Astra Microwave — India's defence RF brain, richly priced
Astra Microwave Products Ltd
Snapshot
Astra Microwave makes the radio-frequency and microwave “brains” that go inside India’s radars, electronic-warfare jammers, missile seekers, satellites and weather systems. Think of it as the engineering shop that builds the bits that see, listen and fight — sub-systems, and increasingly whole systems, for the Defence Research and Development Organisation (DRDO), Bharat Electronics (BEL), the Indian Space Research Organisation (ISRO) and the armed forces. Market cap ₹15,990 cr, share price ₹1,684, 52-week range ₹836–₹1,736 (the stock has roughly doubled off its low). P/E 82.9, price-to-book 12.2x, RoE 16.0%, RoCE 20.2%.
What kind of animal is it? A genuinely good, IP-rich engineering business riding a once-in-a-generation defence tailwind — but a capital-hungry one, with a very thin promoter stake and a working-capital cycle that would frighten you in any other industry. 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 would read the same if the share price doubled or halved overnight. Box 2 is today’s perishable reading of what the crowd is paying.
Box 1 — The business (durable):
| Lens | Result |
|---|---|
| Business-quality score | 15.5 / 23 (Quality 7.5/12 · Growth 5/6 · Longevity 3/5) |
| Buffett rubric | 5.5 / 10 PASS |
| Business bucket | Good (high-IP, but capital-hungry; not Great, not Gruesome) |
| Wealth-creator type | Enduring (structural, not a fad) · Volatile-turning-Consistent |
| Economic Profit | +₹53 cr (RoE 16% − CoE 12% on ₹1,315 cr net worth) — just creating value |
A Good business that is a wealth creator — a deep, hard-to-copy engineering franchise that earns a little above its cost of capital and is finally turning cash. None of that depends on what the share costs today.
Box 2 — The price today (a current phenomenon):
| Reading | Result |
|---|---|
| CMP | ₹1,684 (as of 2026-06-20) |
| Price pillar | 0 / 2 (PEG ≈ 1.8x on 46% trailing growth, ≈ 4x on 20% forward · 5-yr payback ≈ 9x) |
| Margin-of-safety band | ₹600–₹850 (where PEG → ~1x on a realistic ~20% forward growth and payback eases toward sane) |
| Mr. Market’s mood now | Greedy — a defence-sector melt-up + a demerger “value-unlock” story + a blow-out Q4 |
| CMP vs the band | Demanding (≈ 2x the upper edge of the band) |
Today the market is pricing it rich — a mood driven by the defence-stock frenzy, a just-announced space-business demerger, and a record quarter, none of which changes the patient business above.
In plain English
Imagine the Indian military wants a new radar. It doesn’t grow the radar from scratch in one place. A prime — usually government-owned BEL, or a DRDO lab — stitches together dozens of pieces. Some of the most important pieces are the radio-frequency guts: the chips and modules that send out the beam, catch the echo, and turn it into a picture. For thirty years, Astra Microwave has been the trusted shop that designs and builds those guts. Founded in 1991 by a clutch of ex-ISRO and ex-DRDO engineers, it grew up inside India’s defence labs. That heritage is the whole story — you cannot buy your way into being a “qualified vendor” for a missile seeker; you earn it over decades, and once you’re in, you’re hard to throw out.
For most of its life Astra was a humble “tier-2” parts supplier earning thin margins on build-to-print work (you give us the drawing, we make it). The last three years changed the business. India decided to build its own weapons rather than import them — about 75% of the defence capital budget now goes to domestic firms — and a flood of orders arrived. Astra walked away from the low-margin foreign work and leaned into high-value, IP-heavy programs where it owns the design. The result is on the page: sales up from ₹641 cr (FY21) to ₹1,163 cr (FY26), operating margin expanded from 12% to 29%, and net profit nearly seven-fold to ₹193 cr. The order book sits at ₹2,141 cr standalone (₹2,600 cr consolidated), and management guides to tripling revenue over four-to-five years off just five or six big programs — radar (Uttam), missile defence (QRSAM), and two Su-30 fighter-jet upgrades (Virupaksha radar, Angad jammer).
Here’s where its moat is, and it’s real but narrow. The moat is technological breadth and qualification — Astra has, in its own words, “horizontal” capability across radar, electronic warfare, telemetry, space and weather, plus its own in-house microwave-chip (MMIC) division built back in 2005, so it isn’t begging foreign suppliers for the critical bits. A competitor would need a decade and a lot of failed projects to copy that. The moat is narrowing-to-widening as Astra climbs from “parts” to “systems” to (its words) the “Holy Grail” of Astra-owned, Astra-branded products. But the customer is essentially one entity — the Government of India — and that single customer pays slowly, demands the lowest price, and changes its mind on timelines. That’s the crack in the boat.
The strategic move of the moment: Astra is splitting off its space, weather and water-radar business into a separately listed company, Astra Space Technologies, one new share for each you own. Management says it’s for “sharper focus”; the market reads it as “two stocks worth more than one” and has bid the price up.
And that’s the one-line tension. This is a Good, durable, finally-cash-generating business with a real engineering moat — but at ₹1,684, on 83x earnings and 12x book, the price already assumes the triple happens, on time, at fat margins, with the cash actually arriving. The business is the boat; today’s seat is being sold at a premium.
Sitting down with the management
Dear partner — let me tell you who you’d be backing.
The founders are engineers, not financiers, and that matters. B. Malla Reddy, the long-time founder-MD, spent two decades inside ISRO’s systems division and DRDO’s labs before starting Astra in 1991; he holds a master’s in automation from the Indian Institute of Science. The team that runs it today — S. Gurunatha Reddy (MD), Dr. M.V. Reddy (Joint MD) — are cut from the same cloth. This is a company built by people who genuinely love the engineering. When the Joint MD reels off a list of complete radar systems they’ve delivered to ISRO and DRDO, you hear pride, not a sales pitch. That’s the genuinely admirable part: deep domain obsession, a thirty-year track record of delivering hard things, and a stated philosophy — from director Atim Kabra — of “speak less, think long term, deliver on what we promise,” and “only the paranoid survive.” They explicitly say they run the company in three-to-four-year blocks, “not for quarterly applause.” On a sharp analyst who called their growth “lethargic,” Kabra pushed back firmly but politely, defending “2,000 people who work here.” That’s an owner’s temperament.
Now the two things that should give you pause, and they are not small.
First, the promoter owns almost nothing — 6.54%. Read that again. The people running this company hold a thirtieth of it; the public holds 70%, institutions the rest. In Buffett’s language, the captains barely eat their own cooking. Over three decades the founders sold down (some original promoters exited years ago), so this is now effectively a professionally-run, widely-held company more than a promoter-controlled one. That cuts both ways: there’s no controlling family extracting value through related-party games — but there’s also nobody with a fortune riding on the share price the way you are. Alignment runs through salary and reputation, not skin in the game. For a business this dependent on long-term R&D bets, that thin stake is a real flag.
Second, look at the cash, not the profit. For years Astra reported profits that never became cash: operating cash flow was minus ₹25 cr (FY23), minus ₹182 cr (FY24), minus ₹90 cr (FY25). The profit was real but it sat in receivables and inventory — money owed by a slow-paying government and parts stockpiled for long-gestation programs. The good news is FY26 finally flipped hard positive: +₹370 cr operating cash flow, ₹304 cr of free cash flow, debt cut from ₹424 cr to ₹288 cr. Management explained the working-capital monster candidly — defence is brutally working-capital-heavy, the primes don’t pass on advances, receivables are “Grade-1 sovereign credit” the banks happily fund, and the reported gross receivables aren’t netted for ~25% customer advances. That candor is to their credit. But the honest reading is: this is a business that, for most of its history, has had to fund the government to grow, and one good cash year doesn’t yet make a pattern.
On the forensic checklist: no auditor qualifications, clean CEO/CFO certifications, no promoter pledging, no related-party value-leakage flags, dividends paid every year (though the payout ratio has drifted down from ~30% to 12% as they retain to fund growth — reasonable). Remuneration looks proportionate. The receivables ballooning faster than sales is the one “credit P&L, debit balance sheet” pattern that fires — but it’s an industry feature, not fraud, and it’s now reversing.
Would Buffett and Agrawal shake hands on this management? A qualified yes — they’d admire the engineering integrity and the candor, and respect the long-term framing. What would change their mind: the 6.54% stake means they’d watch like hawks for any sign the company is run for the managers rather than the 116,000 owners — and they’d want two or three more years of that positive cash flow before they truly trusted it.
What’s on the horizon (live-issues tracker)
1. The order-book-to-cash conversion — the crux. 🟡 Mixed, improving. (Full interrogation below.)
2. The “triple the revenue in 4–5 years” guidance. 🟡 Early, credible-but-unproven. Management has been consistent across two calls: tripling rests on five-to-six named programs (QRSAM, Uttam radar, Su-30 Virupaksha radar, Su-30 Angad jammer, the Rafael JV, electronic mines), it’s “rear-ended” to FY29–FY30, and they pointedly exclude the export and IP-product upside from the number — a healthy under-promise. FY27 guidance is a more modest 15–20% (₹1,300–1,400 cr). The bear note: these are all government programs, and the company itself admits orders “got shifted from FY28 to FY30-31” because of budget and approval delays. Next milestones: QRSAM order (BEL expected June 2026; Astra ~3-4 months after), Uttam radar HAL negotiation (expected Q2/Q3 FY27).
3. The space / meteorology / hydrology demerger. 🟢 On track, fast-moving. In-principle board nod Feb 2026; full scheme approved 10 June 2026; appointed date 1 April 2026; one new Astra Space Technologies share per existing share; separate listing expected Q3 FY27 (Oct–Dec 2026), pending NCLT/SEBI/exchange approvals. The demerged business is ~₹157 cr revenue (≈13.6% of FY26) and carries the company’s highest margins. Bull: a focused, separately-valued space pure-play could fetch a rich multiple; cleaner accountability. Bear: it strips the highest-margin slice out of the core defence company, and “value unlock” demergers are exactly the kind of catalyst that gets over-priced before any value actually unlocks.
4. The climb to Astra-owned IP products. 🟡 Early but real. Management promises “multiple Astra-branded products with complete IP” hitting the market “before Diwali” FY27, plus a first new-product demo “this quarter.” If real, this moves margins and optionality up a notch. If it slips (as defence timelines do), it’s just talk. Crucially, none of this is in the triple-revenue number — so it’s genuine upside, not baked-in hope.
The crux, interrogated
“This investment works if and only if Astra can turn its booming order book into actual cash, fast enough, while a single slow-paying government customer holds the purse strings.”
The plain-English mechanism. Astra’s danger isn’t that orders won’t come — they’re flooding in. It’s the cash cycle. Picture a builder who wins a contract to build ten houses but must buy all the bricks up front, build over three years, and only gets paid in lumps after the buyer inspects each house — and the buyer is a government department that can take 200+ days to pay and stockpiles its own paperwork. That’s Astra. Its cash-conversion cycle ran 491 days (FY24) and 614 days (FY25) — meaning cash was tied up for over a year and a half between spending it and collecting it. To grow, it had to borrow and burn cash (the −₹182 cr and −₹90 cr operating cash-flow years). The moat (qualification, IP) is real, but a business that must finance its customer to grow is structurally fragile if money ever gets tight.
Test the analogy. Is this like an IT-services firm that also bills big government clients? No — and the difference is the comfort. IT services collect in 60–90 days and carry almost no inventory. Astra carries 390+ inventory days and 216+ receivable days, because the hardware is physical, bespoke, and slow. So the analogy reveals the threat is structural to defence hardware, not a one-off. The reassurance is that the receivables are sovereign — the Government of India does eventually pay, in full. The risk is timing and scale, not default.
Named pressures / who’s on the other side. The “competitor” here isn’t a rival stealing share — Astra’s qualification moat largely protects that. The adversary is the customer’s payment behaviour and the working-capital math itself.
| Pressure | What it is | Current state | Proof point |
|---|---|---|---|
| Slow sovereign payment | BEL/DRDO/MoD pay in 200+ days, primes don’t pass on advances | Debtor days 216 (FY26), down from 273 (FY25) | FY26 OCF flipped to +₹370 cr as old debtors were realised |
| Inventory build-up | Long-gestation programs need parts ordered years ahead; min order quantities | Inventory days 394 (FY26), still rising | Management: numbers are “accumulated over 3 decades,” gross of advances |
| Funding the gap | Working capital growth needs debt; risk if rates/credit tighten | Borrowings cut ₹424→₹288 cr in FY26; credit rating upgraded | One prime bank cut Astra’s interest rate |
| Order timing slippage | Govt programs shift right (budget/approval) | TAM window pushed from FY28 to FY30-31 | Management’s own admission on the Feb 2026 call |
The real-world precedent. This is the standard arc of every Indian defence-hardware vendor — BEL itself, Data Patterns, Bharat Dynamics — and the precedent is reassuring on default (sovereign money arrives) but sobering on valuation: these stocks swing violently because investors keep mistaking lumpy, order-driven, cash-late earnings for smooth compounding. The earnings are real; the timing is not bankable quarter-to-quarter.
Answered follow-on questions. Is the damage to growth or to cash? To cash and to the multiple, not to growth — orders are visible. Which part is protected? The margin and the qualification (hard to dislodge); the cash timing is exposed. Has it actually turned? Yes, once — FY26 was genuinely strong (+₹370 cr OCF, debt down). Is it a pattern? Not yet — three of the prior four years were cash-negative. Is the bet with or against the current? With the current on demand (structural indigenisation), with the current on FY26 cash, but against three decades of working-capital history.
Honest verdict on the crux: Not “too hard” — but not yet proven. The business is real, the moat is real, and FY26 showed the cash can come. The open question is whether FY26 was the new normal or a one-year catch-up on stuck receivables. One more year like FY26 turns this from “Good with an asterisk” toward “Good, full stop.” That single data point — does FY27 operating cash flow stay strongly positive as sales grow — is the whole ballgame.
The watch-list (check next quarter):
- FY27 operating cash flow — does it stay clearly positive as revenue grows ~20%? (The crux.)
- QRSAM and Uttam radar orders — do they actually land in FY27 as guided, or slip again?
- Debtor days — continue falling below 200, or creep back up?
- The “before-Diwali” Astra-branded IP products — shipped, or quietly dropped?
- Demerger listing of Astra Space Technologies — completes in Q3 FY27 on the stated 1:1 terms?
- Promoter holding — any change from 6.54% (a rise would be a strong alignment signal).
QGLP scorecard (the Motilal Oswal lens) — the receipts
QGLP = Quality of Business + Quality of Management + Growth + Longevity + Price. Each line scored 0 / 0.5 / 1.
| # | Question (plain meaning) | Score | Evidence |
|---|---|---|---|
| Q1 | Large opportunity? | 1 | TAM ~₹28,000–30,000 cr FY26-30 for the ecosystem; defence indigenisation (~75% of capital budget to domestic firms) — Jun 2026 concall, snapshot about |
| Q2 | Favourable industry structure? | 1 | Few qualified vendors; OPM expanded 12%→29% over 5 yrs = pricing discipline, not a price war — profit_loss OPM row |
| Q3 | Clear, defensible moat? | 1 | 30-yr qualification + in-house MMIC chip division; RoCE >15% in 6 of last 10 yrs — ratios_table; moat real but partial |
| Q4 | High return ratios (>15%)? | 0.5 | RoCE 20.2%, RoE 16.0% now — but RoCE dipped to 3–13% in FY19–FY22; only recently back above 15% — ratios + ratios_table |
| Q5 | Asset-light / low capital intensity? | 0 | Capital-hungry: 4 cash-negative OCF years out of 5; CCC 491–614 days; must fund customer to grow — cash_flow, ratios_table |
| Q6 | Favourable terms of trade (negative working capital)? | 0 | Debtor days 216 vs payable days 73 → ToT ≈ 296%; it banks its customers — ratios_table |
| Q7 | Unquestionable integrity? | 1 | No auditor qualifications, clean CEO/CFO certs, no pledging, no RPT leakage — AR FY25 sections |
| Q8 | Proven execution track record? | 1 | Delivered complete radar/EW/space systems for 30 yrs; FY26 met its own guidance — Jun 2026 concall |
| Q9 | Growth mindset & vision? | 1 | Rising R&D spend; climbing parts→systems→IP-products; credible long-term framing — concalls |
| Q10 | Superior capital allocation? | 0.5 | Reinvests at improving RoE, no dilution, debt now falling — but the 6.54% promoter stake + years of cash burn temper the one-dollar verdict — balance_sheet, shareholding |
| Q11 | Clear succession plan? | 0.5 | Professional bench (MD, JMD, strategy director); founder long since stepped back — but key-engineer attrition flagged by mgmt — Feb 2026 concall |
| Q12 | Minority interests protected? | 1 | 70% public-held, no controlling family extracting value; steady dividend — shareholding |
| Q13 | Structural sector tailwind? | 1 | Defence indigenisation + 15% higher FY27 defence budget; sector growing well above GDP — concalls |
| Q14 | Volume-led (not just price)? | 1 | Growth is new programs + value-addition mix shift, not commodity price — concalls |
| Q15 | Operating leverage? | 1 | OPM 12%→29% while sales near-doubled — textbook operating leverage — profit_loss |
| Q16 | Manageable leverage? | 0.5 | D/E ≈ 0.22 (₹288 cr debt vs ₹1,315 cr net worth), falling — comfortable now, but working-capital debt will rise with sales — balance_sheet |
| Q17 | Market-share gain potential? | 0.5 | Climbing the value chain (tier-1 systems, DCPP roles); but single govt customer caps it — concalls |
| Q18 | Earnings growth > 15% CAGR? | 1 | PAT 5-yr CAGR 46%; forward guidance 15–20% — profit_loss; from a low FY21 base |
| Q19 | Relevant for next 10–15 yrs? | 1 | RF/microwave is foundational to all radar/EW/space — durable demand — about |
| Q20 | Can extend its moat (CAP)? | 1 | Moving up to owned-IP products widens the moat — Jun 2026 concall |
| Q21 | Long growth runway (GAP)? | 0.5 | Large TAM, but lumpy govt-paced orders that slip right — runway long but bumpy — concalls |
| Q22 | Geographic / product diversification headroom? | 0.5 | Export ambition (Europe, IP products) early; ~92% India revenue today — AR segment note |
| Q23 | Adaptive, resilient culture? | 0 (→0.5 marginal) | Survived the FY19–22 trough; “only the paranoid survive” ethos — call it 0.5 |
| Q24 | Valuation reasonable (PEG)? | 0 | PEG ≈ 1.8x on 46% trailing, ≈ 4x on realistic 20% forward; P/E 82.9 — ratios |
| Q25 | Margin of safety (PEG<1 or payback<1)? | 0 | 5-yr payback ≈ 9x; neither test passed — computed |
Business-quality subtotal (Q1–Q23) = 15.5 / 23 · Quality (Q1–Q12) ≈ 7.5/12 · Growth (Q13–Q18) ≈ 5/6 · Longevity (Q19–Q23) ≈ 3/5. Price (Q24–Q25) = 0/2. Canonical QGLP total = 15.5/25.
The pillar pattern: Growth and the upper half of Quality are the strength; capital-intensity, terms-of-trade, and Price are the weaknesses. This is the classic profile of a Good business — wonderful demand and a real moat, but it must keep feeding the working-capital furnace to grow, and right now the price is the thing standing furthest from a buy-zone.
Buffett lens (the Berkshire-letters read)
| # | Test | Result | Evidence |
|---|---|---|---|
| 1 | Good boat? (business > management) | PARTIAL | Good, not Great: high IP but capital-hungry; RoCE 20% with heavy reinvestment |
| 2 | Moat + pricing power | PASS | OPM expanded through the mix-shift; qualification moat is real; RoE > CoE 6/10 yrs |
| 3 | The See’s test (high returns on little capital) | FAIL | Capex modest (~₹40–50 cr/yr) but working capital is the capital sink — 4/5 cash-negative years |
| 4 | One-dollar test (capital allocation) | PARTIAL | Retained earnings did lift book value & RoE; no dilution — but 6.54% stake + cash-burn years temper it |
| 5 | Owner-oriented, candid management | PASS | Unusually candid on working capital; “we don’t run for quarterly applause”; admits delays by name |
| 6 | Integrity / forensic (profit→cash) | PARTIAL | Profit lagged cash for years (the flag) — but no fraud, and FY26 cash finally caught up |
| 7 | Circle of competence / predictability | PARTIAL | Durable demand, but lumpy — order timing genuinely hard to predict; single govt customer |
| 8 | Mr. Market — gift or trap now? | FAIL | Priced for perfection: 83x earnings, 12x book, near all-time high, defence euphoria |
| 9 | Patience / compounding runway | PASS | Long runway at improving RoE; large TAM, climbing the value chain |
| 10 | The honest red flag | — | See below |
Score: 5.5 / 10 PASS — “a real business with real gaps,” in the Buffett bands. The gaps are exactly where you’d expect for a defence-hardware compounder: the See’s test (capital intensity) and Mr. Market (price).
The See’s test, in prose. See’s Candy was wonderful because it earned huge returns while swallowing almost no extra capital. Astra is the opposite kind of good: it earns a decent return (RoCE 20%) but to grow it must tie up cash for 18 months at a stretch in inventory and government receivables. Its fixed-asset capex is light (~₹40–50 cr a year), but its working-capital appetite is heavy. FY26 was the first year in a while it threw off real free cash (₹304 cr) — and that’s the single most important improvement in the whole story. It is not a See’s. It is a competent capital-user that has just started converting.
The one-dollar test, in prose. Over the last decade Astra retained most of its earnings (payout fell from ~30% to 12%) and grew net worth from ~₹290 cr to ₹1,315 cr. Did each retained rupee create at least a rupee of market value? On the share price, yes — handsomely (the stock is a multi-bagger). On the underlying economics, more modestly: RoE is 16% against a 12% cost of equity, so it earns a ₹53 cr “economic profit” — positive, but a thin spread, and only recently restored after the FY19–FY22 trough when RoCE fell to 3–13%. The verdict: capital allocation is good, not great — and the unusually low promoter stake means the people deciding how to spend your retained rupees have very little of their own riding on the answer.
The framework metrics
- Economic Profit = Net Worth ₹1,315 cr × (RoE 16.0% − CoE 12%) = +₹53 cr → creating value, but a thin spread. (CoE = 12%, the studies’ middle benchmark.)
- Terms of Trade = Debtor days 216 ÷ Payable days 73 ≈ 296% → strongly unfavourable; it finances its customers (the opposite of an FMCG).
- 5-yr Payback = Mcap ₹15,990 cr ÷ projected cumulative 5-yr PAT ≈ ₹1,723 cr (assuming 20% PAT CAGR off FY26’s ₹193 cr) = ≈ 9.3x → far from the <1x multi-bagger signal.
- PEG = P/E 82.9 ÷ growth: 1.8x on the (unsustainable) 46% trailing 5-yr PAT CAGR; ≈ 4x on a realistic ~20% forward. Either way, > 1x.
- RoE − CoE spread = +4% ; RoCE > 15% in 6 of the last 10 years (failed the ≥7/10 moat test by one year, dragged down by the FY19–22 trough).
- Consistent vs Volatile = PAT fell >10% in 3 of 12 years, with one >50% fall (FY19); terminal PAT >> initial. Borderline → historically Volatile, now turning Consistent. Value it more on book/P-of-economics than on a clean P/E.
Peer comparison
Mandatory. From each peer’s screener snapshot, as of 2026-06-20.
| Company | Mkt cap (₹cr) | CMP (₹) | P/E | P/B | RoE | RoCE | OPM | Latest sales (₹cr) |
|---|---|---|---|---|---|---|---|---|
| Astra Microwave | 15,990 | 1,684 | 82.9 | 12.2 | 16.0% | 20.2% | 29% | 1,163 |
| Bharat Electronics (BEL) | 3,12,054 | 427 | 51.5 | ~13.0 | 27.6% | 36.5% | 29% | 27,610 |
| Data Patterns (India) | 26,996 | 4,822 | 98.7 | 15.6 | 16.9% | 23.3% | 40% | 925 |
| Centum Electronics | 5,223 | 3,539 | 51.9 | 15.2 | 26.9% | 25.6% | 14% | -52 NP* |
*Centum reported a net loss in FY26; not directly comparable on P/E.
The read: Astra is neither the cheapest nor the dearest in a uniformly expensive asset class. Defence electronics is priced for the indigenisation boom across the board — every name trades at 50–100x earnings and 12–16x book. On quality of returns, BEL is in a different league (27.6% RoE, 36.5% RoCE, the giant prime) and is, remarkably, cheaper than Astra on P/E. Data Patterns is the closest pure-play peer — higher margins (40%) and similar RoE, but even more expensive. So the relative read tempers the harsh absolute one: Astra isn’t an outlier within its sector. But that’s a sector-allocator’s comfort, not a value-investor’s — the whole basket is expensive, and Astra’s distinctive edge (horizontal IP breadth + in-house MMIC) doesn’t yet show up as best-in-class returns the way BEL’s scale does.
Latest quarter & what’s happening now
Q4 FY26 (reported 27 May 2026): the best quarter in the company’s history — standalone sales ₹488 cr, operating profit ₹161 cr (33% margin), net profit ₹106 cr (up ~44% YoY). FY26 full-year: sales ₹1,163 cr, PAT ₹193 cr, and — the headline — operating cash flow of +₹370 cr versus −₹99 cr the prior year. Dividend ₹2.40/share. Fresh Q4 orders ~₹530 cr; standalone order book ₹2,141 cr (consolidated ₹2,600 cr). [HARD — filed results/concall, 27 May 2026]
Two concall takeaways: (1) the margin step-up is driven by mix — higher-value-add exports to the Rafael JV (~45% gross margin vs single-digit on old build-to-print) and high-margin space work; management guided to sustaining current margins, not pushing higher. (2) The “tripling” thesis rests on 5–6 named programs and explicitly excludes export/IP-product upside. [MEDIUM — guidance]
Live catalysts: demerger scheme approved 10 June 2026, separate listing of Astra Space Technologies expected Q3 FY27 [HARD — filed]; QRSAM order expected after BEL’s (June 2026 guidance) [SOFT]; Uttam radar HAL negotiation in final stages [SOFT]; the stock hit an all-time high ~₹1,736 in mid-June 2026 amid a defence-sector rally [HARD — price].
Where the two lenses agree — and disagree
They agree on the big picture: a Good (not Great) business with a real but narrow moat, riding a genuine structural tailwind, currently priced too richly. QGLP’s 15.5/23 and Buffett’s 5.5/10 land in the same place — quality is there, the See’s/capital-intensity test and the price are the holes.
Where they sharpen each other: the Buffett lens is harder on two things the QGLP checklist can wave through. First, capital intensity — QGLP gives partial credit for “improving RoE,” but Buffett’s See’s test flatly fails a business that funds its customer for 18 months to grow, and that failure is the truest description of Astra’s history. Second, the 6.54% promoter stake — a checklist scores “minority protected” as a positive (no controlling family looting it), but Buffett’s “eat your own cooking” test reads the same fact as a concern (nobody with real skin alongside you). Same fact, opposite sign — and that tension is the single most interesting thing about the ownership here. The honest synthesis: a competent, candid, professionally-run engineering franchise — but one where you are more on your own as an owner than the QGLP score alone suggests.
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: roughly ₹600–₹850. Here’s the arithmetic. To satisfy QGLP’s price pillar you’d want PEG around 1x on a realistic forward growth — call it 20% (management’s own near-term guide), implying a fair P/E near 20–25x, not 83x. On FY26 EPS of ₹20.3, that’s roughly ₹400–₹510; allow generously for the visible order book and the embedded IP optionality and you reach perhaps ₹600–₹850 before the price starts pricing in hope rather than delivery. At ₹1,684 the market is paying roughly twice the top of that band — it has already assumed the triple happens, on schedule, at fat margins, with the cash arriving. (Band, not a point — the future is genuinely uncertain, and defence order timing especially so.)
Mr. Market’s mood: greedy, and you can name why. Three things are pushing the price: (1) a broad defence-stock melt-up — the whole sector is running on the indigenisation story and post-conflict order euphoria; (2) the demerger “value-unlock” narrative — markets love a 1:1 spin-off they can re-rate before any value actually unlocks; and (3) a genuinely record Q4 with the cash-flow turn. All three are real. None of them changes the patient business in Box 1. Two of the three are moods (sector sentiment, demerger hype) that can reverse without a single program slipping.
The plain tension: this is a wonderful-ish business at an unwonderful price — the Good business sitting at a Great-business multiple. Buffett’s question is “who’s the patsy at this price?” — and at 83x earnings with the crowd this excited, you’d want to be honest about whether the next buyer is paying for the engineering or for the momentum. And remember: this reading can flip next month — a defence-sector wobble or a demerger delay — without one rivet of the actual business changing.
Conviction texture
The bull case, at its strongest: India is rearming itself for a generation, and Astra owns a thirty-year, hard-to-copy qualification in exactly the RF/microwave guts that every radar, jammer, missile and satellite needs — plus its own chip division so it isn’t hostage to imports. It’s climbing the value chain from parts to systems to owned IP, margins have structurally re-rated from 12% to 29%, the order book gives multi-year visibility, FY26 finally proved the cash can come, and the triple-revenue guidance excludes the export/IP upside. Buy the boat, ride the wave.
The bear case, at its strongest (test 10’s red flag): strip away the share-price romance and you have a Good — not Great — business that earns only ~4% above its cost of capital, has had to fund a slow-paying government through four of its last five years, whose founders own a startling 6.54%, whose entire growth thesis rests on a handful of government programs that the company itself admits keep slipping right, and which trades at 83x earnings and 12x book at an all-time high in a sector running hot. One cash-positive year does not erase a decade of working-capital pain, and “value-unlock” demergers are a classic top-of-cycle catalyst.
What the numbers actually support: a genuinely good engineering franchise (15.5/23, +₹53 cr economic profit, real moat) that has just turned the corner on cash — bought today at a price that already assumes the best case plays out cleanly. The quality is real; the price has run ahead of the proof.
The two or three things that tip it: (1) does FY27 operating cash flow stay strongly positive — the difference between “Good with an asterisk” and “Good, full stop”; (2) do QRSAM and Uttam actually land in FY27 or slip again; (3) does the demerger create real value or just a momentary re-rating. Watch those, and the verdict writes itself.
No buy/sell/hold — the boat and the seat are described; the decision is yours.
Sources
- Screener: https://www.screener.in/company/ASTRAMICRO/consolidated/ (snapshot fetched 2026-06-20)
- Q4 FY26 concall, 27 May 2026 (BSE filing); Q3 FY26 concall, 13 Feb 2026 (BSE filing) —
_concall_Jun-2026.md,_concall_Feb-2026.md - Annual Reports FY24 & FY25 (governance/segment sections) —
_ar_FY24_sections.md,_ar_FY25_sections.md - Founder background: Crunchbase / Bloomberg / company site profiles for B. Malla Reddy (ex-ISRO/DRDO, M.E. IISc) [MEDIUM]
- Demerger details: TradeBrains, Innovacia, multibagg.ai — board approval 10 Jun 2026, 1:1, listing Q3 FY27, ~₹157 cr/13.6% revenue [HARD-filed / MEDIUM-press]
- Price/rally context: Business Standard, MarketsMojo, TopNews (ICICI Securities target ₹1,725; Goldman Sachs Buy ₹1,455) — June 2026 [HARD-price / MEDIUM]
- Peers: Bharat Electronics, Data Patterns (India), Centum Electronics — screener snapshots, 2026-06-20
- Assumptions: Cost of Equity 12%; forward PAT growth 20% for payback/PEG; net worth = equity + reserves FY26.