TD Power Systems

TD Power Systems makes generators. In a power plant, a turbine or an engine spins, and a generator converts that spinning into electricity. TDPS makes the generator. It does not make the turbine or the engine.

The ratio is fixed at one to one. One turbine takes one generator; one engine takes one generator. TDPS volumes are therefore decided by how many prime movers its customers build. This is the single most important fact about the company. It is a component supplier, and its growth is borrowed from its customers.

Those customers are a short list of large global OEMs — Siemens, INNIO (gas engines), Baker Hughes, Triveni Turbine, and a major US gas turbine maker. The top ten customers account for 60–70% of generator revenue. Exports and deemed exports were 80% of FY26 order inflow.

The product range:

Steam turbine generators are roughly 25% of turnover. The split between gas turbines and gas engines is not disclosed.

The balance sheet is clean. Net worth Rs 1,062 crore, effectively no term debt, interest cover of 47 times. A new plant commissioned in December 2025 cost about Rs 140 crore and was funded from internal accruals. CRISIL upgraded the rating to AA-/Stable in June 2026.

The turbine and engine makers are adding capacity.

AI data centres need firm power that runs continuously. Solar and wind cannot supply that alone, so the world is building gas plants. The companies that make gas turbines and gas engines are expanding to meet it.

Prime mover customers are, on average, planning to double capacity by 2030. INNIO has stated publicly that it intends to triple. TDPS holds a signed capacity commitment agreement with INNIO running to 2030.

Double the prime movers, double the generators. There is no design-around, and the barrier to new entrants in generators is high enough that supply cannot expand quickly from outside the existing set of manufacturers.

Demand is pulling, not being pushed.

The evidence is on the shop floor. TDPS is carrying no finished gas-side machines in inventory. Customers track delivery status twice a week and lift machines on completion. Slippage of three or four days is escalated by the customer. Execution, not order intake, is the binding constraint on revenue.

The mix has flipped to exports.

Exports were around half of sales five years ago. They are close to 70% now, and 80% of Q4 order inflow. Domestic capex has been weak through this period and TDPS grew regardless. The old TDPS was a bet on Indian cement, sugar and steel capex. The current one is a bet on global power generation.

Order inflow has exceeded Rs 600 crore for two consecutive quarters and is now running near Rs 650 crore a quarter. That annualises to roughly Rs 2,600 crore against FY26 revenue of Rs 1,869 crore. The order book is short — execution takes three to six months — so inflow converts to revenue quickly.

Large generators. TDPS put up the stator line for large machines in 2012–13 but held back on the rotor line when demand collapsed. That gap is now the constraint. The plan is to build out rotor manufacturing and buy large machining centres, taking the company up to 200 MW machines.

This is a different market: combined cycle plants, the largest data centres, and small modular reactors, where 200 MW covers most of what is currently classed as small. The technology sits in a wholly owned UK subsidiary, so no royalty leaves the company.

The timing is the issue. Machine tools now carry a 15 to 16 month lead time, because prime mover expansion has filled the order books of machine tool makers worldwide. Full capacity arrives in calendar 2027, and the ramp begins in calendar 2028 — FY29. Capex quantum, revenue potential and anchor customer are all due to be disclosed within two to three months.

Gross margin recovery. Q4 gross margin was 31.4% against a 33–34% norm. A Turkish contract slipped when a shipment of components from India sat in transit for six weeks, and the customer applied a heavy liquidated damages penalty. This shipping scope will not sit with TDPS in future contracts. Roughly 300 bps of gross margin should return.

A professional CEO. Deepak Kumar Sinha has joined as the company's first CEO, from L&T-MHI Power Turbine Generators, and before that GE Power. He sits as number two and runs all operations. Recruiting the executive who ran L&T's large generator joint venture, at the moment of entering large generators, is a coherent signal.

Motors and nuclear. The motor line is being physically separated to improve focus and execution. The NPCIL order is small but qualifies TDPS in a segment with a 100 GW target by 2047. BHEL, Crompton and imports compete here, and ordering is slow and lumpy; this is an optionality, not a plan.

The price. At Rs 1,090 the stock is at 71 times FY26 earnings. Everything below only matters because of this.

Profit is not converting to cash.

Receivables rose 70% against 45% revenue growth. Cash from operations was approximately 21% of EBITDA — around Rs 72 crore against Rs 343 crore of EBITDA and Rs 239 crore of reported profit.

The company's explanation is growth: the orders are short-cycle, inventory has to be built ahead of billing, and retained earnings are consequently funding working capital. For a business growing 45% that is a reasonable explanation, and it is the same one a business with a collection problem would offer. Disclosure on this point is thin — no ageing profile, no split between overdue and current, and no reconciliation of the year-end receivable balance to subsequent collections.

The mechanical consequence is worth sitting with. If FY27 revenue lands at Rs 2,400 crore and receivable days hold at 146, receivables become Rs 960 crore — another Rs 220 crore of cash absorbed against roughly Rs 300 crore of profit. Growth is being financed by the balance sheet, and the balance sheet is a principal reason to own this company.

The FY26 annual report cash flow statement and receivable ageing schedule are the test. Until then this is an open question, not a resolved one.

Customer concentration. The top ten customers are 60–70% of generator sales, with INNIO likely a large share of that. Volumes contracted under the INNIO agreement are not disclosed, nor is whether TDPS's share of that customer's spend is rising. Both are material to the growth case and neither is in the public record.

Copper. The raw material basket is copper, forgings, electrical steel, mild steel and insulation. Copper's share is not disclosed, which for a generator maker suggests it is large. Copper is at $14,000 a tonne, existing hedges are running out, and revised customer prices are only partly in effect.

The company's position is that it will be roughly neutral: a weaker rupee helps an exporter, contracts carry price variation clauses, and imported inputs offset export receipts. That holds for gradual moves. It does not hold for a sharp one — price clauses reset with a lag and material has to be bought today. Management's own stated threshold is a further 20–30% rise in copper, beyond which the response is undecided. A 200 bps margin hit on Rs 2,400 crore of revenue is Rs 48 crore of EBITDA and about Rs 35 crore of profit.

The factory is already full. Utilisation is the principal operating risk the company identifies for itself. At this level of loading, any equipment breakdown becomes a missed delivery. Q4 demonstrated what a single logistics failure does to a quarter. There is no slack in the system until calendar 2028.

This is a cyclical business. Profitability and returns on capital weakened significantly through FY15 to FY18 on low utilisation. Management's own capex conservatism is explained by that downcycle, which ran until 2019. The caution is a virtue in the operator and a risk to the shareholder: the good years are very good, and the bad years were bad enough to halve returns on capital. The price today contains only the good years.

Promoter holding is 26.9%. Institutions hold about half the company — FII 26.7%, mutual funds and AIFs 22.2%. This has not been a problem so far. It becomes one on the first miss.

Management has given two numbers and one boundary.

How reasonable is that?

FY27 looks close to safe. Order inflow is running at Rs 650 crore a quarter, the closing order book is Rs 1,973 crore, execution takes three to six months, and the plant is at full utilisation with customers lifting machines on completion. Rs 2,400 crore requires 28% revenue growth against 51% order inflow growth. The binding constraint is execution capacity, not order availability. A company that cannot keep a single finished machine in its yard is unlikely to miss a revenue number that its order book already covers.

FY28 is a harder number, but it is a capacity statement rather than a forecast, and capacity statements from this management have historically been conservative rather than promotional. The same conservatism that delayed the rotor line for a decade is what makes the Rs 3,200 crore figure worth taking at face value.

Margins are the soft spot. The 18.3% FY26 margin absorbs a Q4 gross margin hit of roughly 300 bps from Türkiye, which should reverse, against copper at $14,000 with hedges expiring. Holding margins flat is fair. Expanding them is not.

Put that guidance into the P&L.

Depreciation stepped up for Rs 100 crore of announced capex. Tax at 27%, in line with FY26.

What that means.

Guidance delivered in full, at the factory's physical ceiling, still leaves the stock at 41 times FY28 earnings. This is not a business that grows into its multiple by FY28. It grows into it, if at all, on what comes after — and what comes after is the large generator line, which produces its first meaningful revenue in FY29.

So the price is not making a claim about FY27 or FY28. It is making a claim about the durability of the multiple and about FY29 onwards.

The case for the multiple holding.

TDPS sits at a bottleneck. Prime mover capacity worldwide is doubling by 2030, the generator is a one-for-one requirement, and the barrier to a new generator supplier being qualified into a Siemens or INNIO programme is measured in years, not quarters. The bottleneck is not TDPS's factory — it is the global inability to add generator capacity fast enough to match the engines and turbines being built. That is a rare position, and markets pay for it.

The cycle length supports this. Data centre power demand in the US alone is a five-to-seven year installation programme. Prime mover expansion decisions are already committed capital at very large organisations, and those organisations do not stop mid-build. The INNIO capacity agreement runs to 2030. This is not a one-year order book dressed up as a trend.

And the large generator entry is real, not aspirational. The stator capacity already exists. The technology sits in-house. The CEO hired to run it built the same product at L&T-MHI. The constraint is a 16-month machine tool lead time, which is a supply problem, not a demand or capability problem.

The case against.

The multiple has to survive the arrival of the earnings. Component suppliers are usually re-rated on the way into a cycle and de-rated on the way out, and by FY28 the visible growth runway ends at the factory wall. The large generator business must then carry the entire valuation, before it has shipped a machine, in a market where Siemens, Baker Hughes, Mitsubishi and Toshiba are the incumbents.

The company's own history is the counterweight. Returns on capital collapsed through FY15 to FY18 on low utilisation. The current management's caution on capex is a direct inheritance of that period, and that caution is precisely why the capacity ceiling exists at Rs 3,200 crore.

Then there is the cash. FY26 converted about a fifth of EBITDA into operating cash. Growth is being funded off the balance sheet. A company earning Rs 300 crore of profit in FY27 while absorbing Rs 200 crore of incremental receivables is compounding book value more slowly than the P&L suggests, and the P/E is being paid on the P&L.

Where that leaves it.

FY28 is the best year the current factory can produce. Everything is sold, everything is built, nothing is left on the table. At Rs 1,090 you are paying 41 times the earnings of that year. So FY28 is not what justifies the price — the large generator business is. That is a coherent bet if the bottleneck is structural and the FY29 entry works. It is an expensive one if either assumption slips. The demand is real, the customers are real, and management has been honest. The valuation contains no margin for the second assumption failing.

Whether that is a price worth paying depends on how much weight you place on a business that has not yet been built.

What would strengthen the case:

Cash conversion. Receivable days back under 120 and operating cash above 60% of EBITDA in FY27. The reported profits then become real, and a higher multiple becomes defensible. This is visible within two quarters.

The large generator business plan. Capex quantum, revenue potential, anchor customer, target margins — all due within two to three months. If capex is modest against Rs 1,062 crore of net worth, incremental returns on capital are high, and there is a named customer with a signed commitment, then the FY29 leg stops being an assumption and becomes an asset.

Committed volumes under the INNIO agreement. The contract runs to 2030. Disclosure of contracted numbers rather than adjectives would materially narrow the cyclicality discount, because it converts a cycle call into a contract.

Guidance upgrades landing. FY27 above Rs 2,400 crore, and FY28 order book growth at or above the 20–25% indicated, would confirm that the bottleneck is holding and that TDPS is capturing it.

What would weaken it:

Receivable days above 145 for a second consecutive period, with the absolute number still climbing. Or anything in the FY26 annual report cash flow statement that fails to reconcile with the growth explanation.

Copper above $16,000 while the company's position remains that the situation is manageable. Watch gross margin, not EBITDA — operating leverage will flatter EBITDA while gross margin tells the truth.

Large generator capex funded with debt, or a capex number materially above Rs 400 to 500 crore. CRISIL's stated downgrade trigger is a large debt-funded capex. Levering the balance sheet at the top of a power cycle, to take on Siemens and Mitsubishi, is the setup that produced FY15 to FY18.

INNIO or the US gas turbine OEM qualifying additional suppliers. This appears as flat order inflow while customer capex keeps rising. It will be visible in the numbers before it is announced.

A miss on FY27. Guidance carries an explicit upgrade bias. If Rs 2,400 crore is not exceeded, something in the order book has broken, and at 71 times trailing there is no cushion.

TenetFour Research. Educational analysis only. Not investment advice or a recommendation to buy, sell, or hold any security. Author may hold a position. Readers are responsible for their own decisions.