All right, so a few weeks back, we were all circumspect when Sun Edison bid at Rs 4.63/kWh for 350 MW solar power plant in Andhra Pradesh under the National Solar Mission.

How could this be? – we asked among ourselves. Perhaps Sun Edison is quoting an unsustainable price?, wondered many; while a humbler few said – perhaps Sun Edison knows something we don’t know?

A few wise eggheads had all along pointed out that many other bids for that allotment were not far from Sun Edison’s.

Now, all these are coming to a head once again – and this time it is SoftBank in the place of Sun Edison, the allotment is once again for 350 MW, once again for AP under National Solar Mission and the bid is once again, Rs 4.63/kWh -not a paise more, not a paise less.

Sounds Jeffrey Archer-ish?

What is so magical about Rs 4.63/kWh for solar? We analyse the 4.63 phenomenon in this post.

Firstly, what indeed does Rs 4.63 represent?

Rs 4.63 represents the tariff that Softbank will receive for every unit they generate and export to the grid.

Well then, this is not exactly the cost of solar power generation, but this rather is the lowest price at which Softbank feels it can sell the power and still make a decent profit.

So, what will be the cost of solar power for Softbank?

In order to answer this question, we will have to first figure out what could have been the inputs for Softbank for them to get a decent returns on a 4.63 tariff.

For this, we took our detailing financial calculations spreadsheet and did a reverse calculation.

The question we asked ourselves was:

With a tariff of Rs 4.63/kWh, what should be the inputs so that Softbank gets a decent IRR – 15% equity and 11% project IRR?

Well, the following is one such scenario under which they indeed can get such an IRR at Rs 4.63

  • Capital cost: Rs 5.3 crores/MW (excluding costs for approvals, interest cost during construction and working capital for 1st two months)
  • Interest: 8% tenure of 12 years with one year moratorium
  • CUF: 19%
  • O&M costs: 5 lacs/annum, 3% escalation
  • Grid uptime: 100% (this means offtaker takes all the power generated by the plant)
  • Panel degradation: 1% per annum for the entire project lifetime

Normal depreciation and taxes (no accelerated stuff). Take other numbers according to convention (insurance etc).

  • Plug in the above numbers in a well prepared financial spreadsheet, and what do you get?
  • Equity IRR: 14.7%
  • Project IRR: 10.8%

Sure, you might get slightly different results for the above inputs as no two financial guys will (as will no two economists) agree on everything, but yours should not be too different from ours.

With that settled, three questions arise.

Q ONE: Is the above the only input combination that can give these equity and project returns for a Rs 4.63/kWh tariff?

Q TWO: Is the above input combination feasible at all under the current circumstances?

Q THREE: What is the LCOE for such an input combination?


Q ONE: Is the above the only input combination that can give these equity and project returns for a Rs 4.63/kWh tariff?

Well, in theory, no.

Anyone who likes to play with spreadsheets can come up with, in theory, a number of other input combinations (mainly with varying capital costs, CUFs and interest rates) and come to the same equity and project IRRs.

But in practice, we feel what we have provided is one of the very few combinations that will make sense in the market.

You will not accept an interest rate of 4% or CUF of 24% for a non-tracker system, surely?

Now, while on trackers, we had assumed that the power plant operates without trackers. With trackers, for a similar combination, the capex could go up by about 25 lacs/MW, opex by 2 lacs/MW annum and CUF could be as high as 21%. Such a combo will again give a similar project & equity IRR, for other inputs such as interest rate remaining the same. (11.1% pIRR, 15.6 Equity IRR).

So, let us say what we have provided above are some of the few plausible scenarios with and without trackers.

Q TWO: Is the above input combination sustainable under the current circumstances?

Well, let us look at some of the thorny numbers presented above:

  1. Is 8% interest rate possible? If it is from Indian banks, forget it – it is a big NO. But it is a Yes, if it is international money – which almost certainly will be the case for Softbank
  2. Is 5.3 crores/MW (without trackers) possible? Yes, just about, if land costs are taken care of by the government. For one of the recent 2 MW projects we worked on, the total cost without trackers came to Rs 5.75 crores/ MW (excluding approval costs, interest during construction etc). In this case as well, we did not include land cost as the land was already owned by our client. Now, is it possible to get 5.75 down to 5.3 when the scale goes from 2 MW to 350 MW? It just could be, don’t you think so?

The rest of the inputs are fairly acceptable (19% CUF, 1% degradation, O&M costs…). One might debate the 100% grid availability assumption as being too naive, but at this scale, the contract could have involved the offtaker agreeing to such a clause.

So overall, the input combination is just about sustainable!

Q THREE: What is the LCOE for such an input combination?

Well, things get interesting here, because we are all interested in knowing what exactly the blooming solar power costs.

For this, we took the above inputs that gave an acceptable IRR and plugged these into our LCOE calculator.

And the result was: tick, tick, tick – Rs 4.1/kWh

So, with Rs 4.63/kWh being the tariff, for a reasonable return (and we expect the folks like Nikesh Arora at Softbank to be reasonable when it comes to financial returns), the LCOE of solar power at these kinds of scale is in the vicinity of Rs 4.1/kWh.

The LCOE number comes out almost identical for both projects with and without trackers, if we were to go with the above inputs.


  1. Just to emphasise once more, we have not considered accelerated depreciation, just simple, civilized, WDV based depreciation
  2. On trackers, we have considered single axis trackers, pretty much the only type of trackers used for solar power plants in south India
  3. While calculating IRRs, we have considered the preliminary expenses etc., as per the accepted norms
  4. We have considered an inverter replacement once every 10 years


Hope the above stuff was of help. You think there is anything inaccurate in my assumptions or inputs? Or think I have missed something? Oryou just have a simple suggestion for improving this post? I will be happy to hear your thoughts on the above – do drop in your comment below or send a note to . Thank you!