Entry point · reviewed August 2026
How to Start a Solar Power Plant Business
Last reviewed August 2026 · models, capital, sequence, failure modes · written to be useful, not encouraging
Pick the business model before anything else, because the five available models need different capital and different skills. Then secure a site with a substation near it, confirm who buys the power, and only then commission a project report and approach a lender. A 1 MW plant needs about ₹1.3 crore of your own money, ₹6 to ₹15 lakh of spending at risk before financial close, and 12 to 18 months.
The order matters more than the effort. Site, grid, offtake, report, finance, build. Almost every stalled project we see started somewhere in the middle of that list.
The honest summary
Owning a MW-scale plant is an infrastructure business with land, a DISCOM relationship and receivables. If you want a passive return, this is not that. If you consume a lot of daytime power at an HT tariff, it is one of the best investments available to you.
Business models
Five different businesses that all look like solar
These are not variations on a theme. They need different capital, different skills and different temperaments, and choosing between them is the first real decision.
| Model | Capital at 1 MW | Return | What it really is |
|---|---|---|---|
| Captive, your own load | ₹1–1.3 cr equity | 22–28% | A cost-reduction project inside an existing business. The best return available, limited only by how much daytime power you actually consume. |
| IPP selling to a DISCOM | ₹1–1.3 cr equity | 8–11% | An infrastructure business. Regulated tariff, 25-year visibility, slow payment cycles, and it lives or dies on the cost of debt. |
| Group captive | ₹1–1.3 cr equity | 14–18% | A structuring business. Consumers hold 26 per cent equity and take 51 per cent of the generation, which buys surcharge exemption and demands annual compliance. |
| RESCO or OPEX developer | ₹4 cr per MW, all in | 12–16% | You own plants on other people rooftops and sell units. Capital-hungry, credit-exposed, and it scales only with a balance sheet. |
| EPC contracting | ₹30–60 lakh working capital | Margin, not tariff | A services business. Far less capital, project risk instead of asset risk, and it earns on execution rather than on generation. |
The question behind the table
Do you want to own an asset for 25 years, or do you want to build things for other people? Both are legitimate businesses and both are profitable. They are not the same job, and the number of people who set out to do the first and end up doing the second is not small.
Readiness scorecard
Eight things that have to be true before you spend money
Tick what is genuinely in place, not what is being discussed. The score is not a game: each unticked item is a specific thing that stops a lender, and the order they get fixed in decides how much money you put at risk.
Capital
What you actually need in the bank
Project cost is not the number that matters to you. Equity plus pre-development spend plus working capital is, and the third one is the one people forget.
| Capacity | Project cost | Equity at 30% | Cash you need |
|---|---|---|---|
| 1 MW | ₹4.05 cr | ₹1.2 cr | ₹1.5 cr including pre-development and working capital |
| 2 MW | ₹7.8 cr | ₹2.3 cr | ₹2.8 cr |
| 5 MW | ₹18.75 cr | ₹5.6 cr | ₹6.6 cr |
| 10 MW | ₹36 cr | ₹10.8 cr | ₹12.5 cr |
Equity, 25 to 30 per cent
Lenders fund 70 to 75 per cent of project cost. Equity funded from another borrowing is visible in the financials and weakens the file.
Pre-development, at risk
₹6 to ₹15 lakh at 1 MW on survey, soil, DPR, yield assessment, legal opinion and application fees. Spent before financial close and lost if the project does not proceed.
Working capital
Four months of receivables at a minimum. A commissioned plant still pays for security and cleaning while waiting to be paid.
Cost overrun cushion
Ten per cent. Lenders often require a sponsor undertaking to fund overruns, which means it has to exist somewhere.
The sequence
Twelve steps, and the order is not negotiable
Each step is cheap to do and expensive to skip. The pattern in failed projects is almost always a step taken out of order, usually money committed to land before the grid position was known.
- 01Choose the business modelCaptive, IPP, group captive, RESCO or contracting. Everything downstream depends on this.
- 02Establish the offtake in principleYour own consumption profile, a consumer who wants power, or the scheme you intend to sell into.
- 03Shortlist sites near a substationIrradiation matters, but proximity to spare grid capacity matters more at this stage.
- 04Ask the DISCOM in writingSpare bay, spare transformation capacity, how much is committed to others, and at what voltage they will accept your capacity.
- 05Check the land before payingClassification, title chain, encumbrance certificate, public paths, number of owners. A fortnight of work.
- 06Secure the siteRegistered lease with an assignment clause, or a sale with clean title. Options are cheaper than purchases while you verify.
- 07Commission the geotechnical reportIt sets the foundation design and the civil cost, and every quotation before it is provisional.
- 08Commission the DPR and an independent yield assessmentFrom someone with no interest in selling you the plant. This is the document lenders read.
- 09Sign the offtakeA tariff in a contract, not a tariff in a conversation. Nothing gets financed without this.
- 10Approach two lendersWith a complete file. Two, so that one saying no is not the end of the project.
- 11Contract the EPC and the O&M togetherPerformance ratio guarantee, milestone payments, retention, and a five-year O&M scope priced with the build.
- 12Build, test and reconcileCommission, run the performance test, then compare the first three months of generation against the DPR and ask questions about any gap.
Money at risk
What you spend before anyone commits to fund you
Between ₹6 and ₹15 lakh at 1 MW, spent before financial close, unrecoverable if the project does not proceed. Knowing the number is what lets you decide how much verification to do first.
| Item | Cost | When |
|---|---|---|
| Land survey and title verification | ₹40,000–1.2 lakh | Before any advance on the land |
| Geotechnical investigation | ₹60,000–1.5 lakh | After the site is secured, before the EPC is priced |
| Detailed project report | ₹1.5–4 lakh | Before approaching lenders |
| Independent energy yield assessment | ₹1.5–4 lakh | Before approaching lenders |
| Connectivity application and fees | ₹50,000–2 lakh | As early as possible, because it is the longest queue |
| Legal opinion on land and contracts | ₹75,000–2.5 lakh | Before signing the lease and the offtake |
| Entity formation and registrations | ₹25,000–75,000 | At the start |
| Total at risk | ₹6–15 lakh | Before financial close |
Spend the cheap money first
The DISCOM enquiry and the title check together cost under ₹2 lakh and answer the two questions most likely to kill the project. Doing them before the ₹8 lakh of reports, and long before any land advance, is the single highest-return decision in the whole sequence.
Failure modes
Six ways this goes wrong, none of them technical
We are asked to look at stalled projects fairly often. The cause is on this list almost every time, and it was almost always visible early.
Land advance before the grid check
Money paid on a parcel, then the DISCOM says the substation is full or the voltage is 110 kV. The land is now worth what it was worth as land.
A lease that is not assignable
Signed, unregistered, or registered without an assignment clause. The lender cannot take security over it and the file stops there.
One holdout owner
Eleven of twelve survey numbers agreed and one in the middle. The layout gets designed around a hole, capacity drops, and the economics change.
Offtake left for later
Everything built around a tariff nobody has signed for. This is the most common cause of a project that is complete on paper and unfundable in practice.
Equity that is itself borrowed
Visible in the financials, and it changes how a credit committee reads the promoter. It also removes the cushion for the cost overrun that usually happens.
No one owning operations
Commissioned, then nobody watching the performance ratio, the cleaning schedule or the meter reading. Losses of five to ten per cent accumulate quietly for years.
The honest section
Three cases where you should not do this
We would rather say this on a public page than in a meeting after you have spent money. If any of these describes you, the plant is probably the wrong instrument.
You want a passive return
A solar plant is an operating business with land, receivables, a DISCOM relationship and equipment. At a feed-in tariff it returns 8 to 11 per cent for that work. A bond returns less for none of it, and the gap is not as wide as the industry implies.
You have no use for the tax shield
With no taxable profits, accelerated depreciation is worth nothing in the year it arises. For a trust, a school or a new entity, letting a developer own the plant and buying units from them is usually the better arrangement.
Your consumption is small or evening-heavy
If your daytime load cannot absorb the generation, the surplus is worth the export rate rather than your tariff, and the return you were shown does not exist. Check the load profile before the capacity.
And the case where you absolutely should
You run a factory or a large commercial site, consume a great deal of daytime power at ₹7 to ₹8.50 a unit, have land or a large roof, and can fund the equity. In that case a captive plant pays for itself in three to four years and then produces power at a marginal cost close to zero for twenty more. Very little else available to an Indian business does that.
Where we fit, and where we do not
What we do, and what we deliberately do not
Blues Renewables works at MW scale as a subcontract EPC and balance-of-system contractor inside other developer projects, and as a turnkey contractor at the rooftop and ground-mount sizes we have built. We do not own plants, bid tenders, hold power purchase agreements or arrange finance. On this page in particular, we have no interest in talking you into a project.
What a feasibility read covers at no cost
- Whether the site and the substation position can actually carry the capacity you have in mind.
- Whether the land classification and title position will survive a lender.
- Whether the numbers you have been shown imply a CUF and an operating cost that are real.
- Which of the five business models fits what you are actually trying to achieve.
What we will tell you to get elsewhere
- A bankable DPR, a financial model or an independent yield assessment.
- Debt arrangement, lender introductions or financial advice.
- Land aggregation, title diligence or conversion filings.
- Tender bidding, open access registration and the DISCOM interface.
Tell us what you are trying to achieve
Not the capacity, the objective. Reducing a factory bill, deploying capital, adding a business line, or using land that is otherwise idle. The right structure follows from that, and it is a twenty-minute conversation rather than a proposal.
If the honest answer is that you should not build a plant, we will say so. It costs us nothing and it saves you the ₹6 to ₹15 lakh of pre-development spend that would have told you the same thing a year later.
Starting out, answered
How do I start a solar power plant business in India?
Decide the business model first, because the five available models need different capital, skills and appetite for risk. Then secure a site with a substation nearby, confirm who will buy the power, and only then commission a project report and approach lenders. In practice the sequence is site, grid, offtake, report, finance, build. Projects that start with equipment selection and work backwards are the ones that stall.
How much money do I need to start?
For a 1 MW plant, about ₹1 to ₹1.3 crore of equity against ₹2.8 to ₹3 crore of debt, plus ₹6 to ₹15 lakh of pre-development spend that is at risk before financial close, plus working capital for four months of receivables. If you are buying rather than leasing the land, add that in full because it usually sits outside the debt. Realistically ₹1.5 crore of available cash for a 1 MW project.
Is a solar power plant a profitable business?
It is a stable business with modest returns when you sell to a DISCOM, and an excellent one when the power displaces your own industrial tariff. Selling at a feed-in tariff of ₹3.20 a unit returns 8 to 11 per cent pre-tax with 25-year visibility. Displacing your own HT tariff at ₹7.80 returns 22 to 28 per cent. The plant is identical; the business model is not.
What licences and approvals are needed?
A project entity, land with a classification that permits a solar plant, connectivity approval from the DISCOM or transmission utility, electrical inspectorate approval for charging, and the offtake contract or scheme allocation. Generation for captive use or sale under open access does not need a generation licence, but the connectivity and inspectorate approvals are absolute prerequisites and both take months.
Do I need my own land?
No, and most projects lease. A registered 25 to 30 year lease at ₹25,000 to ₹60,000 an acre a year with an assignment clause is the norm, because tying up ₹2 to ₹5 crore in land purchase against a ₹4 crore plant changes the return more than any equipment decision. What lenders require is that the lease is registered, long enough to cover the debt, and assignable to them.
Can I start with a smaller plant and expand?
Yes, and it is often the sensible route: build 1 MW, learn the operations, then add capacity on the same land with the same connectivity. The caveat is that the evacuation and switchyard have to be designed for the eventual capacity at the start, which costs a little more upfront and saves a great deal later. Retrofitting a 33 kV bay onto an 11 kV connection is expensive and slow.
What does a solar EPC business need instead?
Much less capital and a completely different skill set: execution capability, a working capital line, technical staff and a track record you can show. An EPC contracting business is a services business with project risk, not an asset business, and it earns on margin rather than on tariff. Many people who set out to own a plant discover the contracting business suits them better, and it is worth being honest about which one you actually want.
What is the most common reason a project fails?
Land and grid, in that order, and almost never the equipment. A title chain with a gap, a lease that is not assignable, a holdout owner in the middle of the layout, or a substation with no spare capacity. Each of these can be checked in a fortnight, for very little money, before anything is committed. The projects that fail are usually the ones where the checking happened after the advance was paid.