Candlepay

Remote-lock meters and UPI instalments for small-town solar battery dealers

IoT Tier 2–3 Cities

TL;DR Candlepay lets a small-town solar dealer sell a home inverter, battery or solar kit on daily or weekly UPI payments. A module in the unit switches it off when payments lapse and back on when they resume. The dealer or a partner NBFC carries the credit, and Candlepay charges per unit for the module and software.

The problem

Where the grid runs short of hours, families buy their own backup. A 2025 study of Bahraich in Uttar Pradesh found households there averaged 8.1 hours of grid supply a day in 2019[1]. An inverter with a battery or a small solar kit sorts out evenings and exam nights, but it costs tens of thousands of rupees upfront.

Local dealers already sell on informal credit to customers they know, and they chase dues on foot. They have no way to stop a unit working when a buyer stops paying, so they either say no to credit or sit on bad debts. We found no public figure on how big these dealer credit books are or how often they go bad. That's the first thing to measure.

The idea

What it does, in regulatory terms: payment-linked device control and collection software. Candlepay does not lend. In dealer mode the dealer sells its own stock on instalments, which counts as a trader's credit sale as long as finance stays a small side of its business. The RBI's 50-50 test decides when an entity becomes an NBFC[2]. Bigger books go through a partner NBFC, and then Candlepay is a Lending Service Provider under the Digital Lending Directions 2025, with repayments paid straight into the NBFC's account[3].

A relay-and-controller module sits inside or beside the inverter or battery. Each UPI payment, by AutoPay or QR, adds a few days of credit to the unit. When credit runs out the module cuts output, and a fresh payment brings it back within minutes. The deposit is set from the buyer's history with that dealer, the unit's own payment record if they're a repeat buyer and a rough area signal from NASA's night-light data.

We reframed the seed, which scored households by satellite. Night-light data is 750 m resolution[4] and describes a village rather than a home, so here it only nudges the deposit.

Who pays

The dealer pays a per-unit fee for the module and a small monthly software charge, and gets it back by pricing the instalment plan slightly above cash. Where an NBFC funds the book it pays a servicing fee per active account instead. The household pays the dealer or the NBFC, never Candlepay. No public price list for similar services exists in India, though a global PAYG software vendor sells similar software to distributors elsewhere[5].

Why now

The Digital Lending Directions of May 2025 set clear duties for service providers working with lenders[3], so an NBFC can sign one up knowing the rules. The DPDP Rules notified in November 2025 settle how payment and device data can be used[6]. Rooftop solar is spreading too - PM Surya Ghar had reached over 5 million homes by August 2026[7], which widens the base of small solar installers.

Market signal

There's no published figure for India's dealer-financed inverter, battery or home-solar sales. The narrowest firm number is the vendor base - 29,469 active vendors under PM Surya Ghar in August 2026[7], many of them small installers who also sell batteries. Adjacent proof of the model comes from Africa, where one asset-finance firm reports over 10 million customers buying devices on daily or weekly payments[8]. Treat that as proof the mechanism works and not as an Indian market size.

Who's already doing this

One long-running Indian solar company links customers to local banks and finance institutions instead of locking devices[9]. A global PAYG software vendor supplies software to more than 200 distributors worldwide but shows no Indian presence[5]. An African asset-finance firm finances and locks its own phones in five African markets[8]. A metered pay-as-you-go LPG startup was bought by a big off-grid solar company in 2023[10]. Nobody offers a lock-and-collect layer that independent Indian dealers can run on their own credit or a lender's.

How hard is it?

Difficulty Hard (55/100)
Starting capital ₹1.4–2.6 Cr
Time to launch 9–12 months

The module is simple electronics, but it has to fit many inverter brands without voiding warranties and it has to survive tampering. The harder work is legal - keeping dealer mode a genuine credit sale and meeting every LSP duty when an NBFC funds the book.

Licences and approvals

Heavily regulated. Candlepay doesn't lend, but it works right beside lenders, so RBI rules shape the whole design.

  • NBFC registration with the RBI under Section 45-IA, needed by any company whose principal business is loans or hire purchase under the 50-50 test. Candlepay must never fund or buy receivables[2].
  • RBI Digital Lending Directions 2025. When an NBFC funds the book, Candlepay is its Lending Service Provider, with repayments paid straight to the NBFC, before you sign a lender[3].
  • WPC equipment type approval for GSM and Bluetooth modules, by self-declaration, before you sell the module[11].
  • DPDP Act and Rules 2025 before launch, for payment and device-usage data[6].
  • Payment aggregator. Collections run through a licensed aggregator, so Candlepay never holds customer money itself.

Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.

How AI can power this business

  • Deposit setting. Combine a buyer's payment history with the dealer and a rough area signal to suggest a deposit, which the dealer can accept or change.
  • Tamper detection. Flag modules that report odd voltage patterns or go silent after a missed payment, so the dealer visits before the unit is lost.
  • Payment reminders. Time UPI reminders to each buyer's usual pay day in their own language, which cuts lockouts.
  • Dealer dashboards. Summarise each dealer's book, overdue units and likely defaults in plain language every Monday.

Don't let a model make credit decisions or switch off a family's power on its own. Lending calls sit with the dealer or NBFC, and switch-offs need written grace rules.

Want to work out the AI layer for Candlepay? Talk to Yash →

Biggest risks

  • Tampering. Buyers or the local electrician bypass the relay, and the dealer loses both the unit and the payments.
  • Regulatory drift. If Candlepay or a dealer starts funding receivables at scale then it needs an NBFC licence it doesn't have.
  • Switch-off backlash. Cutting a family's power during an exam or illness invites complaints and bad press unless grace rules are written into the contract.

Our take

I'd back a founder with hardware and lending-operations experience who can sign up dealers in one district. If you secretly want to run a lending book then this isn't the vehicle. Prove two numbers first - tampering stays under 5% of units in a pilot, and on-time weekly payment beats the dealer's current informal credit.

Questions founders ask

Plan for roughly ₹1.4–2.6 Cr, mostly on the module, firmware and legal work. A one-district dealer pilot keeps early costs down.

Do I need an RBI licence to start a pay-as-you-go solar business in India?

Not if you never lend or buy receivables. You'll act as a Lending Service Provider when an NBFC funds the book, and your module needs WPC approval.

Can I start a PAYG device business in another country?

Yes, the model already runs at scale in Africa. Lending, device and data rules differ by country, so get local legal advice before launch.

Download the kit

The free Candlepay validation kit has the line-by-line capital breakdown, the effort scoring, the NBFC, digital-lending and data-protection checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for dealers and buyers, and 5 assumptions to test. Download the kit →

Talk to Yash

Thinking about building Candlepay or something close to it? The make-or-break work is a tamper-proof module and a structure that keeps you clear of running a lending business, and that structure is something I can help you get right. If you need a hand, you know where to find me. Talk to Yash →

Sources

  1. Sandwell et al., "Opportunities for decentralised solar power to improve reliability, reduce emissions and avoid stranded assets", Nature Communications 16:8061, 2025-08-28. https://doi.org/10.1038/s41467-025-62948-8
  2. RBI, "FAQs on Non-Banking Financial Companies", accessed 2026-10-21. https://www.rbi.org.in/Scripts/FAQView.aspx?Id=92
  3. RBI, "Reserve Bank of India (Digital Lending) Directions, 2025", 2025-05-08. https://www.rbi.org.in/Scripts/NotificationUser.aspx?Id=12848&Mode=0
  4. NASA Earthdata, "Black Marble", accessed 2026-10-21. https://www.earthdata.nasa.gov/data/projects/black-marble
  5. Angaza, official site, accessed 2026-10-21. https://www.angaza.com/
  6. PIB, "DPDP Rules, 2025 notified" (backgrounder), 2025-11-17. https://static.pib.gov.in/WriteReadData/specificdocs/documents/2025/nov/doc20251117695301.pdf
  7. Mercom India, "PM Surya Ghar Powers 5 Million Homes, Total Capacity Hits 15 GW", 2026-08-04. https://mercomindia.com/pm-surya-ghar-powers-5-million-homes-total-capacity-hits-15-gw
  8. M-KOPA, official site, accessed 2026-10-21. https://m-kopa.com/
  9. SELCO India, official site, accessed 2026-10-21. https://selco-india.com/
  10. PayGo Energy, official site, accessed 2026-10-21. https://www.paygoenergy.co/
  11. Bureau Veritas CPS, "India WPC ETA through self-declaration", 2019-04-12. https://www.cps.bureauveritas.com/newsroom/india-wpc-wing-introduce-equipment-type-approval-eta-through-self-declaration-saral

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.

Reviews

No reviews yet