neighbours:
- Airgauge, sensors that score room ventilation in clinics and classrooms, at ₹1.68–3.1 Cr over 4–6 months.
- Prackit, offline science practicals for state board schools, at ₹60 L–1.2 Cr over 6–9 months.
- Fundvalues, a values check on your mutual funds, at ₹50–90 L over 6–9 months.
- Geoloop, one shared cooling plant for a new housing society, at ₹2.7–5.6 Cr over 18–30 months.
Geoloop is the one to study if you like energy. It's the hardware-heavy version of the same idea - run one system well instead of hundreds badly. It also shows the risk. One Hyderabad plant's chillers are rated at 0.6 kW/TR but the plant averages 2.07 kW/TR over a year (source: BEE and GIZ, 2023), which wipes out the savings. Software that only looks good on the spec sheet fails the same way.
Licences
Tariftimer is software, so it needs no licence to operate. The rules that matter are the ones its customers' equipment sits under.
- ToD tariff rules under the Electricity (Rights of Consumers) Amendment Rules 2023 set the bands. Each state commission sets its own numbers, so your schedule logic has to read the tariff for each state.
- Equipment and grid rules for the battery and solar sit with the factory and its installer. Don't take that duty on.
- The baseline: company or LLP registration with the MCA, and GST once you cross the threshold.
The first 90 days
Days 1–30. Visit ten factories in one industrial cluster that already have solar. Ask for three months of power bills and genset fuel records. If you can't show a saving on paper from their own bills then the product won't show one in real life.
Days 31–60. Run the schedule by hand for two plants. A spreadsheet and a daily phone call to the plant manager is enough. You're testing whether they'll follow a schedule, not whether your code is pretty.
Days 61–90. Measure the bill against the same month last year, and against the plan. If the gap is real then turn the spreadsheet into software and charge for it.
How AI fits
AI is useful for forecasting solar output and the plant's load from past data. It's not useful for the final call on starting a genset during a critical batch. Keep the plant manager able to override everything, and log why they did.
The risks
- Thin savings. Where the gap between peak and solar-hour prices is small, the fee may eat most of the gain.
- Uneven rollout. ToD depends on smart meters, and states are moving at different speeds.
- Vendor lock. Battery and inverter makers may offer their own scheduling for free.
My take
I like this one. The buyer has already spent money and wants a return, which is the easiest sale there is. Start in one state with clear ToD bands, prove a saving on real bills, and let the plant managers sell it to their friends in the next cluster.
For more, see the sustainability ideas and the India idea hub.
Our estimate for Tariftimer is ₹65 L–1.2 Cr over 6–9 months. A hand-run pilot costs far less.
Does energy scheduling software need a licence?
No operating licence. It works inside ToD tariff rules set by each state commission, and the factory keeps its own equipment approvals.
Who should I sell to first?
A factory that already has solar and a genset and has just added a battery.
Talk to Yash
If you're building energy software for factories and want to check the savings case before you write code, I can help you plan the pilot. Talk to Yash →
For illustration only. Ideas are hypothetical and names fictional; sourced facts 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.
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.