← Projects

Research Brief · August 2026

Cloning One Bharat Charge:
what the idea really is,
and what it would take.

A teardown of onebharatcharge.com — an EV-charging aggregation play for India — checked against the live 2026 market, the competitive field, the protocol stack, RBI's payment rules, and the actual unit economics. Written to be built from, not pitched from.

The idea
One app + one wallet across every Indian charging network, via OCPI 2.2.1 and Beckn/UEI.
Is it real?
The problem is real and unsolved. The solution is already attempted by 6+ funded players.
Hardest part
Not the code — it's CPO commercial consent, wallet licensing, and session volume.
Recommendation
Don't clone the consumer app. Clone the rails underneath it. See §5–§6.

Contents

What's in this brief

Fourteen segments. If you're short on time, read §04, §06 and §10 — the case against, the way around it, and the arithmetic.

Every heading is also reachable from the sticky bar at the top of the page.

00 — Plain English

Every term in this document, decoded

EV charging has an unusually dense acronym soup, and the original site uses all of it without explanation. Read this once and the rest of the brief is straightforward.

The 30-second version

Someone owns the physical charger (CPO). Someone sells you access to lots of chargers through an app (eMSP — this is what One Bharat Charge wants to be). OCPI is the language those two speak so your app can start a stranger's charger and the money ends up in the right place. UEI is India's free, government-backed version of that same idea. Everything else below is detail.

Who's who in the value chain

CPOCharge Point Operator
The company that owns and runs the physical chargers — buys the hardware, pays the electricity bill, fixes it when it breaks. Tata Power, ChargeZone, Statiq.
eMSPe-Mobility Service Provider
The company that gives drivers access to chargers they don't own — an app, an account, a bill. It owns the customer, not the hardware.
Aggregator
Everyday word for an eMSP that pulls many CPO networks into one app and one payment method. Exactly what this whole business idea is.
Roaming
Using a charger belonging to a network you never signed up with, and it just works — because your provider has an agreement with theirs.
Roaming hub
A middleman that connects many CPOs to many eMSPs at once, so nobody has to build a separate integration with every counterparty. One plug instead of fifty.
Site host
Whoever owns the land the charger sits on — a mall, hotel, office park, petrol pump. They usually take a cut of revenue or rent for the parking bay.

The physical stuff

Charging station
One physical location. It may contain several chargers — which is why national "station count" numbers and "charge point" numbers differ so much and get misquoted.
Charge point / EVSEElectric Vehicle Supply Equipment
One individual charging unit that one car plugs into at a time. The unit that actually matters for capacity.
Connector
The physical plug type. India uses CCS2 and CHAdeMO for fast DC, Type 2 for AC, plus the Indian government's own Bharat AC-001 and Bharat DC-001 standards for low-cost 2W/3W charging.
AC vs DC charging
AC is slow (hours) and cheap — homes, offices, overnight. DC fast charging is minutes-to-an-hour and expensive — highways, public top-ups. Public charging revenue is overwhelmingly DC.
kWhkilowatt-hour
The unit of electricity sold — the "litre" of EV charging. A typical Indian electric car takes ~15 kWh in a public top-up. Tariffs are quoted per kWh.
SoCState of Charge
How full the battery is, as a percentage. The "fuel gauge". A trip planner that does SoC simulation predicts your battery level at each stop along a route.
Utilisation
What fraction of the day a charger is actually being used. The number that decides whether a CPO makes money. Most Indian public chargers sit painfully low.
Uptime
Percentage of time a charger actually works. India's weak spot — of 27,737 registered stations, roughly 5,000 aren't operational. This gap is Wedge 01 in §5.

The protocols (how the software talks)

OCPIOpen Charge Point Interface
The open standard that lets a CPO's system and an eMSP's system talk: share station lists, check availability, start a session remotely, and exchange the bill afterwards. Version 2.2.1 is the current industry norm; 2.3 exists.
OCPPOpen Charge Point Protocol
Easily confused with OCPI, and different. OCPP connects a charger to its own operator's backend. OCPI connects two companies. OCPP is vertical, OCPI is horizontal.
OCPI modules
OCPI is split into named parts you implement one at a time: locations (where the chargers are), sessions (a charge in progress), cdrs (the receipt), tariffs (pricing), tokens (who's allowed to charge), commands (start/stop remotely), credentials (the handshake).
CDRCharge Detail Record
The itemised receipt a CPO sends after a charge ends: kWh delivered, duration, price, taxes. It's what everyone bills against — and it can arrive seconds, hours, or never after the session. That unreliability is a core engineering problem (§8).
Session
One car plugged in, from start to unplug. The atomic unit of this entire business — revenue, take-rate and volume are all measured per session.
Token
In OCPI, the credential proving a driver is allowed to charge — an RFID card, an app identity, a fleet driver's ID. Fleet management is largely token management.
Credentials handshake / token_a
The one-time setup where two companies swap API keys and agree to trust each other. The "48-hour onboarding" the original site advertises is essentially this plus a data sync.
Beckn
An open, India-built protocol for letting any buyer app discover and transact with any seller, without a central marketplace owning the middle. The same machinery underneath ONDC.
UEIUnified Energy Interface
Beckn applied to energy — India's open, Ministry-of-Power-endorsed network for EV charging discovery and payment. Live since April 2024, ~13,000 sessions/day.
BAP / BPPBeckn Application / Provider Platform
On a Beckn network, the BAP is the buyer-side app (what you'd build) and the BPP is the seller-side platform (the CPO). Rough equivalents of eMSP and CPO.
ONDCOpen Network for Digital Commerce
The government-backed open e-commerce network built on Beckn. Relevant here only as precedent and shared infrastructure — it's why UEI has political backing.
Interoperability
The umbrella goal: any car, any app, any charger, one payment. Every protocol above exists to deliver it. Its absence is the problem this whole business claims to solve.

Money and payments

Tariff
What the driver is charged — per kWh, per minute, a flat session fee, or a mix. Set by the CPO, not by you.
Take rate / platform fee
Your cut of each transaction, as a percentage. Realistically 3–8% here. The heart of §10.
GMVGross Merchandise Value
Total value of transactions flowing through the platform — not your revenue. Your revenue is the take rate applied to GMV. Confusing the two is the classic marketplace pitch-deck sin.
Reconciliation / settlement
Matching every charge to every payment to every payout, and paying each CPO the right amount. "Automated CDR reconciliation" means doing this with no human in the loop — genuinely valuable, and where trust is won or lost.
Pre-auth & capture
Block an estimated amount on the card/UPI mandate before charging starts, then take only the real amount once the CDR arrives, releasing the rest. Standard practice at petrol pumps abroad; the licence-free alternative to a wallet.
PPIPrepaid Payment Instrument
RBI's term for a stored-value wallet. A semi-closed PPI — money you top up and spend at third-party merchants — is a licensed activity with net-worth, escrow and KYC obligations. This is why §9 recommends skipping the wallet at launch.
Wallet float
The pile of customer money sitting unspent in wallets, and the interest earned on it. A real revenue line at enormous scale; a rounding error before that, and escrow rules limit it.
UPI / UPI Autopay / RuPay
India's instant bank-to-bank payment rail; the recurring-mandate version of it; and the domestic card network. Near-zero transaction cost on UPI is a large part of why Indian unit economics can work at all.
Payment aggregator / gateway
Razorpay, Cashfree and similar — they take the payment and route money to the right party. Using one keeps you out of most payments regulation, provided funds don't rest with you.
CACCustomer Acquisition Cost
What it costs in marketing to get one user. At ₹10–25 revenue per session, paid app installs almost certainly don't pay back — which is why §5 argues for fleets over retail.
Asset-light vs asset-heavy
Asset-heavy players buy chargers (capital-intensive, slow, defensible). Asset-light players sell software over someone else's chargers (fast, cheap, easy to copy). This idea is firmly asset-light — which is both its appeal and its weakness.

Policy, bodies and law

MoPMinistry of Power
Sets the rules for charging infrastructure. Its Guidelines of 17 Sept 2024 made operating a charging station an unlicensed activity and laid down interoperability standards.
FAME-IIFaster Adoption & Manufacturing of EVs, phase II
The older central subsidy scheme, including ₹912.50 Cr for public charging. Largely wound down — quoting it as current policy dates a pitch.
PM E-DRIVE
The successor scheme, and the one that matters now: ₹2,000 Cr for charging infrastructure, targeting 72,300 public stations by 2028.
NITI Aayog
The central government's policy think tank. No regulatory power, but its positions heavily shape EV and open-network policy direction.
DPDP Act 2023Digital Personal Data Protection
India's data-protection law. Location history and charging behaviour are personal data, so consent, purpose limitation and breach reporting all apply to this product.
EV penetration
EVs as a share of all vehicles sold in a period. India crossed 12% overall for the first time in June 2026 — the single strongest number in the whole thesis.

How one charging session actually flows

Every term above, in the order it appears in real life:

01 · DISCOVER
Driver opens the eMSP app. It shows chargers pulled from many CPOs via OCPI locations and Beckn search.
02 · AUTHORISE
Driver taps Start. The app pre-authorises ~₹400 on UPI and sends the driver's token to the CPO.
03 · START
eMSP sends OCPI START_SESSION. The CPO's backend tells the charge point over OCPP to unlock and deliver power.
04 · CHARGE
A session object streams live kWh and SoC back to the app until the driver unplugs.
05 · BILL
CPO issues a CDR — 14.2 kWh, ₹284. The app captures ₹284 and releases the rest of the hold.
06 · SETTLE
Reconciliation matches CDR to payment, pays the CPO its share, and keeps the take rate — about ₹15.

01 — Teardown

What onebharatcharge.com actually is

Read carefully, the site is a fundraising and CPO-recruitment landing page — not a live consumer product surface. That distinction matters enormously for anyone copying it.

SITE

A single static HTML page

~87 KB, one file, no framework, no build step. Google Fonts (Playfair Display / DM Sans / DM Mono). Nine anchor sections: Opportunity, Platform, Investors, For CPOs, Bharat, USP, App, About, Contact.

PRODUCT

Android-only, claimed live

The page states the Android app is "fully functional" with station map, booking, live sessions and a unified wallet. No Play Store link, no iOS, no web app, no public API docs, no pricing page.

COMPANY

AuraCodings, Noida

An independent studio positioning itself around telecom infra, enterprise architecture and AI/RAG. Three contact addresses on the onecharge.in domain — investors@, partners@, hello@.

What it claims to have built

Claims vs. the 2026 record

Several headline numbers on the site are stale or soft. If you reuse this narrative, fix these first — an investor who checks will find them in one search.

"India has 25,000+ public EV charge points"
Government told Parliament in July 2026 there are 52,718 public EV charging stations, incl. 16,561 fast chargers for cars. Separately, of 27,737 registered stations at March 2026, only ~22,753 were operational — the gap between installed and working is itself a product opportunity.
"₹50,000 Cr annual charging revenue by 2030"
A projection, not a fact, and unsourced on the page. Directionally plausible but treat as a TAM slide, not a plan input. Your revenue is a take-rate on sessions, not on this number.
"OCPI 2.2.1 is required for FAME-II funding"
Overstated. The MoP Guidelines of 17 Sept 2024 set interoperability standards and make charging an unlicensed activity; FAME-II is largely superseded by PM E-DRIVE (₹2,000 Cr for charging, target 72,300 stations by 2028). OCPI is the de-facto industry standard — not a hard statutory mandate.
"The aggregation layer does not yet exist at scale"
It partly does. UEI (Beckn-based, government-endorsed) already clears ~13,000 sessions/day and 60 MWh/day across 30 member organisations. ChargeZone reaches 13,500+ stations through OCPI roaming with Statiq, Bolt, Kazam and Pulse Energy. ElectricPe advertises 25,000+ stations from 60+ networks.

02 — Market reality, mid-2026

The demand side is finally real

This is the strongest part of the thesis. India crossed double-digit EV penetration in 2026 and charging supply is being funded aggressively. The market timing argument holds up.

52,718
Public EV charging stations in India, incl. 16,561 fast chargers for cars
Govt. to Parliament, Jul 2026
12%+
EV share of all vehicle retail, first time crossed — June 2026 (3,06,220 units)
EVreporter, Jun 2026
2.45M
EVs sold in FY2026, all four segments in double-digit growth; 8.27% FY penetration
Autocar Professional
1,93,633
Electric 4W registrations in FY26, up 91.3% YoY — the segment that actually uses public DC charging
FY26 registration data
1.40M
Electric 2W in FY26 (+21.8%), 11.2% penetration by July 2026 — but mostly home-charged
EVreporter, Jul 2026
72,300
Public stations targeted by 2028 under PM E-DRIVE (₹2,000 Cr allocated)
Ministry of Heavy Industries

The number that should drive your product decisions

Public charging demand is overwhelmingly a four-wheeler and commercial-fleet phenomenon. 1.4M electric two-wheelers sold in FY26, but they charge at home off a 5A socket. The ~194k electric cars — plus e-3W fleets, e-cabs and last-mile delivery — are what fill a DC bay. Sizing your app to "40M EVs by 2030" is a category error; size it to public sessions.

Supply-side structure

The market is genuinely fragmented, which is the premise the whole idea rests on — and that premise checks out. No single CPO holds above 20% network share; the top five (Tata Power, Tecso ChargeZone, Sharify, Fortum, Kazam) together hold roughly 55–60% of active public charge points. A driver genuinely does need multiple apps.

03 — Competitive field

You would be the seventh entrant, not the first

This is the part the original site does not mention at all. Every element of its pitch — one app, one wallet, OCPI roaming, UEI membership — is already shipped by someone with funding.

PlayerTypeScale / tractionCapitalThreat to a clone
UEI / Beckn
Alliance
Open network ~13,000 sessions/day, 60 MWh/day, 30 members. Launched Apr 2024, MoP-endorsed. Public good Existential
It commoditises discovery + roaming — the exact layer you'd charge for.
ElectricPe Pure aggregator 25,000+ live stations, 60+ partner networks. Founded 2021. Has diversified into EV retail, finance, servicing, 126+ stores. ~$8M+ disclosed Direct
Same pitch, five years of head start. Note the pivot into retail — the aggregator alone didn't pay.
Pulse Energy Aggregator + CPO SaaS Positions as India's largest aggregator of fast charging; carries Tata Power EZ, Electreefi, PlugShare and others. Vocal UEI advocate. VC-backed Direct
Owns both the consumer app and the CPO back-end — the strategy you'd want.
ChargeZone CPO turned network 13,500+ stations reachable via OCPI roaming deals with Statiq, Bolt, Kazam, Pulse Energy. ~$97.8M High
Proves CPOs can roam bilaterally and skip the aggregator entirely.
Statiq CPO, app-first 7,000+ charge points, 60 cities. Won a BPCL tender. $18M round Feb 2026 (Tenacity Ventures, Y Combinator, Shell Ventures). $45.5M total High
Best consumer app UX among CPOs; drivers may never feel the need for a layer above it.
Kazam CPO SaaS / asset-light ₹40 Cr revenue FY25, tracking ₹100 Cr run-rate, guided to profitability by early 2026. IFC-backed. Strong in 2W/3W and residential. IFC + VC High
The only one with proven revenue discipline. Owns the low-cost network layer.
Tata Power EZ, Jio-bp Pulse, BPCL, Ather Grid, Bolt.Earth OEM / oil-major captives Balance-sheet-funded, bundled with vehicle or fuel-retail footprint. Corporate Structural
Zero incentive to let an aggregator own their customer. Hardest to sign.

The one genuinely encouraging data point

Discovery through aggregators reportedly drives 40–60% of first-time sessions at a station. That is real leverage over a CPO — it is the single strongest argument you have in an integration call, and it is why CPOs sign roaming deals at all.

04 — The honest case against

Five reasons pure aggregation is a hard business

None of these are reasons not to build. They are the things that will actually decide the outcome, and none of them appear on the original site.

01

Aggregation is not a moat

The site argues OCPI integration creates "high switching costs." It does the opposite — OCPI is a standard. A CPO that integrated with you in 48 hours can integrate with your competitor in 48 hours, and with a roaming hub in one. Standardisation lowers switching costs for everyone, including away from you.

02

UEI is a public utility eating your layer

Beckn/UEI exists precisely to make cross-network discovery and settlement a free public protocol, backed by the Ministry of Power and the people who built UPI and ONDC. Building a private toll booth on a rail that is being deliberately made free is a strategically exposed position. Better to build on UEI and sell what UEI doesn't: reliability, UX, fleet tooling.

03

You need CPO consent, and they don't need you

The "48-hour onboarding" flow assumes a CPO wants to be listed. Tata Power, Jio-bp and Ather have their own apps, their own wallets, and their own customer relationship to defend. The independents will list with you happily — but they're also the ones with the worst uptime. Coverage of the good chargers is the bottleneck, and it is commercial, not technical.

04

The take-rate is thin and the volume is small

A CPO's gross margin is roughly ₹8–16/kWh. Your slice of that is a few percent of the session. At a realistic ₹15–25 per session, ₹10 Cr of revenue needs on the order of ~5–7 million sessions a year — versus the entire UEI network's ~4.7M/year today. See §10 for the full arithmetic.

05

The wallet is a licensed activity

A stored-value prepaid wallet spendable at third parties is a semi-closed PPI under RBI's Master Direction — net-worth thresholds, escrow rules, audits, KYC. "Wallet float revenue" is not a casual revenue line; it is a regulated business. There is a clean way around this — see §9.

06

Android-only caps the addressable user

Public DC charging skews to premium four-wheeler owners — disproportionately iOS in India. Shipping Android-only excludes a meaningful slice of exactly the highest-value cohort.

05 — Where a copy could actually win

Five wedges, ranked

If you build the same consumer aggregator, you are competing on app-store marketing spend against six funded teams. These are the positions where the same codebase wins instead.

WEDGE 01 · STRONGEST

Sell reliability, not discovery

~5,000 of India's 27,737 registered stations are non-operational. Every EV driver's top complaint is arriving at a dead or blocked charger. Discovery is solved and free; trust is not. Build a verified-uptime layer: OCPP/OCPI telemetry, crowdsourced check-ins, a per-station reliability score, and a "we'll refund your detour" guarantee. That is a consumer reason to open your app instead of Google Maps — and it is defensible, because it's a data asset that compounds.

WEDGE 02 · BEST ECONOMICS

Go B2B: fleets, not drivers

e-cab, e-3W and last-mile delivery fleets charge daily, on contract, at predictable volumes — the opposite of the retail driver who charges twice a month. Sell a fleet console: driver RFID/token management, per-vehicle spend limits, consolidated GST invoicing, charge-vs-route planning, reimbursement. Subscription + per-session, invoiced monthly. No PPI licence needed. Ten fleets beats a hundred thousand app installs.

WEDGE 03

Be the OCPI hub, not the app

Instead of a consumer brand, be the roaming/settlement middleware: one integration for a CPO, reach into every eMSP, automated CDR clearing and dispute handling. Charge per settled CDR. This is unglamorous, hard to displace, and the layer Pulse Energy is already reaching for — which is evidence it works, not that it's taken.

WEDGE 04

Own the highway corridor

The trip planner is the one genuinely under-served feature on the site's list. Nobody in India does credible SoC-aware NH routing with live availability, real tariffs, and a booked slot at arrival. Intercity range anxiety is the actual blocker to 4W EV adoption. Narrow, emotional, marketable — and a legitimate reason for a CPO to want to be in your index.

WEDGE 05

White-label the app for OEMs and CPOs

Tata, Mahindra, MG, Ather and every mid-size CPO need a charging app and none want to build one. Licence yours: their brand, your rails, your roaming index underneath. Recurring revenue, and it converts your worst competitors into distribution.

Suggested combination

Build the consumer app as the demo and data-collection surface, monetise through Wedge 02 (fleets) from month one, and let the reliability dataset (Wedge 01) become the thing that's actually hard to copy. Skip the prepaid wallet at launch (§9).

06 — The credits & partner network

Sell the 40 minutes, not the electrons

A driver on a DC fast charger is physically immobilised for 30–45 minutes, within walking distance of one specific set of shops, with nothing to do. There is almost no other retail context this captive. The charging session is the loss-leader. The dwell time is the business.

The whole idea in one comparison

You take ₹15 on a ₹300 charging session — a 5% slice of a low-margin commodity. A ₹400 meal at the dhaba 80 metres away, at a standard 15% food-service commission, is ₹60. The samosas are worth four times the electricity. And unlike the electricity, nobody is currently competing for them at that location, at that moment.

30–45min
Captive dwell time per DC session — currently monetised by nobody
The asset you actually own
3–8%
Your take rate on charging — a regulated, commoditised, contested slice
§10
10–20%
Standard commission on food & retail — 3x the rate, on a bigger basket
Indian F&B norms

How the credit works — and the one rule that keeps it legal

Call the currency Bharat Miles for now (placeholder — but the airline-miles mental model is exactly right, and drivers will understand it instantly). Four mechanics define it:

MECHANIC 01 · EARN

Miles accrue on every session

Roughly 2% of session value — a ₹300 charge yields ~6 Miles at ₹1 each. Bonus multipliers for the behaviour you want: off-peak charging, first session at a new CPO, filing a station reliability report, a completed highway trip.

MECHANIC 02 · THE RULE

Earned only. Never purchasable. Never cashable.

This is not a design preference, it is the compliance boundary. The moment a user can buy credits and spend them at a third-party merchant, you have re-created a semi-closed PPI and dragged the RBI licensing problem from §9 back in. Earned-only, non-transferable, non-refundable, expiring loyalty points sit outside that perimeter. Confirm the current position with counsel — but design to this line from day one.

MECHANIC 03 · REDEEM

Bound to the session window and the geo-fence

Miles are spendable during the charge, within ~500 m. This single constraint does three jobs at once: it manufactures urgency, it guarantees the merchant the footfall is a real nearby customer, and it makes credit-farming fraud structurally hard. Unspent Miles fall back to a smaller general balance.

"38 minutes left. ₹120 in Miles. Café Rasoi is a 90-second walk."

MECHANIC 04 · FUND

Someone else pays for the reward

The failure mode of every loyalty scheme is funding rewards out of your own thin margin. Here you don't have to — the merchant is buying guaranteed, timed, walking-distance footfall, and the CPO is buying utilisation on a dead bay. See the funding table below.

Who funds a Mile

FunderWhat they're buyingCost to youUse it for
Merchant-funded
The default
Guaranteed footfall in a known 40-minute window, from a customer already standing 80 m away. Far better targeting than any ad they can buy. Zero Everyday redemption. This should be 80%+ of all Miles spent.
CPO-funded
The clever one
Utilisation on dead chargers. Indian CPOs' single biggest pain is bays sitting idle. Let them bid Miles to route drivers to under-used sites — a demand-routing auction. Zero Filling low-utilisation stations. Also your strongest CPO sales pitch: you bring demand, not just a listing.
OEM / co-brand Owner delight and post-sale engagement. Tata or Mahindra funds Miles for its own drivers as a bundled ownership perk. Zero Launch campaigns, cohort acquisition, white-label deals (Wedge 05).
Platform-funded
Use sparingly
Behaviour you want and can't otherwise buy: off-peak shifting, reliability reports, first session, new-corridor seeding. Real Treat as a marketing line with a hard cap. It is CAC by another name — but far cheaper CAC than app-install ads.

The insight that makes merchant acquisition tractable

Your addressable merchant market is not "restaurants in India." It is "businesses within 500 m of a charger you already index" — and you know the exact coordinates of every one of those chargers, because you pulled them over OCPI.

Map the buffer, and the national list is a few thousand merchants; the NH-48 / NH-44 / NH-19 / NH-8 charging corridor is a few hundred. That is a sales list a two-person BD team can work through in a year — as opposed to the unwinnable land-grab against Zomato and Swiggy in the general case.

Partner categories, ranked by fit

CategoryWhy it fits the 40-minute windowPriority
Highway restaurants, dhabas & food plazasPerfect duration match. Zero existing digital demand-gen. Desperate for footfall. Driver is hungry, stationary and has no alternative within 20 km.Beachhead
Cafés & QSR near urban DC hubsA coffee is exactly a 30-minute purchase. High repeat frequency. Easy to sign, easy to redeem.Beachhead
Highway hotels & resortsBig basket, high margin, and they'll often install the charger themselves to get the listing — turning a merchant into a site host.Phase 2
Convenience / fuel-station retailChargers increasingly sit at BPCL/HPCL/IOC forecourts. The shop is already there. Institutional deals, not one-by-one BD.Phase 2
Malls, cinemas, family entertainmentDwell time far exceeds the charge — the charger becomes their parking amenity. They pay for anchor footfall.Phase 2
Car care: wash, detailing, tyres, serviceContextually perfect — the car is already parked and idle for 40 minutes. Genuinely additive service, high ticket.Phase 3
EV-adjacent: insurance, accessories, tyres, AMCMiles as a discount on things only EV owners buy. High intent, and the data to target it is already yours.Phase 3

Revenue streams the partner network unlocks

StreamModelNote
Redemption commission10–20% of the partner bill when Miles are redeemedThe core line. Bigger per event than the charging fee itself.
Merchant listing / placement₹2,000–8,000 per outlet per monthPredictable recurring revenue that does not depend on redemption volume — the healthiest line on this table.
CPO demand-routing auctionCPOs buy Miles to pull drivers to idle baysSells the one thing CPOs desperately need. Pure margin; costs you nothing to fulfil.
Sponsored placementPromoted merchant in the in-session panelOnly ~3 slots exist per location, which makes them scarce and therefore valuable.
Dwell-time footfall analyticsReports to site hosts, mall operators, oil majors"Chargers here drove 1,840 walk-ins last month." That's the number a site host renews a lease on.
Co-branded OEM programmesOEM buys Miles in bulk for its ownersLarge single contracts; pairs naturally with white-labelling.
BreakageMiles that expire unredeemedReal accounting reality — but never build the plan on it. A programme optimised for breakage is one drivers stop trusting.

What it does to the unit economics

Reworking the §10 model with a commerce layer attached. Attach rate is the share of sessions where the driver actually redeems at a partner — assumed conservatively.

Charging revenue per session (from §10)₹15
Partner attach rate — urban site~12%
Partner attach rate — highway site (captive, hungry, no alternative)~35%
Average partner basket₹350 – ₹500
Commission at 15%₹52 – ₹75
Blended commerce revenue — urban (0.12 × ₹52)≈ ₹6
Blended commerce revenue — highway (0.35 × ₹75)≈ ₹26
Less platform-funded Miles (capped)−₹3
Revenue per session: urban / highway₹18 / ₹38
Sessions/day needed for ₹10 Cr — charging fees alone (§10)~15,200
…with commerce, urban mix at ₹18~15,200 → 15,200
…with commerce, highway-weighted at ₹38~7,200
Volume requirement roughly halves — plus recurring listing fees on top≈ 2.1× better

Don't oversell this to yourself

The commerce layer roughly doubles revenue per session; it does not make a volume-starved platform viable. 7,200 sessions/day is still a very large number in 2026 India. What it genuinely changes is three things: it makes each session worth fighting for, it gives you a merchant-funded reason for drivers to open your app instead of Google Maps, and — via listing fees — it adds recurring revenue that doesn't wait for session scale.

Note also the urban row barely moves. This business is a highway-corridor business. That is a real strategic constraint, and it happens to align exactly with Wedge 04.

The second-order play: Miles as a grid instrument

Once Miles can shift driver behaviour, you own a demand-response lever — and that has buyers well beyond restaurants. India's grid peaks in the evening; solar is abundant midday. CPOs pay punishing demand charges at peak.

SELL TO CPOs

Peak-shaving

Offer 3× Miles for charging 11am–4pm. The CPO's demand charges fall, its solar utilisation rises, and it funds the Miles happily because the saving exceeds the reward.

SELL TO DISCOMs

Load shifting

Distribution companies fear EV load stacking onto the 6–10pm peak. A platform that can provably move charging demand by time of day is a policy-relevant asset, not just an app.

SELL TO MERCHANTS

Off-peak footfall

Restaurants have their own dead hours. Align Miles multipliers with the merchant's empty 3–6pm window and you're arbitraging two idle assets against each other with the same token.

The cold-start problem, and how to sequence around it

This is now a three-sided network — CPOs, drivers, merchants — and each side is waiting for the other two. That is the single biggest risk in this section. The sequencing matters more than the design.

Step 1 · From day one, before any merchant exists

Accrue Miles with nothing to spend them on

Start issuing Miles from session #1. It costs nothing, it's an unfunded liability of trivial size at low volume, and it quietly builds a balance drivers want to use. Show the balance prominently. Say "partners coming soon" honestly.

Step 2 · Once you have a few hundred sessions/week in one corridor

Turn the liability into the sales pitch

Walk into highway restaurants with a specific, local, verifiable number: "1,340 drivers stopped within 500 m of you last month and they are holding ₹2.1 lakh in credits they can only spend nearby." That is a far better opening than any commission deck, and it is a claim only you can make.

Step 3 · Density before breadth

Sign every merchant around ten chargers, not one around a hundred

A driver needs to find a partner at their stop, not somewhere in India. Saturate one corridor — say Delhi–Jaipur on NH-48 — until redemption is near-guaranteed at every stop on the route. Then market that corridor as a complete experience and repeat.

Step 4 · Flip it into CPO acquisition

Use merchants to win the chargers

Once the corridor works, the pitch to a reluctant CPO inverts: you're no longer asking for a listing, you're offering a footfall-and-utilisation package their own app cannot produce. This is the point where the merchant network stops being a feature and becomes the moat.

Why this is the answer to the UEI problem

§4 argued that the aggregation layer has no moat: OCPI makes integrations cheap and UEI makes discovery free. Both are true, and neither touches this. UEI will never build a merchant network. It is a protocol, not a business — it standardises the transaction and stops there.

A local, physical, contract-by-contract network of a few thousand merchants next to specific chargers is slow to assemble, geographically defensible, and gets more valuable with every driver. It is precisely the kind of asset a protocol cannot commoditise — and the only thing in this entire brief that a well-funded competitor can't replicate in a quarter.

What has to be true for this to work

  1. Session density in one corridor first. Below roughly 300 sessions/week within a corridor, no merchant will sign. Don't start the BD before you have it.
  2. Redemption has to be frictionless. QR at the counter, instant confirmation, no app-switching, works on a dhaba's ₹6,000 Android phone with poor connectivity. Offline fallback is mandatory on highways.
  3. Merchants must be paid fast. Weekly settlement minimum. Small highway businesses run on cash flow and will drop out over a 30-day cycle.
  4. Fraud controls from day one. Bind redemption to a verified CDR, a live geo-fence and a time window. Cap per-session and per-day. Watch for merchant-collusion patterns.
  5. Clean accounting for the liability. Unredeemed Miles are deferred revenue with an expiry policy and a breakage estimate. Get the treatment agreed with your auditor before issuance, not after.
  6. GST clarity. Whether a Mile is a discount, consideration, or a voucher changes the tax treatment for you and the merchant. Settle this with counsel before the first partner contract.

07 — Product scope

What to build, and in what order

CapabilityPriorityNotes
Map + station discovery
live availability, connector type, power, tariff
P0Table stakes. PostGIS + a tile provider. Cache aggressively — CPO APIs are slow and rate-limited.
OCPI 2.2.1 eMSP
versions, credentials, locations, sessions, cdrs, tariffs, tokens, commands
P0The core asset. Build it as a standalone service with a conformance test suite from day one.
Session start/stop + live telemetryP0Remote start over OCPI commands; poll sessions for kWh/SoC. Handle the failure modes — this is where the UX is won or lost.
Payments (UPI per-session, no wallet)P0Pre-auth via UPI mandate or card, capture on CDR. Avoids PPI licensing entirely. See §9.
Station reliability scoreP0Your differentiator. Start collecting from session #1 — it can't be backfilled.
Fleet console
tokens, spend limits, GST invoicing, reports
P1Where the revenue is. OCPI tokens module already models driver credentials.
Beckn BAP on UEI
search / select / init / confirm / status
P1Joins you to the public network for free coverage. Also strong credibility signal.
CDR reconciliation + settlement ledgerP1Double-entry, immutable, dispute workflow. Get this wrong and CPOs leave.
Trip planner with SoC simulationP1Valhalla/OSRM + elevation + a per-model consumption curve. High marketing value.
CPO SaaS dashboardP2Utilisation, revenue, downtime alerts. Cheap to add once you hold the data; a retention hook.
iOS appP2Do not repeat the original's Android-only choice for long. Use a cross-platform stack so this is nearly free.
8-language localisationP2Cheap and good optics, but the 4W public-charging cohort is largely English-comfortable. Don't front-load it.
Prepaid walletP3Only after a PPI licence or a licensed BaaS partner. Not a launch feature.

08 — Technical architecture

A stack that survives contact with real CPOs

The protocol work is well-specified and genuinely not that hard. The difficulty is that every CPO's OCPI implementation is subtly wrong in a different way, so the architecture must isolate that mess behind an adapter boundary.

   ┌─────────────┐   ┌─────────────┐   ┌──────────────┐
   │ Mobile app  │   │ Fleet web   │   │ CPO dashboard│
   │ (RN/Flutter)│   │  console    │   │   (Next.js)  │
   └──────┬──────┘   └──────┬──────┘   └──────┬───────┘
          └────────────┬────┴─────────────────┘
                       ▼
              API Gateway  · JWT + refresh rotation, rate limits, idempotency keys
                       │
   ┌───────────┬───────┴────┬─────────────┬──────────────┐
   ▼           ▼            ▼             ▼              ▼
Discovery  Session     Payments     Ledger        Routing
PostGIS     orchestr.    pre-auth /    double-entry   Valhalla +
+ Redis     state m/c    capture       CDR recon      SoC model
   │           │            │             ▲
   └─────┬─────┘            │             │
         ▼                  ▼             │
  PROTOCOL ADAPTER LAYER  ── normalise every CPO quirk here ──┘
   ┌──────────────┬──────────────┬─────────────────┐
   │ OCPI 2.2.1   │  Beckn BAP   │  Direct REST    │
   │ eMSP module  │  on UEI      │  per-CPO shims  │
   └──────┬───────┘──────┬───────┘────────┬────────┘
          ▼              ▼                ▼
     CPO OCPI       UEI / ONDC       legacy CPO
     endpoints      network          private APIs

Concrete choices

LayerPickWhy
MobileReact Native or FlutteriOS parity for near-zero extra cost. The original's Android-only stance is a self-inflicted wound.
BackendGo, or Node/NestJSOCPI is I/O-bound HTTP plumbing. Either works; pick what your team ships fastest in.
Primary DBPostgres + PostGISGeospatial radius/bbox queries, plus a strict relational ledger. One database until it hurts.
CacheRedisAvailability state has ~30–60s useful life. Never call a CPO on the map-pan path.
EventsNATS or KafkaSession state changes and CDR arrival are async and must be replayable for reconciliation.
MapsMappls (MapmyIndia) or GoogleMappls has materially better Indian addressing/POI density; Google has better familiarity. Abstract behind an interface.
Routing / SoCValhalla or OSRM, self-hostedYou need elevation-aware costing and custom consumption curves. Hosted routing APIs won't give you that.
PaymentsRazorpay or CashfreeUPI mandates / pre-auth, RuPay, cards. Both support the auth-then-capture flow that charging needs.
AuthPhone OTP + JWT w/ refresh rotationIndia-standard. Don't build email/password.

The three engineering traps

1. CDR arrival is asynchronous and unreliable. A session ends; the CDR may arrive seconds later, hours later, or never. Your money flow cannot assume it. Design an explicit reconciliation state machine with timeouts, estimated-vs-final amounts, and a dispute path.

2. Pre-auth sizing. Hold too little and you eat the shortfall; hold too much and UPI users see a scary block. Model it on connector power × expected dwell, and refund the delta the instant the CDR settles.

3. Every CPO's OCPI is non-conformant somewhere. Missing tariffs, wrong timezones, stale EVSE status, UNKNOWN connector types. Write a per-CPO conformance suite and a quirks registry before you write feature #2.

10 — Unit economics

The arithmetic the pitch deck skips

The original lists five revenue streams and no numbers. Here is what one session is actually worth, using published Indian margin ranges. Treat these as an order-of-magnitude model to challenge, not a forecast.

One four-wheeler DC session

Energy delivered (typical 4W DC top-up)~15 kWh
Retail tariff to driver₹18–22 / kWh
Session GMV₹270 – ₹330
CPO gross margin on energy (published range)₹8–16 / kWh
Aggregator take-rate, realistic negotiated band3 – 8%
Payment gateway cost (UPI ≈ 0, cards ~2%)−₹0 to ₹6
Net platform revenue per session₹10 – ₹25

What that means at scale

Target annual platform revenue₹10 Cr
At ₹18 net revenue per session÷ 18
Sessions required per year~5.6 million
Sessions required per day~15,200
For comparison: entire UEI network today~13,000 / day

Read that last line twice

To make ₹10 Cr on session fees alone, you would need to route more sessions than the whole government-backed UEI network currently clears in a day. Session take-rate is a real revenue line but it is not a business on its own at 2026 volumes. It becomes one around 2028–2030 if penetration keeps compounding.

This is precisely why ElectricPe diversified into EV retail, financing and 126+ physical stores. It is the strongest single piece of evidence in this whole brief.

Revenue stack that actually closes the gap near-term

StreamModelViability now
Fleet subscriptions₹X per vehicle/month + per-session fee, invoicedStrong — contracted, predictable, no app-store CAC, daily usage
CPO SaaS dashboardPer-charge-point per-monthStrong — Kazam has proven Indian CPOs pay for software
White-label app licensingSetup fee + annual licenceMedium — long sales cycle, large contracts
Session platform fee3–8% of GMVMedium — real but volume-starved until ~2028
Reliability/utilisation dataReports to CPOs, OEMs, site hosts, lendersMedium — genuinely scarce data; needs scale first
Wallet floatInterest on prepaid balancesWeak — needs a PPI licence; escrow rules constrain it; immaterial below huge scale

11 — Build plan

Twelve months, small team

Assumes 2 backend, 1 mobile, 1 full-stack/web, plus a founder doing CPO business development full time. That last role is not optional — it is the actual bottleneck.

Phase 0 · Weeks 1–4

Prove you can get data before you write an app

Sign 2–3 independent CPOs to a non-binding test integration. Complete one real OCPI credentials handshake and pull live /locations. Register interest with the UEI alliance. If you cannot get three CPOs on a call in a month, stop here — that signal is worth more than six months of code.

Phase 1 · Months 2–4

OCPI eMSP core + read-only app

Full 2.2.1 eMSP: versions, credentials, locations, tokens, sessions, cdrs, tariffs, commands. Per-CPO conformance suite and quirks registry. Ship a map app that only finds chargers — no payments. Start recording reliability data immediately.

Phase 2 · Months 4–6

Transact end-to-end

Remote start/stop, live session telemetry, UPI pre-auth → capture-on-CDR, the double-entry ledger, and the reconciliation state machine. Launch in one city, not sixty-five. Target: 100 clean sessions with zero manual settlement intervention.

Phase 3 · Months 6–9

Fleet console — turn on revenue

Driver token management, spend limits, consolidated GST invoicing, usage reports. Sell to 3–5 e-3W or e-cab fleets in your launch city. This is the first month you have real income. Beckn BAP goes live on UEI in parallel for free coverage.

Phase 4 · Months 9–12

Reliability score + trip planner + iOS

Publish the station reliability index — it's your marketing and your moat in one artefact. Ship SoC-aware NH corridor routing. Ship iOS. Add the CPO dashboard as a retention hook. Expand city two and three only once city one's unit economics are legible.

Phase 5 · Months 12–18

Light the credits & partner network (§6)

Miles have been accruing silently since Phase 2 — that's deliberate, and by now drivers hold a balance you can point at. Pick one corridor, walk the 500 m buffer around every charger in it, and sign every viable merchant. Ship QR redemption with an offline fallback and weekly merchant settlement. Do not open a second corridor until redemption at the first is near-guaranteed at every stop.

Indicative cost

Engineering, 4 people × 12 months (India market rates)₹1.2 – 2.4 Cr
Maps / routing / infra / gateway fees₹10 – 25 L
Legal, GST, DPDP, CPO contracting₹8 – 15 L
Launch-city marketing & fleet BD₹15 – 40 L
Year-one total (excludes any PPI licence path)₹1.5 – 3.0 Cr

12 — Risk register

What kills this, and what to do about it

RiskSeverityMitigation
Major CPOs refuse to list
Tata Power, Jio-bp, Ather defend their own apps
Critical Lead with the 40–60% first-session discovery stat. Offer them fleet demand they can't source alone. Accept partial coverage and be honest in-app about what you don't cover.
UEI makes your layer free Critical Join it rather than fight it. Monetise above the protocol — reliability, fleet tooling, UX, settlement ops — not the routing itself.
Session volumes stay too low to fund the company Critical Do not depend on take-rate before 2028. Fleet subscriptions and CPO SaaS carry the P&L in years 1–2.
Reconciliation errors burn CPO trust High Immutable double-entry ledger, daily automated recon reports, a named dispute SLA. Over-invest here; it's the thing that keeps partners.
PPI licensing blocks the wallet High Launch pay-per-session. Treat the wallet as a year-two feature behind a licensed partner.
A funded incumbent copies your reliability layer Medium Data compounds and can't be backfilled — start collecting from session one. Publish the index publicly to make it the reference.
Poor CPO API quality degrades your UX
you get blamed for their downtime
Medium Show freshness timestamps and confidence, never a false "available". Honesty about stale data is a feature drivers will notice.
CAC exceeds lifetime value on retail drivers Medium Don't buy installs. Acquire through fleets, OEM bundles and highway-corridor content.
Three-sided cold start on the credits network
no drivers → no merchants → no drivers
Critical Accrue Miles from day one with nothing to spend them on, then use the accumulated balance as the merchant pitch. Saturate one corridor before opening a second. Never launch the partner network before ~300 sessions/week in that corridor (§6).
Credits get classified as a prepaid instrument High Hard design rule: earned-only, never purchasable, non-transferable, non-cashable, expiring. Get the structure reviewed before the first Mile is issued, not after.
Credit-farming and merchant collusion Medium Bind every redemption to a verified CDR, a live geo-fence and the session time window. Per-session and per-day caps. Monitor redemption-to-session ratios per merchant.
Merchants churn over slow settlement Medium Weekly payouts minimum. Small highway businesses run on cash flow and will quietly stop honouring Miles long before they tell you.
Tariff/policy change compresses margins Low You take a percentage, not a spread. Regulated tariffs actually reduce your risk.

13 — Before writing code

Seven questions to answer first

  1. Can you get three CPOs on a technical call in 30 days? Everything downstream depends on this and nothing else de-risks it.
  2. Which single city? Delhi-NCR has density and the studio is already in Noida; Bengaluru has the fleet ecosystem. Pick one and say no to the rest.
  3. Retail drivers or fleets first? This brief argues fleets. If you disagree, write down what your CAC and session frequency assumptions are.
  4. Are you building a company or a licensable product? White-label is a genuinely different business with a different codebase shape. Decide before the first architecture commit.
  5. Wallet or pay-per-session? Answering "wallet" adds a regulated workstream and months. Be sure it's worth it.
  6. What is the honest coverage number on day one? If it's 400 chargers from three independents, the app must be useful at 400 chargers. Design for that, not for 52,718.
  7. What happens if UEI absorbs the whole roaming layer in 18 months? If the answer isn't "we're fine, we sell X", X doesn't exist yet.

Bottom line

The problem is real, the timing is defensible, and the technical build is a solved, well-documented four-to-six-month exercise for a competent team. The idea's weakness is that it is a distribution and commercial-agreements business wearing a protocol-engineering costume. Copy the codebase premise; do not copy the go-to-market. Win on reliability data and fleet contracts, and let discovery be the free thing you give away.

Sources

Research trail

All figures retrieved August 2026. Market projections are third-party estimates; verify anything load-bearing before it enters a deck.