CG Power — a phoenix priced like a prophet
CG Power and Industrial Solutions Ltd
Snapshot
CG Power makes the unglamorous iron-and-copper guts of the electricity system — the transformers and switchgear that move power around the grid, and the electric motors and drives that turn that power into work inside factories. It is now controlled by the Murugappa Group (via Tube Investments), one of India’s most respected old-money houses, which bought it out of a near-death fraud in 2020 and rebuilt it. Market cap ₹1,51,781 cr, price ₹964 (52-week range ₹526–971, so it sits near its all-time high), trading at a P/E of 123 and a price-to-book of 19×. Return on equity 20.8%, return on capital 27.0% — and it is essentially debt-free. What kind of animal is it? A high-quality, turbo-charged capital-goods compounder riding a once-in-a-generation power boom — wearing a price tag built for a software company. As of 2026-06-20, from screener snapshot.
The verdict in two boxes — the business first, the price second
Keep them apart on purpose. Box 1 describes the company you’d own; it would read the same if the share price doubled or halved tomorrow. Box 2 describes only what Mr. Market is charging for it today.
Box 1 — The business (durable):
| Lens | Result |
|---|---|
| Business-quality score | 18 / 23 (Quality 8.5/12 · Growth 5.5/6 · Longevity 4/5) |
| Buffett rubric | 6 / 10 PASS |
| Business bucket | Good, edging toward Great (high returns, but turning capital-hungry) |
| Wealth-creator type | Enduring (structural moat + clean owner) · becoming Consistent (post-2020 record only ~5 years clean) |
| Economic Profit | +₹701 cr (RoE 20.8% − CoE 12% on ₹7,970 cr net worth) — clearly creating value |
A Good business, edging toward Great, that genuinely creates wealth — run by people you’d trust with your money — independent of what it costs today.
Box 2 — The price today (a current phenomenon):
| Reading | Result |
|---|---|
| CMP | ₹964 (as of 2026-06-20) |
| Price pillar | 0 / 2 (PEG ≈ 4–6× · 5-yr payback ≈ 12–14×) |
| Margin-of-safety band | ₹250–₹400 (where PEG nears 1× / payback nears 1× on ~25% growth) |
| Mr. Market’s mood now | Greedy — a power-capex supercycle + a “first Indian chip plant” story have pulled in euphoric ownership |
| CMP vs the band | Very demanding (≈ 2.5–3.5× the band) |
Today the market prices it rich — a mood driven by a real power boom and a glamorous semiconductor dream — and that mood can deflate while the business above carries on unchanged.
In plain English
Picture the electricity that runs your country as water in a vast plumbing system. CG Power makes two of the most important fittings in that system. First, transformers and switchgear — the big boxes at substations that step voltage up and down and safely switch the flow on and off. Second, the electric motors and drives that sit inside factories and turn that electricity into spinning shafts that grind cement, pump water, and run machines. It’s a boring, essential, century-old business. Boring and essential is usually a good place to make money.
For most of the last decade this was a sad story. Under its previous owner, the Avantha Group’s Gautam Thapar, the company was looted — regulators later found over ₹1,600 crore of hidden liabilities and money quietly funneled out to the promoter’s other ventures. The shares nearly went to zero. Then in 2020 the Murugappa Group — a 125-year-old Chennai family famous for doing things slowly, honestly, and with very little debt — stepped in, paid off the borrowings, cleaned the books, and put in real managers. The turnaround has been spectacular: revenue has more than doubled, the company is debt-free, and it now earns 27 paise of operating profit on every rupee of capital. That is a very good return.
What’s powering it now is a genuine boom, not a fad. India and much of the world are spending enormous sums upgrading their electricity grids — for renewables, for electric vehicles, and most of all for the data centres that artificial intelligence is hungry for. All of those need transformers, and the world is short of them. CG just won its largest-ever order — ₹900 crore of transformers for a US data centre — and its order backlog is up 59% to over ₹17,000 crore. It has roughly tripled its transformer-making capacity in two years and still hears customers say “give us more.” The moat is real but young: it’s a trusted brand, huge factories that take years and approvals to replicate, and a ~38% share of low-tension industrial motors. The moat is widening as it adds capacity faster than rivals.
So here is the tension, and it’s the whole report in one line: this is a good — arguably becoming great — business, run by trustworthy owners, that the stock market is pricing as if nothing can ever go wrong. At 123 times earnings and 19 times book value, you are paying today for a decade of flawless growth in advance. The business deserves a premium. It does not obviously deserve this premium. And quietly, in the background, the same management is making one very large, very different bet — a ₹7,600 crore semiconductor packaging plant — that is far outside its comfort zone and is the one thing that could turn a brilliant capital allocator into a careless one. Watch that, not the order book.
Sitting down with the management
If Buffett and Raamdeo Agrawal sat across from these people, they’d relax within five minutes — and then lean forward on one question.
The promoter is the Murugappa Group, through Tube Investments. This matters enormously, because the single most important fact about CG Power is who owns it now versus who owned it before. The old owner committed fraud; the new one is the opposite kind of animal. Murugappa is a fourth-generation, ~125-year-old conglomerate whose codified values literally put “integrity above all else,” and whose finance jewel, Cholamandalam, compounded its market value some 60× under the man who now chairs CG — Vellayan Subbiah [SOFT — EY, 7 Jun 2024]. Subbiah is an IIT-Madras engineer and Michigan MBA who was named EY World Entrepreneur of the Year 2024, only the fourth Indian ever to win it [HARD/SOFT — EY, 7 Jun 2024]. His method is a “platform” approach: take a controlled base, reinvest with discipline, and bolt on smart acquisitions. He turned CG from debt-laden wreck to profit machine inside roughly a year [SOFT — Business Standard, 7 Jun 2024]. The day-to-day is run by Amar Kaul (MD & CEO since 25 Jul 2024), a 30-year operations man from Ingersoll Rand and Bharat Forge [HARD — Business Standard, 8 Jul 2024], and CFO Susheel Todi.
How have they spent the owners’ money? Mostly well. They wiped out the debt and got to net-cash by March 2022. They’ve poured capital into transformer and switchgear capacity at exactly the moment demand exploded — capacity up from ~17,000 MVA to ~110,000 MVA (planned) in two years — and the returns prove it was the right call (Power Systems margins jumped from 19% to ~22% as those plants filled). The bolt-ons are sensible adjacencies: G.G.Tronics (railway signalling, 55% for ₹319 cr) sits next to their existing rail business. And the discipline on the big ₹3,000 cr equity raise (QIP) in July 2025 is exactly what you want to see — independent monitoring confirms zero deviation in how the money was used through March 2026 [HARD — ScanX, 2026]. That single fact — audited proof that fresh capital went where they said it would — is the loudest possible “we’re not the old owners” signal.
Do they talk straight? Yes. Read the concalls and you hear a CEO who answers “what’s your market share?” with “38–39% in LT motors, ~19–20% in larger ones,” who admits margin pressure in railways and motors plainly, and who declines to over-disclose (“I’m not allowed to give exact IE3/IE4 numbers”) rather than spin a number. There’s no chronic guidance-and-miss pattern. Promoter holding is high (~56%) and zero pledged — real skin in the game, real alignment [SOFT — Trendlyne, 2026].
Now the lean-forward question. There are two honest concerns. First, key-man and family risk: CG’s fate is tied to one allocator, Subbiah, and to Murugappa’s own family dynamics — in Feb 2026, reports of a family settlement said he’d step back from Cholamandalam to focus on TII and CG (Chola denied any exit) [SOFT — Outlook Business, 26 Feb 2026]. Second, and bigger, the semiconductor bet (below). Would Buffett and Agrawal shake hands on this management? Yes — with eyes open. The one thing that would change their mind: if the ₹7,600 cr chip plant drifts from “calculated adjacency” into a capital-hungry, subsidy-dependent moonshot that they keep feeding with shareholders’ money. The very discipline that earns the handshake is the thing that bet could break.
What’s on the horizon (live-issues tracker)
1. The semiconductor (OSAT) plant — THE CRUX. 🟡 early, unproven.
The crux in one sentence: this investment compounds cleanly if and only if the ₹7,600 cr semiconductor bet either works at a decent return — or is kept small enough not to poison the capital-allocation record that justifies the whole premium.
The plain-English mechanism (with an analogy). “OSAT” stands for outsourced semiconductor assembly and test. It is not chip-making (no fancy lithography). It’s the back end: you take finished silicon wafers and package them into the little black chips that go on circuit boards, then test them. Think of it as the bottling-and-labelling plant at the end of a drinks factory — essential, high-volume, but a commodity service business. You earn a fee per chip, you compete on yield, cost, and reliability, and you must run the line near-full to make money. That is a very different animal from CG’s core. Test the analogy: is OSAT really “just another factory” CG can run with its manufacturing DNA? Partly — CG is genuinely good at running high-volume plants. But the demand side is alien: customers are global chip designers, the technology shifts yearly, yields take time to ramp, and you’re a price-taker on a global cost curve. So the analogy holds for the operations and breaks for the economics and the customer. The manufacturing risk is modest; the “will the orders and the yields show up at a good margin” risk is real.
The named-competitor / named-threat map:
| Who | Backed by | Posture | Proof point |
|---|---|---|---|
| Kaynes Semicon | Kaynes Technology (listed) + ISM subsidy | Sanand OSAT, same scheme | Already started production at Sanand — India’s second chip unit [SOFT — CRN Asia, 2026] |
| Tata Electronics | Tata Group (deep pockets) | Assam OSAT + Dholera fab | Far larger balance sheet; strategic national champion |
| Micron | US giant + ISM | Sanand ATMP | Global incumbent; captive volumes |
| Global OSATs (ASE, Amkor) | Decades of scale | The world cost curve | They define the price CG must beat |
The threat is not that CG can’t build the plant — it’s that OSAT is a global commodity where India is a high-cost late entrant, and CG will be competing for the same customers and the same subsidy pool as Tata and Kaynes. The damage, if it comes, is to return on capital (a low-margin business soaking up ₹7,600 cr), not to CG’s grid business.
The real-world precedent. Every OSAT-building nation — Taiwan, Malaysia, Vietnam, China — got there with years of losses, government money, and patient scale-up before the back-end turned cash-positive. The incumbents that won (ASE, Amkor) did it over decades on enormous volume. The honest read of precedent: OSAT is a real, durable business — but a thin-margin, scale-and-patience one, and first-generation national entrants rarely earn their cost of capital for years.
Answered follow-on questions. Is the bet de-risked? Substantially — ~₹3,500 cr of central ISM subsidy + ~₹1,400 cr state support cover roughly two-thirds of the cost, and CG owns ~92% (so it controls it, with Renesas/Stars as real partners). How big is the downside to CG? Manageable: even a total write-down of CG’s own share is a fraction of its ₹1.5 lakh-cr market cap and is dwarfed by the grid order book. Is anyone actually buying chips yet? Not really — chip revenue is “a few quarters away” (the ₹500 cr “semi” revenue today is the acquired Renesas RF business, not packaged-chip output). The G1 plant (0.5m units/day) opened Aug 2025; G2 (14.5m/day) is due end-2026. Honest verdict: not “too hard” — but genuinely unproven. The subsidy and the small relative size keep the downside contained; the upside and the timing are unknowable. Treat the semiconductor story as a free option, not as something you should pay 123× earnings for.
2. The power & transformer supercycle — 🟢 firing on all cylinders. Power Systems sales grew 46% in FY26 with margins expanding ~280 bps; the order backlog is up 91%. Management calls it “Amrit Kaal for the power sector.” Capacity tripled and is still being added. Data-centre and US export demand is a new, large leg. This is the engine and it is running hot — the only worry is that the base is now high, so the eye-popping growth rates will naturally cool.
3. Motors & railways margin repair — 🟡 mixed, improving. The industrial side lagged: motors took commodity-cost hits, railways is structurally low-margin. Management has pushed through ~17.5% of price increases over a year while holding ~38% market share, and is leaning on cost programs and a services push. Trajectory is up but slow.
The watch-list (what tells you the thesis is working or breaking):
- Semi: real packaged-chip revenue and segment margin turning positive (vs today’s ~₹110 cr drag) — milestone, not yet hit.
- Semi discipline: any new equity raise to fund chips = warning; self-funded ramp = reassuring.
- Power orders: backlog holding/growing above ₹17,000 cr and US grid (not just data-centre) approvals landing.
- Margins: Power Systems holding ~22%; Industrial PBIT margin climbing back toward double digits.
- Working capital: WC days have jumped 35→82 — watch they don’t keep climbing as the order book grows.
QGLP scorecard (the Motilal Oswal lens) — the receipts
| # | Question | Score | Evidence |
|---|---|---|---|
| 1 | Large opportunity? | 1 | Grid capex supercycle, data-centre transformers, motors, semis — vast and growing 10y+ |
| 2 | Industry structured favourably? | 0.5 | Transformers/switchgear a consolidated oligopoly (stable margins); motors/railways competitive |
| 3 | Clear, defensible moat? | 0.5 | Brand + scale + approvals + ~38% motor share, but high-RoE record only ~5 yrs old (post-2020) |
| 4 | Return ratios > 15%? | 1 | RoE 20.8%, RoCE 27.0%; RoCE 42–61% in FY22–24, now 27% (snapshot ratios) |
| 5 | Asset-light / low capital intensity? | 0.5 | Core is reasonably capital-efficient, but capex is now heavy (₹647 cr CWIP, ₹3,605 cr investing outflow FY26) |
| 6 | Terms of trade favourable? | 0.5 | Debtor 86d vs Payable 105d ≈ 82% (slightly favourable), but WC days 35→82 rising |
| 7 | Unquestionable integrity? | 1 | Murugappa owner; 0% pledge; audited zero-deviation QIP usage (the prior fraud predates this owner) |
| 8 | Proven execution? | 1 | Fraud-to-debt-free-to-₹1,199 cr PAT turnaround; 21% FY26 sales growth delivered |
| 9 | Growth mindset & vision? | 1 | Capacity 3×, exports, semis, railways — clear reinvestment ambition |
| 10 | Superior capital allocation? | 0.5 | Core bets excellent; but large unproven semi bet + QIP dilution temper this |
| 11 | Clear succession plan? | 0.5 | Strong pro bench (Kaul/Todi), but strategy rests on one allocator (Subbiah) + family dynamics |
| 12 | Minority interests protected? | 1 | Clean QIP, dividends restored (17% payout), no RPT leakage under TII |
| 13 | Structural tailwind? | 1 | Power capex >1.5× GDP growth; ₹9.2 lakh-cr grid plan to 2032 |
| 14 | Volume-led growth? | 0.5 | Power is volume/execution-led; motors ~50% price-led this year |
| 15 | Operating leverage? | 1 | Power Systems OPM 19%→22% as sales rose 46% — textbook leverage |
| 16 | Manageable leverage? | 1 | Net-cash; borrowings ₹118 cr vs ₹7,970 cr net worth |
| 17 | Market-share gain potential? | 1 | Adding transformer capacity faster than rivals; ~38% motor share held through price hikes |
| 18 | Earnings growth > 15%? | 1 | Standalone PBT-ex-EI +34% FY26; consol PAT-before-EI +27%; sales 3-yr CAGR ~21% |
| 19 | Relevant for 10–15 years? | 1 | Electricity infrastructure is an “Inevitable” — demand won’t disappear |
| 20 | Extend its CAP? | 0.5 | Moat widening via capacity, but young; commodity-cost exposure caps pricing power in parts |
| 21 | Sustain its GAP? | 1 | Long runway — grid, EV, data-centre, exports all early |
| 22 | Diversification headroom? | 1 | Exports (just doubling off a ~5% base), semis, new geographies |
| 23 | Adaptive, resilient culture? | 0.5 | Survived a near-death fraud and rebuilt — but culture noted as traditional/“old bossy” (Glassdoor) |
| Business quality (Q1–Q23) | 18 / 23 | Quality 8.5 · Growth 5.5 · Longevity 4 | |
| 24 | Valuation reasonable (PEG)? | 0 | P/E 123 ÷ ~27% growth ≈ PEG 4.5× |
| 25 | Margin of safety (payback/PEG <1×)? | 0 | 5-yr payback ≈ 12–14× even at 25% growth |
| Price pillar (Q24–Q25) | 0 / 2 | Priced for perfection | |
| Canonical QGLP total | 18 / 25 | (headline is the 18/23 business score) |
The pillar pattern: Quality is solid and Growth is the standout — this is a business firing on the things the studies reward (returns, leverage, a structural tailwind, a clean owner). The only thing standing between this and a clear wealth-creator buy-zone is Price, which scores a flat zero. Quality and Growth carry the day; Price is the entire argument against owning it today.
Buffett lens (the Berkshire-letters read)
| # | Test | Verdict | Evidence / Buffett line |
|---|---|---|---|
| 1 | Good boat? | PARTIAL | Good (high RoCE) but turning capital-hungry — “what business boat you get into…“ |
| 2 | Moat + pricing power | PARTIAL | Real brand/scale; power has price-variation clauses, but motors are commodity-cost exposed |
| 3 | See’s test (high returns, little capital) | PARTIAL | Earns well, but growth now eats cash (FCF −₹72 cr FY26 on heavy capex) |
| 4 | One-dollar test (capital allocation) | PASS | Retained rupees rebuilt book value 6× and market value many times; clean QIP use |
| 5 | Owner-oriented, candid | PASS | Straight answers, admits margin misses, high unpledged promoter stake |
| 6 | Integrity / forensic | PASS | OCF ≈ PAT over the cycle; profit becomes cash; no balance-sheet games under TII |
| 7 | Circle of competence / predictability | PARTIAL | Grid business is predictable; semiconductors are technology-fragile and outside the circle |
| 8 | Mr. Market — gift or trap? | FAIL | Priced for perfection: P/E 123, P/B 19×, near all-time high, euphoric ownership |
| 9 | Patience / compounding runway | PASS | Long, high-RoE reinvestment runway in grid/exports |
| 10 | The honest red flag | (see prose) | The price itself is the bear case |
Count: ~6 / 10 PASS (PARTIALs at ½). A real business with real gaps — the gaps are predictability (test 7, the chip bet) and price (test 8).
The See’s test, spelled out. See’s Candies was Buffett’s ideal because it threw off cash without needing much back. CG was trending that way in FY22–24 (RoCE 42–61%, fat free cash flow). But FY26 tells the other story: free cash flow turned slightly negative (−₹72 cr) because the company is plowing money into transformer plants and the chip facility. That’s the difference between a “Great” cash fountain and a “Good” compounder that must keep feeding the machine to grow. It earns high returns — but right now it’s reinvesting, not distributing. Fine while the reinvestment earns 27%; the watch is whether the semi spend dilutes that.
The one-dollar test, spelled out. Have retained earnings created at least a rupee of value each? Emphatically yes. Net worth went from negative reserves (−₹2,081 cr in FY20) to +₹7,655 cr of reserves in FY26, while the company stayed debt-free and lifted profit from a ₹1,331 cr loss to a ₹1,199 cr profit. Every retained rupee has multiplied. This is the test CG passes most clearly, and it’s why the management section is warm despite the price section being cold.
The framework metrics
- Economic Profit = Net Worth ₹7,970 cr × (RoE 20.8% − CoE 12%) = +₹701 cr. Genuine value creation above the cost of owners’ money — top-tier. (CoE 12%, the Indian mid-point the studies use.)
- Terms of Trade = Debtor days 86 ÷ Payable days 105 ≈ 82% — modestly favourable (suppliers part-fund it), but deteriorating as the order book grows.
- 5-yr Payback = Mcap ₹1,51,781 cr ÷ projected 5-yr cumulative PAT ≈ 12–14× (at 20–25% PAT growth off ₹1,199 cr). Far above the 1× multi-bagger line.
- PEG = P/E 123 ÷ ~27% growth ≈ 4.5×. Far above the 1× discipline line.
- RoE − CoE spread = ~8.8 points, and strongly positive; RoE > 15% in roughly 5 of the last 10 years (the earlier years were the loss-making Avantha era). So the uncommon profit is real but young.
- Consistent vs Volatile: the 15-year PAT row has huge swings (₹-1,331 cr in FY20 to +₹1,428 cr in FY24) — that’s a Volatile history on paper. But the volatility is the old company dying and a new one being born. The post-2020 record (FY21→FY26) is steady and rising. So: Volatile on the old book, Consistent on the new — value it on earnings power, but respect that the clean track record is only ~5 years long.
Peer comparison
| Company | Mkt cap (₹cr) | CMP (₹) | P/E | P/B | RoE | RoCE | OPM | FY26 sales (₹cr) |
|---|---|---|---|---|---|---|---|---|
| CG Power | 1,51,781 | 964 | 123 | 19.0 | 20.8% | 27.0% | 13% | 12,418 |
| Hitachi Energy India | 1,64,383 | 36,880 | 160 | 31.8 | 21.9% | 29.0% | 15% | 8,148 |
| ABB India | 1,53,591 | 7,248 | 101 | 19.6 | 22.4% | 29.9% | 14% | 13,093 |
| Siemens India | 1,33,887 | 3,760 | 56 | 9.7 | 16.3% | 21.2% | 11% | 24,846 |
| Transformers & Rectifiers (TARIL) | 10,732 | 358 | 41 | 7.1 | 19.1% | 23.3% | 15% | 2,509 |
What it says. CG Power is expensive on its own — but it is not the most expensive in its asset class. Hitachi Energy India trades richer on every metric (160× P/E, 32× book), and ABB is in the same zip code. The entire Indian power-equipment shelf has been bid up on the same supercycle story. So a sector allocator could fairly call CG “mid-pack, not the dearest.” A patient value investor sees a whole shelf priced for perfection and notes CG’s distinctive edge: it’s the debt-free, founder-respected, fastest-capacity-adding domestic champion with optionality (semis, exports) the MNCs don’t price the same way. The relative read (reasonable-for-the-shelf) and the absolute read (very expensive) genuinely disagree — because they answer different questions: “cheapest transformer stock?” vs “is any transformer stock cheap?” Right now the honest answer to the second is no.
Latest quarter & what’s happening now
Q4 FY26 (reported 6 May 2026): consolidated sales ₹3,442 cr (+25% YoY), PAT ₹362 cr (+32%). A record fiscal year — highest-ever standalone revenue, PBT, and order book. Order backlog ₹17,107 cr, up 61%. Two concall takeaways: (1) Power is on fire — “the game has just started,” capacity headed to ~110,000 MVA, the ₹900 cr US data-centre order is the largest in CG’s history and management implies more US/grid orders behind it [HARD/MEDIUM — concall 6 May 2026]; (2) the chip plant is still pre-revenue — packaged-chip output is a few quarters out, today’s semi line is a ~₹110 cr drag on margins from talent build-out [HARD]. Interim dividend ₹1.30/share declared. Live catalysts: G2 chip plant commissioning by end-CY2026 [MEDIUM]; 400 kV GIS commercialisation in FY27 [MEDIUM]; greenfield switchgear plant (₹748 cr) [SOFT — board-approved 29 Oct 2025].
Where the two lenses agree — and disagree
They agree on the spine: high returns (Q4/test 4 PASS), a structural tailwind and long runway (Q13/Q21/test 9 PASS), a clean owner-operator (Q7/test 5–6 PASS), and a price that fails the discipline test (Q24–25 zero / test 8 FAIL). Both say: Good-toward-Great business, wonderful owner, unwonderful price.
Where they diverge — and it’s the signal: the QGLP checklist happily scores Growth 5.5/6 and barely dings the semiconductor adventure (it reads as “growth mindset, vision”). Buffett’s test 7 (circle of competence) flags it hard — semiconductors are exactly the “lots of technology, we won’t understand it” business he avoids. The checklist sees ambition; the letters see a predictability risk. Trust the Buffett flag here: the chip plant is the one thing that could break the capital-allocation record the whole premium rests on. The numbers can’t yet see it because it isn’t earning (or losing much) yet.
The price as a current phenomenon
This section judges the price, not the business. The business verdict above is already settled.
The margin-of-safety band. The framework’s arithmetic is brutal here. To satisfy QGLP’s Price pillar — PEG near 1× and a 5-yr payback heading toward 1× — even on a generous 25% earnings-growth assumption, you’d need a price roughly in the ₹250–₹400 range (a P/E in the 30s–40s, like the cheaper end of the peer shelf, e.g. TARIL at 41×). At ₹964 the stock is ~2.5–3.5× that band. This is not a forecast that it will fall there; it’s the arithmetic of where the price discipline the studies demand would be satisfied.
Mr. Market’s mood. He is greedy on this name, and you can see exactly why. A real, multi-year power-capex supercycle (the “Amrit Kaal for power”) + a glamorous “India’s-first-chip-plant” narrative + a celebrated owner + a near-all-time-high chart = euphoric ownership. None of that is fake. But greed prices in a decade of perfect execution today. Note that FIIs have been trimming (16% → 12% over two years) while domestic institutions pile in (11% → 18%) — a tug-of-war that often marks a fully-valued name.
The plain tension. A wonderful business can sit at an unwonderful price, and that is precisely this case: the boat is good and getting better; the seat is being sold at a scalper’s markup. The reverse — a gruesome business at a bargain — does not apply here at all. And remember: this entire price reading can change next week — a market wobble, an FII exodus, a chip-plant disappointment — without a single transformer order being cancelled. The business box would not move. The price box would.
Conviction texture
The bull case, strongest form: You’re buying the debt-free, best-governed, fastest-growing domestic champion of the iron backbone of electrification — at the exact moment grids, EVs, exports and AI data centres all demand more of what it makes, with a respected owner who has already proven he can allocate capital and a free option on Indian semiconductors on top. Quality compounders in a supercycle rarely look cheap; pay up and let it run.
The bear case, strongest form (test 10’s red flag): You are paying 123× earnings and 19× book for a cyclical capital-goods company — however good — whose eye-popping growth is partly a low-base turnaround that must now lap high numbers, whose free cash flow has gone negative as capex balloons, whose working capital is creeping up, and whose management is simultaneously sinking ₹7,600 cr into a commodity, technology-fragile, subsidy-dependent semiconductor business it has never run. If growth merely normalises to 15% and the multiple de-rates toward its peers’ cheaper end, the stock can fall by half while the business keeps doing fine. The single strongest reason this might not compound from here is simply the starting price.
What the numbers actually support: a genuinely good, value-creating business (EP +₹701 cr, RoE 21%, debt-free, 21% growth) — and a price that offers no margin of safety on any framework metric. The quality verdict and the price verdict point in opposite directions, and that’s the whole story.
The two or three things that would tip it: (1) a market or sector de-rating that brings the price toward the ₹250–₹400 band — turning a great business into a great purchase; (2) the chip plant either proving out at a good return (raises the quality verdict) or being kept small and self-funded (removes the bear’s sharpest point); (3) the power order book and margins holding — the engine staying hot. No buy/sell/hold here — just: own the business gladly, mind the seat price.
Sources
- Screener snapshot: https://www.screener.in/company/CGPOWER/consolidated/ (fetched 2026-06-20)
- Concalls: Q4 FY26 — 6 May 2026 (IIFL); Q3 FY26 — 27 Jan 2026 (IIFL), via cgglobal.com
- Annual Reports: FY25 & FY24 (BSE filings)
- Turnaround / fraud: SEBI orders — BusinessToday 18 Sep 2019; Business Standard 4 Oct 2022 (₹10 cr Thapar fine, 5-yr ban)
- Ownership / QIP: Business Standard 21 Dec 2020 (TII control); NSE filing Jul 2025 (₹3,000 cr QIP @ ₹660, >3× subscribed); ScanX 2026 (zero-deviation fund use)
- Semiconductor: BusinessToday 28 Aug 2025 (CG Semi G1, ₹7,600 cr); CRN Asia 2026 (Kaynes Sanand start); company concall (G2 end-2026)
- Management: EY 7 Jun 2024 (Subbiah World Entrepreneur); Business Standard 8 Jul 2024 (Amar Kaul appointment); Outlook Business 26 Feb 2026 (Chola exit reports denied); Trendlyne 2026 (0% pledge); Glassdoor (culture)
- Peers: screener.in snapshots — Hitachi Energy India (POWERINDIA), ABB India, Siemens, Transformers & Rectifiers (TARIL)
- Assumptions: Cost of equity 12%; PAT-growth projection 20–25% for payback; tags HARD (filed) / MEDIUM (guidance) / SOFT (announced/news/opinion).