Choices that cost growth or profit, with the bill left in

Accepting instalments for ₹1,710 crore of overdue utility dues

In June 2022, three months after our listing, two state distribution companies owed us ₹1,710 crore of bills more than sixty days overdue: ₹1,240 crore from the larger, which bought power from 520 megawatts of our older wind and solar plants, and ₹470 crore from the smaller. Receivables stood at 238 days of revenue. Our contracts also entitled us to late payment surcharge, by then ₹412 crore, which I had never recognised as income because I could not say when, or whether, it would be paid.

That month the Ministry of Power notified the Electricity (Late Payment Surcharge and Related Matters) Rules. They allowed utilities to clear accumulated dues in monthly instalments, stopped further surcharge accruing on the frozen balance while instalments were paid, and tied continued access to power markets to paying current bills on time. Three options reached our board. The general counsel and our chairman, a member of the founding family, wanted to keep pressing the full surcharge claim before the state commission. Aldermoor's nominee director proposed selling the receivables to a bank at a discount of about 11 per cent, which would have raised roughly ₹1,520 crore at once. I recommended accepting the instalment schedules both utilities offered and, with the larger one, settling the disputed part of the surcharge claim.

The cost was real. We waived ₹148 crore of surcharge in the settlement, gave up the immediate cash a sale of receivables offered, and accepted that the last instalments would arrive in 2026. The chairman agreed on condition that a single missed instalment would restart the litigation, which was the right condition.

Both utilities paid every instalment on schedule. Overdue dues fell to ₹180 crore by March 2026 and receivable days to 64. Because the surcharge had never been booked, the waiver touched no reported profit. The board now sees receivables by buyer every quarter, and our bid committee works within an exposure limit for each utility.

I would do one thing differently. I offered the smaller utility the same long schedule as the larger, although its record was far better and it would probably have cleared its dues within a year without any concession. The longer schedule cost us roughly ₹35 crore of surcharge that would otherwise have accrued. Treating the two alike was administratively tidy and commercially careless.

Pricing currency cover into a bid that we then lost

In early 2023 we prepared a bid for 600 megawatts of solar capacity in a central auction for inter-state projects. Modules made up about 38 per cent of project cost and were priced in dollars; the tariff would be fixed in rupees for twenty-five years. About sixteen months would separate the bid from the module payments. Our practice until then had been to bid at the spot exchange rate and decide on hedging after the power purchase agreement was signed.

I proposed that the bid include the full cost of forward cover for the module payments over that window. Treasury estimated it at about 2.5 per cent of module cost, or roughly two and a half paise per unit over the life of the tariff. Our chief business officer disagreed firmly. Competitors, he said, were bidding at spot and hedging later or not at all, and a rule that added even two paise would lose us auctions we could otherwise win. He was right about the arithmetic of the auction. My concern was that a winning bid's equity return would then rest on sixteen months of currency movement that nobody could predict.

The Risk Management Committee supported the policy. We bid ₹2.62 per unit and lost by three paise to a bid of ₹2.59. Our project teams had spent five months on that bid, and some of the best of them were openly unhappy.

The aftermath was not a vindication. Over the following sixteen months the rupee moved less than the cover would have cost, and in hindsight a bid at spot would have done better in that auction. I told the board so. The decision stood for a different reason: it kept our equity return from depending on a currency forecast. The board made it a standing rule, and in 2024 we won two auctions totalling 900 megawatts on those terms, mostly with domestic modules bought in rupees. When the rupee fell by about 4 per cent in a few months from late 2024, those projects were unaffected.

What I would change is how I argued it. I presented the policy as a risk rule instead of showing the business development team, project by project, how much of their expected return a 5 per cent currency move would remove. When I later did show them, most of the resistance ended.

Capping new bids at about one gigawatt a year

By the spring of 2023 our installed capacity had reached 3.1 gigawatts and net debt was 7.4 times EBITDA. Our chairman had announced an ambition of ten gigawatts by 2030, which implied winning and building about two gigawatts a year. At that pace our equity needs would have exceeded operating cash flow by about ₹1,100 crore a year. Our ratings had sat at A+ for three years.

I proposed two things to the board. The first was a written ceiling: no new commitment unless net debt to EBITDA, projected through the peak of construction, stayed at or below seven times, which in practice meant about one gigawatt of new bids a year. The second was to sell 49 per cent of a 900-megawatt operating solar cluster to a long-term investor, keeping control, and to use the proceeds as equity for new projects. The chairman and our CEO opposed the ceiling, arguing that auction volumes would not stay high for ever and that the market rewarded visible growth. Two of our four independent directors agreed with them. The Audit Committee chair supported me, and after two meetings the board approved both proposals, with a review after twenty-four months.

We gave up a good deal. We stayed out of auctions for perhaps 1.5 gigawatts over two years that we could probably have won, and our share price trailed a sector index by about 18 per cent in 2024–25. Corvanne Pension Investments paid ₹1,640 crore for the minority stake, a price some directors thought low.

Operating capacity still grew to 4.2 gigawatts by March 2026, and net debt fell to 6.5 times EBITDA. Two domestic rating agencies raised us to AA- in 2024, and our weighted average borrowing cost fell from 9.4 to 8.4 per cent between March 2023 and March 2026, worth about ₹200 crore a year in interest at present debt levels. The ceiling is now part of the budget the board approves each year, not an argument we repeat every auction season.

I would have proposed the ceiling a year earlier, when nothing was at stake, so that it did not read as a verdict on the chairman's ambition. I also misjudged how our project teams would hear it: as a statement that their work mattered less than a ratio.

Leaving a change-in-law claim out of our pre-listing profits

In October 2021 the GST rate on solar generating equipment rose from 5 to 12 per cent. We had 1.1 gigawatts under construction for central intermediaries, and the change added ₹236 crore to the cost of those plants. Our power purchase agreements provided compensation for changes in law, and we filed petitions with the Central Electricity Regulatory Commission. We were also five months from listing.

Our CEO, our head of investor relations and the bankers managing the listing wanted the claim recognised in the 2021–22 accounts as income and a receivable, on the ground that compensation for a tax change was all but certain. I did not agree. Under Ind AS 37 an asset that depends on a future decision is disclosed, not recognised, until its realisation is virtually certain, and we did not know how the commission would compute compensation, over what period, or whether the buyers would dispute the figure. I recommended disclosing the claim in the offer document as a contingent asset, with its amount and the process ahead. The statutory auditors agreed. The Audit Committee accepted my recommendation, and one member asked me to put my reasoning in writing, which I did.

Our 2021–22 profit was ₹236 crore lower than the alternative would have shown. The issue was priced in the lower half of its band. The bankers believed the claim was one reason, and I cannot prove they were wrong.

The commission ruled in 2023. It allowed compensation worth ₹204 crore in present-value terms, payable as a monthly annuity over thirteen years at a rate it set rather than as a lump sum, and it disallowed some of the costs we had claimed. Had we booked ₹236 crore in 2021–22, we would have reversed ₹32 crore in our second year as a listed company. We recognised the award once the buyers accepted the computation, and it is now collected with our monthly bills.

The policy that followed is simple. Claims for change in law, curtailment and late payment are disclosed every quarter with their amounts and status, and recognised only on an order and an accepted computation, or on receipt.

I would do one thing differently: write that policy and take it to the Audit Committee before any claim arose, so that the argument concerned a principle, not a figure in an offer document.