Eternal — The Year Blinkit Broke Even and Revenue Tripled on an Accounting Switch

The State of Play

Eternal is the company you still think of as Zomato — it renamed itself in 2025 and now sits as a holding company over four businesses: Zomato (food delivery), Blinkit (10-minute grocery, the “quick commerce” everyone is fighting over), District (booking tables, movies and concerts), and Hyperpure (supplying restaurants). FY26 was a year of two headlines that pull in opposite directions: reported revenue nearly tripled to ₹54,364 crore, while net profit actually fell to ₹366 crore. The revenue explosion is mostly an accounting change, not a tripling of the business; the profit dip is the price of a land-grab in quick commerce. Underneath both, the genuinely important thing happened — Blinkit, the cash-burning bet that has defined the company for three years, crossed into profit. This is the story of how management talked through that year, quarter by quarter, and quietly changed what it was asking investors to judge it on.

The Company

Eternal makes money four ways, and they could hardly be more different. Zomato, the original food-delivery marketplace, connects diners, restaurants and a fleet of gig riders, and takes a cut — a “take rate” — of each order’s value. It is the profit engine and, at roughly 44% of revenue in FY25 (down from 81% in FY22), no longer the main event. Blinkit is the growth engine and the battleground: a network of small neighbourhood warehouses called “dark stores” (closed to the public, picked by staff) that promise grocery delivery in 10–15 minutes. District is the going-out business — restaurant reservations, movie and event tickets. Hyperpure supplies fresh produce and kitchen goods to restaurants. The two big ones — food delivery and quick commerce — are measured not by revenue but by Gross/Net Order Value (GOV/NOV), the total rupee value of everything ordered through the platform, of which Eternal keeps a slice.

The financial profile is a pure growth-stock one. At a ₹2.47 lakh crore market value and a ₹256 share price (down from a 52-week high of ₹368), it trades at an almost meaningless 676× trailing earnings and ~8× book, pays no dividend, and earns a return on equity of barely 1%. Profit is real but tiny — ₹366 crore in FY26 — and leans on ₹1,396 crore of non-operating “other income” (interest on its large cash pile), which means the operating businesses together barely broke even at the group level. This is a company valued entirely on what it will earn years from now.

One structural oddity: Eternal has no promoter. There is no founding family or controlling block — founder Deepinder Goyal holds a modest stake, and the company is professionally run and widely held. Underneath that open register, ownership has changed hands dramatically in two years: foreign institutions have cut their stake from ~54% to ~33%, and domestic institutions have nearly tripled theirs from ~16% to ~36%, overtaking foreigners to become the largest block as the stock de-rated.

The Story So Far

The backdrop: from cash-burning bet to thin profit

The annual reports trace a company spending its way into a new business. FY23 was the year Zomato bought Blinkit; quick commerce showed up as a new segment with a partial-year ₹806 crore of revenue, the company was still loss-making, and Goyal took no salary (a voluntary 36-month waiver). FY24 revenue grew 71% to ₹12,114 crore, quick commerce nearly tripled, and the company turned its first full-year profit (₹351 crore). FY25 revenue grew another 67% to ₹20,243 crore; Blinkit’s segment loss narrowed almost to zero (₹21 crore, from ₹253 crore), food delivery threw off ₹1,541 crore of segment profit, the company earned ₹527 crore — and Zomato renamed itself Eternal. The set-up for the year ahead: a profitable food-delivery business funding a quick-commerce business standing right at the edge of profitability, in a market filling up with well-funded rivals (Zepto, Swiggy’s Instamart, and increasingly Amazon, Flipkart and JioMart).

July 2025 (Q1 FY26) — the model switch that would distort everything

The first call of the year announced the change that explains the whole year’s optics. Blinkit was shifting from a marketplace (3P) model, where independent sellers own the goods on the shelves and Blinkit just connects them to buyers, to an inventory (1P) model, where Blinkit buys and owns the stock itself. Management framed it modestly — a compliance simplification worth about 100 basis points of margin over two or three quarters — but flagged the accounting consequence that matters: once Blinkit owns the inventory, its reported revenue stops being a thin commission and becomes close to the full order value. Revenue would, in the CFO’s words, “start to look like NOV.” A line item was about to balloon for reasons that had nothing to do with the business growing.

The underlying numbers were a mixed picture. Blinkit was running above $5 billion of annualised GOV with its loss narrowing — adjusted EBITDA margin improving from −2.4% to −1.8% of order value — but its contribution margin had slipped to 3.9% from 4.9% a year earlier, and management pointedly refused to give a breakeven date:

“That number is not important… it’s a function of the weighted average of mature stores and new expansion.” — Akshant Goyal, CFO

Food delivery, meanwhile, had decelerated hard: order-value growth halved to 13% year-on-year, from 27% the year before. The dark-store count stood above 1,500 (up from ~600 a year earlier), heading, management said, “eventually” to 3,000.

October 2025 (Q2 FY26) — the revenue line balloons, the breakeven goalpost moves

This is the quarter the accounting switch became visible: reported quarterly revenue jumped from ₹7,167 crore to ₹13,590 crore — nearly doubling — as Blinkit’s inventory ownership rose from ~80% to ~90% and gross order value started flowing through the revenue line. The take rate optically “jumped” ~300 basis points for the same reason. None of this was the business doubling; it was the same business booked differently.

The real growth underneath was genuinely strong — Blinkit’s order value grew 137% year-on-year, with ~1,800 dark stores now en route to 3,000 by March 2027. But two cautions landed. Food delivery softened further to ~14–15%, and management re-cast its long-standing “20% growth” as a long-term aspiration rather than a near-term number. And on Blinkit, management explicitly de-prioritised hitting EBITDA breakeven as a milestone — a quiet walk-back of the Street’s hope for breakeven in the second half of the year, against competition it was now watching warily.

January 2026 (Q3 FY26) — breakeven, ahead of the lowered expectation

Then the bet paid off. The quarter’s marquee result, which analysts opened the call by congratulating management on: Blinkit reached breakeven, and Hyperpure did too. Blinkit’s contribution margin expanded ~90 basis points and its segment EBITDA ~130 basis points sequentially — and, tellingly, it did so even as the take rate fell ~20 basis points and throughput per store dipped 6–7% (newly added long-tail products turn over slower than core staples). The gains came from operating leverage below the gross-profit line, not from charging more. The company had deprioritised the goalpost in October and walked through it in January.

Management used the moment to plant a long-term flag — high confidence in Blinkit reaching a 5–6% margin on order value eventually, with a couple of cities already at 5%, and a target return on capital above 40% — while flatly refusing to guide the near term, because competition had turned, in its word, “irrational.”

“Margins… [are] a multi-variable problem with no linear correlation with just one variable.” — Akshant Goyal, declining to project the next quarter

Two other threads moved. The inventory shift was now ~90% complete. And the growth ambition quietly inflated and hedged at once: hitting ~100% growth would now require 3,500–4,000 stores, not 3,000, and even that was “conditional on a rational market.” District, the going-out business, saw losses spike on an unplanned membership launch (“District Pass”), with breakeven pushed four-to-six quarters out. Albinder Dhindsa, Blinkit’s founder, was elevated to group CEO.

April 2026 (Q4 FY26) — one number to be judged on

The year-end call (held 28 April) completed the shift in how management wanted to be measured. Gone were per-metric promises; in their place, a single anchor: group EBITDA of $1 billion by FY29. Quick commerce was guided to a 60%+ growth rate over three years — and, notably, FY27 growth was explicitly walked down from the earlier 100% to roughly 70–80%, the price of staying disciplined against irrational rivals. The 3,000-store target for March 2027 was reaffirmed (“firmly on track”), though store additions had gone flattish. Food delivery was settling at ~15% order growth, ~20% order-value growth, and a ~5.5% margin. Quick commerce, management confirmed, was now “steadily” profitable; even tiny Hyperpure was posting a small positive margin and was folded into the FY29 goal.

The philosophy was stated plainly, and it is the key to reading everything else: optimise for absolute profit dollars, not margin percentage, and keep reinvesting every incremental gain into growth as long as the returns hold.

“We’ll continue doing that without worrying about what it does to the margin.” — Akshant Goyal

When an analyst reverse-engineered what the $1 billion target implies for Blinkit’s margin — roughly 3–3.5% over the next few years, on the way to 5–6% — the CFO’s response was “broadly the math is fine.” That is as close to hard guidance as the year produced.

Where Things Stand

The reconciliation across the year is unusual, because the most important promises were the ones management stopped making. The hard commitments it did keep, it kept or beat: the inventory transition landed roughly on its two-to-three-quarter timetable, and Blinkit reached breakeven in Q3 — ahead of where the October call had reset expectations. But across the year the company steadily withdrew from specifics: a breakeven date it wouldn’t give, then a margin trajectory it wouldn’t guide, then a 100% growth rate walked down to 60–80%, all of it folded into one distant number — $1 billion of group EBITDA by FY29 — and a stated preference for absolute profit over any ratio an analyst might hold it to.

What a reader is left holding is a gap between optics and economics. The revenue line tripled and means much less than it appears to — most of it is the Blinkit accounting switch, while the figures that matter (order value growing ~60%+ at Blinkit, ~20% at food delivery) sit elsewhere. Reported profit fell and is propped up by interest income, because the operating businesses are being deliberately run near breakeven to fund the quick-commerce build-out. The bet is visibly working at the unit level — Blinkit profitable, mature cities at 5–6% margins, food delivery steady — and just as visibly unfinished, with growth guidance trimmed, store additions flattening, and competition management itself calls “irrational” and expects to persist for about three more years. The valuation (676× earnings, ~8× book) prices the FY29 promise as if it is already in hand; the foreign investors halving their stake and the domestic ones doubling theirs suggest the market is, at minimum, changing its mind about who wants to own that promise. Whether the $1 billion arrives on schedule is, by management’s own deliberate framing, the only scoreboard that now matters.

The Four Checks

1. Quality and moat. Half a moat, with the other half under construction in a war zone. Food delivery is the real thing — a near-duopoly with Swiggy, network effects between diners, restaurants and riders, 800+ cities, settling into ~20% order-value growth at a ~5.5% margin without anyone seriously contesting it anymore. Blinkit is the bigger business by order value and the weaker claim: dark-store density and scale matter, mature cities already earn 5–6% margins, but the field is crowded with funded rivals (Zepto, Instamart, and now Amazon, Flipkart and JioMart) whose behaviour management itself calls “irrational” and expects to persist for roughly three more years. The slipped customer frequency and an admitted loss of order share in Q3 FY26 show the position is contestable. A decent moat in one segment, an unproven one in the segment the valuation rests on.

2. Returns on incremental capital and runway. On the reported numbers, dismal: ROCE of 2.97% and ROE of 1.19% as of the June 2026 snapshot, with three-year ROE averaging 1.35% and FY26 free cash flow negative at −₹1,114 crore. That is by design — the operating businesses are run near breakeven to fund the build-out — but the rubric asks what reinvested capital earns, and right now the honest answer is “not yet measurable at the group level.” Management’s claimed destination is a Blinkit ROCE above 40% with working capital inside 18 days, and the runway (3,000+ dark stores, 60%+ quick-commerce growth guided for three years) is genuinely long. The trajectory supports some benefit of the doubt — Blinkit went from a ₹253 crore segment loss to breakeven in two years — but the high incremental returns are a promise, not a record.

3. Capital allocation for the stage. Broadly rational for a land-grab phase, and judged on actions it scores better than the headline numbers suggest. Every rupee of food-delivery profit and interest income is redeployed into dark stores, warehousing and automation — capex per store deliberately rising, automation pursued “only where ROCE is trackable.” No dividend, correctly, and no buyback history to assess. The discipline shows in what they declined to do: FY27 quick-commerce growth was walked down from 100% to 70–80% rather than matching irrational discounting, and store additions went flattish when the math stopped working. The quibbles are real, though — an ESOP pool above 20 crore shares is a standing dilution overhang (management says no near-term dilution), borrowings are rising, and a cash pile generating ₹1,396 crore of other income is what keeps the P&L positive at all.

4. Price. Demanding by any reading. As of the June 2026 snapshot the stock trades at ₹244 — down a third from its ₹368 high — for a ₹2.36 lakh crore market cap, 644× trailing earnings and 7.6× book, against a 1.19% ROE and a profit that fell in FY26 and leans entirely on interest income. Even taking management’s own FY29 anchor at face value — $1 billion of group EBITDA, roughly ₹8,800 crore — today’s price is about 27 times an EBITDA number three years away that depends on irrational competitors turning rational on schedule. The FII stake halving from 54% to 33% while DIIs tripled to 36% says the market is already arguing about this. The de-rating has taken the price from absurd to merely heroic; it has not made it cheap.

Sources

Built entirely from primary documents fetched from screener.in on 2026-06-05 (a logged-out session; the financial tables and concall list were fully current):

• Earnings-call transcripts: Q1 FY26 (call 21 Jul 2025), Q2 FY26 (16 Oct 2025), Q3 FY26 (21 Jan 2026), and Q4 / full-year FY26 (28 Apr 2026). Important: Eternal’s calls are Q&A-only — management gives no prepared remarks and publishes the headline financials, segment GOV/NOV and take-rate tables in a separate “shareholders’ letter” that is not filed as the transcript. So segment figures here are what management spoke aloud in Q&A; absolute consolidated revenue/profit figures are taken from the screener snapshot’s quarterly tables. (The Q4 transcript was recovered from the company’s investor-relations CDN after the exchange link returned a non-PDF page.)
• Annual reports (BSE): FY23, FY24, FY25 — the trimmed extracts were compliance-heavy and largely yielded the segment-revenue tables rather than the founder letters and full MD&A, so the multi-year arc here leans on those segment numbers plus the snapshot.
• Screener snapshot: company profile, key ratios, pros/cons, and the quarterly / multi-year P&L / shareholding tables (figures through the March 2026 quarter).

Full audit trail — the per-quarter and per-report research digests, raw transcripts, and the snapshot — lives in vault/Sources/Earnings/Eternal Ltd/.