Astra Microwave — a quiet year, sold on the loud ones coming

The State of Play

Astra Microwave closed FY26 with record numbers — ₹1,157 crore of standalone revenue, a ~55% gross margin its CFO admits is “probably as best as it gets,” and operating cash flow that swung from minus ₹99 crore to plus ₹370 crore. And yet the year’s actual growth, about 10%, was roughly half of what management guided at the start. The entire investment case, repeated with increasing elaborateness on every call, is that FY26 was the lull before five or six big defence programs — QRSAM, the Uttam radar for Tejas, two Su-30 upgrade programs, the Rafael JV — fire together and triple the company by FY30–31. The order book ended March 2026 at ₹2,141 crore standalone, nearly twice the year’s revenue.

The Company

Astra designs and builds radio-frequency and microwave electronics — the “super components,” sub-systems and increasingly full systems inside radars, electronic-warfare suites, missile seekers, and satellites. Its customers are DRDO labs, defence PSUs like BEL and HAL, and ISRO; 13 of the 14 Doppler weather radars installed in India are Astra’s. The competitive edge management cites most is backward integration down to the chip: in-house MMIC design (the microwave equivalent of a custom processor), with foundry contracts in Taiwan and France, building up through transmit-receive modules to entire radars. A joint venture with Israel’s Rafael — Astra Rafael Comsys (ARC) — makes software-defined radios for the armed forces.

Structurally it is an odd duck: promoter holding is just 6.54%, frozen there for years — screener flags it as a con — so the company is effectively institution-and-public owned, with FIIs creeping from 4.7% to 7.6% and DIIs to 15.4% over two years. The model is working-capital brutal, as defence electronics tends to be: 216 debtor days, around 394 inventory days, a cash-conversion cycle of 537 days. Revenue is violently seasonal — the March quarter is always the giant (₹258 → ₹354 → ₹408 → ₹488 crore over the last four years) and June the runt. The multi-year arc is a recovery story: sales bottomed at ₹293 crore in FY19 with profit of just ₹10 crore, then climbed every year to FY26’s ₹1,163 crore (consolidated) and ₹193 crore net profit. The annual reports add one quiet structural fact the concalls explain: the reported export line collapsed from ₹231 crore in FY22 to under ₹10 crore by FY25 — a deliberate exit from low-margin build-to-print contract work, not lost business.

The Story So Far

Q1 FY26 (call: August 14, 2025) — big guidance, bigger pipeline talk

The June quarter delivered ₹197 crore of revenue, up 28%, on an order book of ₹1,891 crore. Management’s stated FY26 ambitions were at their most expansive here: order inflows of ₹1,300–1,400 crore, and —

“We are expecting growth of around 18% to 20% year-on-year.” — management, Q1 FY26 call

— with EBITDA margins held around 20% “with slight improvement.” The boldest claim of the year also came on this call: three software-defined-radio contracts worth about $100 million would be in place before March 2026, “with 100% probability,” since they were single-party tenders. The pipeline recitation was long — the Virupaksha AAAU development win from DRDO (the active antenna array that is 55–65% of an entire radar), an expected ₹2,000 crore BEL fire-control radar program where Astra supplies half the subsystems, anti-drone RFIs by the dozen, a new wholly-owned space subsidiary, a ground-penetrating radar, $50 million of MMIC chip exports over five years. Strategy director Atim Kabra framed total identified potential at “in excess of ₹4,000–5,000 crores over the next 5–6 years.”

Q2 FY26 (call: November 13, 2025) — the year softens, the roadmap appears

The September quarter brought ₹213 crore of revenue and a half-year up just 7.2% — nowhere near an 18–20% pace. The FY26 revenue number quietly took its first haircut: guidance was now “₹1,150 crores to ₹1,200 crores,” which against FY25’s ₹1,051 crore is roughly 10–14% growth, not 18–20%. Nobody on the call framed it as a cut.

What management offered instead was the first-ever multi-year roadmap, from Kabra: FY27 around ₹1,400–1,500 crore as QRSAM revenues begin, FY28 ~₹1,650 crore as Uttam starts, FY29 ~₹2,000 crore with Su-30 upgrades, FY30 ₹2,250–2,500 crore with everything in production —

“We have a very good chance of doubling our turnover, 2x plus over the next 3 to 4 years… That will be quarter billion dollars on current exchange rate that we will exceed.” — Atim Kabra, Q2 FY26 call

— plus a $1 billion aspiration with, explicitly, no timeline. Asked whether the growth was back-ended to FY27–28, management answered: “Absolutely.” There were honest notes too: the counter-drone business had come second in three bids and lost all three (the product was too costly and generic), and gross margins of 45–50% were attributed squarely to product mix. When an analyst asked why Astra’s quarterly numbers lag peers like Data Patterns, management pushed back: “Look at us in blocks of years.”

Q3 FY26 (call: February 13, 2026) — best quarter ever, sharpest question ever

The December quarter was genuinely strong: record revenue of ₹258 crore, EBITDA margin of 30.9%, order book over the ₹2,000 crore mark at ₹2,226 crore — ₹1,477 crore of it from defence PSUs and DRDO, ₹369 crore from the weather business. FY26 guidance settled at “approximately 10% growth” — the original 18–20% now formally history — while order-inflow guidance ticked up to ₹1,400–1,450 crore. The four-year frame grew grander: a total addressable market of ₹25,000–30,000 crore, of which —

“MV Reddy and SG Reddy are very confident of giving us a number of around INR8,000 crores to INR10,000 crores worth of new order booking and concurrent sales happening to the tune of about INR7,500 crores plus over the next 4 years.” — Atim Kabra, Q3 FY26 call

The sharpest exchange of the year came from Varun Bahl of Ploutos, who called 10% growth in the middle of India’s defence capex boom “underwhelming” and management execution “lethargic.” Management blamed R&D-heavy programs, inspection delays and design changes, and promised bulk production orders would lift execution “drastically.” Kabra called the framing “offensive” to the company’s 2,000 employees and repeated the house line: Astra is run on a 3–4 year view, not quarterly. Meanwhile the $100 million SDR promise from August was nowhere near landed — the man-portable SDR was still in final trials, bids “possibly opened by March.”

Q4 FY26 (call: May 27, 2026) — guidance met (the revised one), and a demerger

The March quarter did its usual heavy lifting: ₹490 crore of execution, up 16%, closing the year at ₹1,157 crore — squarely inside the November-onwards guidance, about half the August guidance. The quality improvements were real and notable: operating cash flow of +₹370 crore against –₹99 crore the prior year (long-stuck receivables finally collected), a credit-rating upgrade, one bank already cutting its lending rate, a ₹2.40 dividend. Gross margin for the year reached ~55%, and management took the unusual step of talking it down: “you should factor in slightly lower numbers… This is probably as best as it gets.”

The forward story was restated as a tripling: FY26 turnover ×3 by roughly FY30–31, on just five or six programs — QRSAM (first production units already flowing), Uttam (HAL negotiations near final, order expected Q2–Q3 FY27), the Su-30 Virupaksha radar and Angad EW pod, the Rafael JV (guided to over ₹600 crore of FY27 sales, 50%+ growth), and electronic mines — with exports and BEL programs explicitly excluded from the math. Capex stays at a routine ₹40–50 crore a year; the constraint is working capital, not factories. Two new threads: multiple Astra-owned, Astra-branded IP products promised “before Diwali,” and a board-approved demerger of the space, meteorology and hydrology business, detailed scheme due “in the next few weeks.”

The ledger: said vs. delivered

Kept: FY26 revenue guidance — once revised — was met almost exactly (₹1,150 guided, ₹1,157 delivered); margins didn’t just hold at 20% EBITDA but beat it (25% for 9M, 30.9% in Q3); the working-capital improvement promised vaguely in November arrived emphatically by May. Moved: the headline growth number walked down from 18–20% (August) to ₹1,150–1,200 crore (November) to “approximately 10%” (February) without ever being called a cut. Missed or slid: the “$100 million of SDR contracts before March ‘26 with 100% probability” — by the May call, ARC’s entire FY26 order booking was ₹546 crore ($65 million) and the man-portable SDR tender was still unresolved; the BEL ₹2,000 crore fire-control radar orders expected “by next couple of quarters” in August had not been announced by May. The pattern: near-term promises slip; the long-term numbers get bigger and more precise each quarter — ₹4,000–5,000 crore of potential in August became ₹8,000–10,000 crore of bookings by February became a 3x revenue commitment by May, all of it back-loaded to FY29–FY31.

Where Things Stand

Astra enters FY27 guiding 15–20% growth (₹1,300–1,400 crore of sales), with visibility — by its own account — of ₹1,600 crore-plus of order bookings, and an order book of ₹2,141 crore standalone, ₹2,600 crore consolidated. The near-term watch-list is unusually concrete: the Uttam order from HAL in Q2–Q3 FY27, QRSAM main contracts 3–4 months after BEL’s (BEL guided June), branded IP products before Diwali, the ARC JV’s promised 50% growth year, and the demerger scheme due within weeks of the May call. The structural picture is a company whose operations — margins, cash conversion, order book — improved markedly through FY26 even as its growth underwhelmed its own initial telling, now priced (P/E ~70, 10x book) almost entirely on the back-ended tripling its management has now committed to in writing, quarter after quarter. FY27 is the first year that story produces checkable numbers.

The Four Checks

1. Quality and moat. A genuinely capable business with a real but contestable edge. The moat is qualification and depth: backward integration down to in-house MMIC chip design, decades of clearance with DRDO, BEL, HAL and ISRO, and entrenchment in programs (13 of India’s 14 Doppler weather radars are Astra’s) where a new supplier can’t simply bid in. What limits it is the customer set — a handful of state monopsonies that own the programs, set the timelines, and can nurture rivals like Data Patterns or build in-house at BEL. The FY26 counter-drone losses (second in three bids, product “too costly and generic”) show the edge does not travel beyond the qualified niches. A defensible position inside a structurally favoured industry, not a fortress.

2. Returns on incremental capital and runway. ROCE is 20.2% and ROE 16.0% per the June 2026 snapshot, and the trend is the right shape — ROCE has climbed from 3% at the FY19 trough through 17%, 19%, 19% to 20% over the last four years, with operating margin up from 18% to 29% in the same stretch. The runway, if management’s program math holds, is long: an order book of ₹2,141 crore standalone against ₹1,157 crore of revenue, and a claimed ₹25,000–30,000 crore addressable market. The drag is structural — every rupee of growth must be financed through a 537-day cash conversion cycle (216 debtor days, ~394 inventory days), which is why a record year still needed a working-capital miracle (+₹370 crore operating cash flow after –₹99 crore) to look clean. Good and improving returns with real runway, permanently taxed by the working-capital model.

3. Capital allocation for the stage. Mostly rational for a build phase. Capex is deliberately light at ₹40–50 crore a year — management is explicit that the constraint is working capital, not factories — debt has been reduced, a credit-rating upgrade followed the cash-flow swing, and the dividend payout has been walked down from 36% in FY21 to 12% in FY26 as returns improved, which is the correct direction. The portfolio moves have been sensible too: the deliberate exit from single-digit-margin build-to-print exports (₹231 crore of revenue given up), the Rafael JV now guided past ₹600 crore of sales, and a board-approved demerger of the space and meteorology business. No buyback history is visible in the data, and the dividend is token (0.15% yield); the standing oddity is governance rather than allocation — a promoter holding frozen at 6.54%, so nobody with skin owns these decisions. Rational with quibbles.

4. Price. Demanding to the point of pre-payment. As of the June 2026 snapshot the stock trades at ₹1,449 — within 2% of its 52-week high — at a P/E of 71.2 and 10.4 times book, for a business that just grew 10% (half its own opening guidance) and earns a 16% ROE. The five-year profit CAGR of 46% is real, but the market is paying now for the back-ended tripling management has promised for FY30–31, on programs (Uttam, QRSAM main contracts, the SDR tenders) that have a documented habit of slipping. At this multiple the next five years of the story are already in the price; FY27 merely gets to confirm or dent it.

Sources

• Concall transcripts (4): Q1 FY26 (Aug 14, 2025), Q2 FY26 (Nov 13, 2025), Q3 FY26 (Feb 13, 2026), Q4 FY26 + full-year (May 27, 2026) — BSE filings, converted to markdown.
• Annual reports (3): FY23, FY24, FY25 high-signal sections. Note: all three extracts were thin on narrative (no chairman’s letters or MD&A prose survived trimming); their main contribution is the geographic revenue split showing the deliberate export/BTP exit.
• Screener.in snapshot: consolidated quarterly and annual tables, ratios, shareholding — fetched 2026-06-05 (logged-out session).
• Research files: vault/Sources/Earnings/Astra Microwave Products Ltd/ — raw transcripts, AR sections, snapshot, per-document digests (not published).