The EV Charging Station Business Plan
💡 EV Charging Station Business Plan: Key Highlights
- Utilization, not capex, is what gets a plan rejected. Independent estimates put Indian public-charger utilization at 1–5%; a plan opening at 20% is claiming 4–20× the national average.
- Your gross margin per kWh has a legal ceiling. The Ministry of Power’s 2024 guidelines cap the service charge at ₹11/₹13 per unit for DC and ₹3/₹4 for AC, excluding GST and land cost.
- One 60 kW DC charger earns roughly ₹5,200 of gross margin per month for each percentage point of utilization at a ₹12/kWh service charge — which makes break-even arithmetic, not opinion.
- Energy costs 86% more outside solar hours. Supply is billed at 0.7× average cost of supply from 09:00–16:00 and 1.3× the rest of the day — a single blended tariff throws away the biggest controllable lever in the model.
- The tariff regime expires on 31 March 2028, well inside a typical loan tenor. A flat energy-cost line to year seven is an unstated assumption a lender will find.
- PM E-DRIVE support is narrower than most plans claim — at commercial sites it covers 80% of upstream infrastructure only, on 2022 benchmark costs, and pays after you have already spent.
Most EV charging station business plans fail on a single line, and it is almost never the capital expenditure line. Setup cost is the easiest number in the document to get right — vendors quote it, and a reviewer can verify it in an afternoon. What does not survive scrutiny is the revenue build, because it rests on an assumption the author rarely labels as one: how often the charger is actually in use.
India now has 29,151 public charging stations operated by roughly 200 charge point operators, and independent estimates put capacity utilization somewhere between 1% and 5%. A plan that opens at 20% utilization in month one is quietly claiming four to twenty times the national average. Experienced lenders know that, which is why the conversation moves so quickly from your spreadsheet to your assumptions.
This article is about the document itself — its structure, its assumption register, and how each number is made defensible. It is written for first-time CPOs, franchise and dealership partners, fuel retailers and energy retailers preparing a board case or a lender pack. We do not re-derive the numbers here: EV charging station setup costs covers capex and the licensing and approvals process covers the regulatory route. This is the layer above both.
Estimated capacity utilization of India’s public charging fleet — the number most plans overwrite without saying so
Regulated ceiling on the DC service charge per unit, excluding GST and land — the upper bound on your gross margin
When the capped single-part supply tariff for public charging stations is currently set to lapse
What An EV Charging Station Business Plan Has To Prove
A business plan is not a description of your business. It is a sequence of claims, each of which a credit committee will try to break. The eight sections below are the order most Indian lenders and boards expect; the useful discipline is to write the proof obligation at the top of each before writing a word of content. A section that cannot state what it proves is filler, and it weakens everything around it.
| Section | What it has to prove |
|---|---|
| 1. Executive summary | The ask, the use of funds, and the one number the entire case depends on — stated, not buried. |
| 2. Demand evidence | That demand exists at your sites. National EV sales growth is context, not evidence. |
| 3. Site portfolio & rollout | That every site is real: land control, sanctioned load, DISCOM feasibility, commissioning date. |
| 4. Technical & operating plan | Charger mix and power rating per site, the uptime target, and who physically fixes a fault at 2 a.m. |
| 5. Regulatory position | That you know charging is an unlicensed activity, and that your pricing sits inside the 2024 service-charge ceilings. |
| 6. Financial model & assumptions | That every input is sourced, ranged and challengeable — the assumption register below. |
| 7. Risk & sensitivity | That you have already stress-tested yourself harder than the lender intends to. |
| 8. Funding structure & exit | That the loan tenor is longer than the payback period, and what the lender recovers if it is not. |
The proof obligation per section. Sections 6 and 7 are where plans are won or lost; sections 1–5 exist to make them believable.
Two deserve a note. Section 3 is where first-time plans quietly overstate progress — “site identified” and “site secured with a sanctioned load” are separated by months of DISCOM process and a very different risk profile, and a reviewer will ask which you mean. Section 8 tests the structure itself: a model showing six-year payback against a four-year term loan has answered its own question in the wrong direction.
The Assumption Register: Every Number The Model Rests On
The assumption register is the heart of the document: one table listing every input, the range you consider defensible, and the reasoning behind it. A single number per input signals false precision; a range with a stated basis signals you have thought about what could be wrong. Build it before the spreadsheet, because the register is what the model consumes.
Two entries carry disproportionate weight. The first is utilization — and the single most common reason a charging plan is rejected is that utilization is entered as a constant. Real networks ramp: a site opens near zero, builds as it enters navigation apps and roaming networks, and plateaus at a level set by its catchment. A flat 18% from month one is not an optimistic assumption, it is a missing one. Model it month by month for at least 24 months, and state what drives each step — a fleet contract, a corridor opening, a competitor’s site closing.
The second is the energy tariff, which is not one number either. Under the Ministry of Power’s 2024 guidelines, supply to a public charging station is billed at 0.7 times the average cost of supply during solar hours (09:00 to 16:00) and 1.3 times outside them. That is an 86% swing in your input cost depending on when the kWh is delivered. A model with one blended energy price has thrown away the largest controllable margin lever in the business before the first charger is even installed, and a plan whose load profile is deliberately weighted toward solar hours is materially more defensible than one that is not.
| Assumption | Defensible range | Basis and what moves it |
|---|---|---|
| Utilization ramp | 1–3% at launch → 8–15% by month 18–24 | National estimates sit at 1–5%. Anything above that band needs a named reason: an anchor fleet, a captive site, a corridor monopoly. Never flat. |
| Sessions per day | Derived, not assumed | Utilization × 24 h ÷ average session duration. At 10% utilization and a 45-minute DC session, that is ~3.2 sessions per gun per day. If your stated sessions and utilization disagree, the model is internally inconsistent. |
| Average kWh per session | 18–28 kWh (4W DC); 2–4 kWh (2W/3W AC) | Battery capacity × the usable state-of-charge window, typically 30–80% on a public fast charger. Do not model full-pack charges. |
| Energy purchase tariff | ₹6–12/kWh across states | Model 0.7× average cost of supply for 09:00–16:00 and 1.3× outside, per the 2024 guidelines — not one blended rate. |
| Service charge realised | 70–90% of the regulated ceiling | Ceilings are ₹11 (solar) / ₹13 (non-solar) per unit for DC and ₹3 / ₹4 for AC, excluding GST and land. Subscriptions, launch discounts and aggregator commissions all pull the realised figure below the cap. |
| Post-2028 tariff | State it explicitly | The capped single-part supply tariff runs only to 31 March 2028. Years 3 onward of a seven-year model sit outside it. Either model a two-part tariff with demand charges from FY2028-29 or flag the assumption. |
| Availability / downtime | 92–97% uptime | Every percentage point of downtime is a percentage point off revenue — and a chargeback if you have signed an uptime SLA with a site host or fleet customer. |
| Payment processing | 0% to ~1% of collections | UPI and RuPay debit carry zero MDR in the covered bands; cards do not. Model by expected payment mix — a highway corridor with fleet cards looks nothing like an urban mall. |
| Software / platform | Recurring, per charger per month | An operating line, never a one-time item. Includes session management, settlement, roaming and the reporting your lender will ask to see quarterly. |
| O&M, spares, insurance | 3–6% of EVSE capex per year | Guns and cables are consumables at high session counts. Add a replacement provision rather than treating breakage as an exception. |
| Land / lease | ₹1/kWh (public land) or negotiated rent | On public land the 2024 guidelines set a ₹1/kWh share to the land-owning agency, paid quarterly. On private land it is rent or a revenue share — and it sits outside the service-charge ceiling. |
An assumption register lists the range and the reasoning, never a single number. Ranges are what let a reviewer test you instead of doubting you.
Platform cost is the line most often set to zero, and it is the one a lender will probe hardest, because it is the line that determines whether the reporting in the plan can actually be produced. If the model promises quarterly utilization and uptime reporting, something has to generate it — which is why the cost of a charging management platform like YoCharge belongs in the operating column from month one, alongside O&M rather than buried in it.
The Financial Model: One Site Versus Ten
Keep the model monthly for the first 24–36 months and annual thereafter, with three linked sheets: a P&L, a cash flow, and the assumption register feeding both. Resist the urge to model a network as one blended site — the whole point of a multi-site case is that the costs behave differently, and a lender reads the ten-site model specifically to see whether you understand how.
The divergence shows up in three places. Shared operations costs genuinely are shared, so per-site overhead falls as sites are added — until a step change. Headcount does not scale smoothly: one operations engineer covers a handful of sites, and the eighth triggers a hire that appears as a cliff in the P&L, not a slope. And working capital, negligible in a one-site plan, becomes the binding constraint in a rollout: energy is bought before revenue is collected, and subsidy arrives long after capex is spent.
| Model line | One site | Ten sites |
|---|---|---|
| Revenue build | One utilization curve | Ten independent ramps with different start dates — never one average applied to all |
| Energy cost | Solar / non-solar split by load profile | Same split per site; profiles differ sharply between highway, mall and depot |
| Site fixed cost | Rent, connection, insurance, O&M | Same per site — this genuinely does not benefit from scale |
| Shared ops & software | Allocated in full to one site | Spread across ten; per-site cost falls, total cost rises |
| Headcount | Usually the founder | Step changes, not a slope — model the hire, not an average |
| Working capital | Negligible | The binding constraint: energy paid before collections, subsidy reimbursed after spend |
| Capex phasing | Single month | Staggered — and each site’s payback clock starts at its own commissioning date |
Where the one-site and ten-site cases diverge. A ten-site model that is simply a one-site model multiplied by ten is the fastest way to signal inexperience.
Subsidy deserves its own treatment in the cash flow rather than the P&L. Under PM E-DRIVE, support at ordinary commercial locations — malls, streets, market complexes, highways — is 80% of upstream infrastructure only, with the charger itself unsupported; the higher categories that also cover equipment are reserved for government premises and public-sector sites. The subsidy is calculated on the lower of actual cost or a benchmark published in August 2022, and 70% of it releases only after the DISCOM demand note has been paid and the chargers procured. It is a reimbursement with a lag, not a reduction in what you must fund on day one. What CPOs can actually claim in 2026 sets out the routes in detail.
Payback, IRR And Break-Even Utilization: The Metrics That Get Asked
Four metrics carry almost every conversation. Payback period — months to recover capex from operating cash flow, reported per site and for the portfolio, because they differ. Project IRR — the return on the asset before financing, which is what tells a lender whether the business works independently of how it is funded. EBITDA per charger per month — the cleanest unit-economics figure and the easiest to benchmark. And break-even utilization — the one most plans omit and the one most likely to be asked for first.
Break-even utilization is arithmetic, not judgement. One 60 kW DC charger running at 100% would deliver 43,200 kWh a month, so each percentage point of utilization is 432 kWh. At a realised service charge of ₹12/kWh, that is roughly ₹5,200 of gross margin per month for every point of utilization. Against ₹50,000 of monthly site fixed cost — rent, insurance, O&M, platform, share of supervision — break-even lands near 9.6%. Set against a national band of 1–5%, that single line explains more about this industry’s economics than any market-size chart, and stating it yourself is far stronger than having it extracted from you.
Sensitivity should then run on the two variables that actually move the answer: utilization and the realised margin per kWh. Everything else is noise by comparison.
| Utilization | kWh/month | At ₹8/kWh | At ₹10/kWh | At ₹12/kWh |
|---|---|---|---|---|
| 5% | 2,160 | ₹17,280 | ₹21,600 | ₹25,920 |
| 10% | 4,320 | ₹34,560 | ₹43,200 | ₹51,840 |
| 15% | 6,480 | ₹51,840 | ₹64,800 | ₹77,760 |
| 20% | 8,640 | ₹69,120 | ₹86,400 | ₹1,03,680 |
Illustrative monthly gross margin, one 60 kW DC charger. Green clears a ₹50,000 monthly site fixed cost; amber does not. Note the shape: moving from 10% to 15% utilization is worth more than a ₹4/kWh price increase — which is why operational levers beat pricing levers in almost every charging plan.
Present the grid, not a single downside case. A reviewer wants the break-even frontier and where your base case sits against it — and the shape of that frontier is itself an argument: utilization work pays better than price increases, which your plan should then reflect in how the money is spent. Profit margins by site type is the right benchmark for where your base case should land.
The Questions Lenders And Boards Actually Ask
These come up in nearly every credit conversation, and none are hostile — they are the standard tests for an asset class with a short track record and thin operating data. A plan that answers them pre-emptively, in its own risk section, converts an interrogation into a review. Each answer below is the shape of a good response, not a script.
Show the downside case already built, with the month cash turns negative and the specific action taken at that point — tariff change, site mothballing, a renegotiated lease. “It would be difficult” is not an answer; a trigger and a response is.
Point to the clause. A charger is normally treated as removable equipment rather than a fixture of the land, so the default position on early termination is that you remove it — which means your model needs a relocation cost and the dead months around it, not a residual-value line.
The right answer is a number: project IRR with and without support. If the case only works with subsidy, say so and show the timing risk, because PM E-DRIVE proposals are routed through central ministries or state nodal agencies and CPOs are then selected by open tender — the support is a pipeline you compete in, not a form you file.
Public charging has very low switching costs, so model a utilization haircut rather than assuming loyalty. What genuinely defends a site is contracted demand — a fleet, a captive depot, a housing society — and roaming reach that keeps you visible in other operators’ apps.
The question fewest plans expect. The capped single-part supply tariff expires inside a normal loan tenor, so a flat energy line to year seven is an assumption about future regulation. Model the cliff, or state plainly that you have not.
Underneath all of these sits a funding-route question: is charging being underwritten as equipment finance or as infrastructure? The distinction decides your tenor, and tenor decides whether a six-year payback is financeable at all. Credit-enhancement facilities exist for exactly this gap — the SIDBI–Shell Foundation risk-sharing facility covers charging infrastructure on a second-loss basis, with lenders absorbing the first 3% of portfolio losses and 75% cover on the next 10%. Knowing which instruments your lender can access is part of arriving prepared; investment models for charging businesses covers the equity side.
How An EV Charging Business Plan Falls Apart
Most rejected plans fail in one of a handful of recognisable ways, and all of them are avoidable at the register stage rather than the review stage. The three below account for the majority of what reviewers actually flag.
The defining error. It usually happens in reverse: the author finds break-even at ~10%, enters 18% to make the returns work, and never revisits it. A reviewer who computes your break-even in thirty seconds will see exactly what happened. Roughly nine in ten EV owners charge primarily at home or at work, which caps public demand structurally — your ramp has to respect that ceiling, not argue with it.
Two versions of the same mistake. Sanctioned load is billed whether or not it is used, so an oversized connection at a low-utilization site quietly destroys the margin — and the current capped single-part tariff, which softens this, is not permanent. A model that runs to year seven on today’s tariff structure is assuming a regulatory decision that has not been made.
Both are recurring, both scale with sessions rather than sites, and together they frequently exceed the entire net margin at low utilization. A plan that shows ₹0 for platform costs is also implicitly claiming it will produce uptime and utilization reporting by hand — which is the same plan that later cannot answer the lender’s quarterly questions.
The pattern is worth naming. Every one of these failures is a plan hiding an assumption rather than declaring one — and hidden assumptions are exactly what a credit process is built to find. An EV charging station business plan that opens at 3% utilization, ramps slowly, states energy cost by time of day, prices inside the regulatory ceiling and still shows a return is more fundable than one reaching a better headline IRR by declining to be specific. Once the network is live, that discipline becomes reporting: published assumptions turn into variances you explain, which is why multi-site operators end up running the model and the charge point operator platform as one system. On protecting the utilization the plan promised, charger placement and site selection is where the upside is won.
Frequently Asked Questions
Sources: Ministry of Heavy Industries — PM E-DRIVE EV PCS Operational Guidelines | Ministry of Power EV Charging Infrastructure Guidelines, 2024 | Business Standard — India’s EV charging stations cross 29,000 | Autocar India — Charging network grows, usage remains low | RMI — De-Risking Lending for a Brisk EV Uptake
Put real platform numbers in your financial model
Software, settlement and reporting are operating lines in every credible charging business plan — and the ones reviewers test hardest. See what those lines actually look like for a network your size before you commit them to the spreadsheet.
Talk to Sales