Ten Thousand
Small Shops.

Eternal is a retail business trading on an internet multiple, and the mismatch is the opportunity. Blinkit turned its first positive adjusted EBITDA in the December 2025 quarter — ₹4 crore on ₹13,300 crore of net order value. The number is immaterial. What it establishes is that a mature Indian dark-store cluster earns a real operating margin, which converts quick commerce from a subsidised habit into a retail format with a knowable steady state.

Once that is true, the valuation question changes shape. It stops being "what multiple does a platform deserve" and becomes "how many stores, at what throughput, at what margin". We model the dark-store estate from 1,301 at FY25A to 6,000 by FY30E and 8,500 by FY35E, with net order value per store per day rising from ₹6.71 lakh to ₹12.30 lakh, and Blinkit's adjusted EBITDA margin on order value moving from -3.3% to 6.4%. Consolidated B2C net order value compounds from ₹62,291 crore to ₹5,31,396 crore.

That build produces three valuations that do not agree, and the disagreement is informative. A nine-year FCFF discounted cash flow returns ₹251. A sum-of-the-parts on FY30E segment metrics returns ₹336. A three-route relative cross-check returns ₹267. We weight them 30 / 50 / 20 for a target of ₹297. The SOTP carries the most weight because consolidated multiples on a four-segment business with margins ranging from 2.0% to 5.9% are close to meaningless.

The most useful single line in the model is the reverse DCF. At ₹242.87 the market is implying 6.2% perpetual growth against India's ~10.5% nominal GDP. The market is not pricing heroic assumptions into Eternal; it is pricing sub-GDP terminal growth into the country's fastest-scaling retail format. That is the mispricing we are underwriting.

India's quick commerce market sits at roughly a ₹1.3 lakh crore annualised run-rate on category GMV, roughly doubling year on year. Blinkit holds about 48% of national order value, with Instamart near 24% and Zepto near 22% as of the December 2025 quarter. On a three-player structure those shares are stable and the category earns a spread. That is not the structure the market currently has.

Reliance, Walmart-backed Flipkart and Amazon all entered during CY2025. None of them needs quick commerce to earn its cost of capital in the next five years; all three need it to defend a catchment. The category's economics are proven. The industry's economics are not yet defended. That single sentence is the bear case, and it is why our terminal margin stops at 6.4% rather than the double digits a genuine oligopoly would support.

Eternal's structural advantage in that fight is not product. It is funding capacity. Food delivery generates a 5.3% margin on order value in FY26E and throws off cash while Blinkit consumes it, and the group carries ₹18,911 crore of net cash after lease liabilities. A competitor without an annuity segment funds the same subsidy war from equity issuance.

The revenue build is a store roll-forward, not a growth rate. Average stores in operation multiplied by 365, multiplied by net order value per store per day, gives segment order value. Revenue is order value times a monetisation rate. Segment adjusted EBITDA is order value times a margin. Three assumptions per year drive roughly 61.1% of the equity value.

The post-IPO history matters because it dates the inflection precisely. Revenue compounded from ₹1,994 crore in FY21A to ₹20,243 crore in FY25A — a 78% CAGR. Reported EBITDA crossed zero in FY24A, at ₹43 crore, and operating cash flow turned positive in the same year. Everything before that is a funding story; everything after is an operating one.

The line to watch is the cash conversion cycle. It has run structurally negative for the entire period — (513) days in FY21A narrowing to (54) days in FY25A — because the business collects from consumers instantly and pays suppliers on terms. Growth is self-funding at the working-capital line. The narrowing is not deterioration; it is the mix shifting toward owned inventory, which carries stock but still collects on delivery.

We forecast ten years explicitly to FY35E rather than the conventional five. The reason is mechanical: a business still opening a thousand stores a year has free cash flow that is not representative of its steady state, so a short explicit window pushes an unreasonable share of value into the terminal calculation. At nine discounted years the terminal value is 70% of enterprise value. At seven it was 76%. The extension does not manufacture value; it relocates it from an assumption into a forecast that can be checked.

Margins fade rather than inflect. Blinkit reaches 5.1% on order value by FY30E and 6.4% by FY35E. Store additions taper from 1,000 a year at peak to 300 by FY35E. Food delivery order-value growth decays from 17% to 7.5%. Nothing in the forecast requires an acceleration.

WACC is built up, not assumed. Peer betas are unlevered at each peer's own capital structure and relevered at Eternal's, giving a relevered beta of 0.99. Against a 6.8% 10-year G-Sec and a 6.0% India equity risk premium that produces a cost of equity of 12.70%. With lease liabilities as the only debt at a 11.0% discount rate, the blended WACC is 12.70% and terminal growth is set at 6.5%.

The terminal value implies an exit multiple of 12.6x FY35E adjusted EBITDA. Cross-checked the other way, applying a 12.0x exit multiple instead of Gordon growth returns ₹244 per share — ₹7 below the base case. The two methods agree closely, which means the perpetual growth rate is not doing unreasonable work.

This is the primary method and carries a 50% weight. Eternal is four businesses with different growth, margin and capital profiles; a consolidated multiple on the blend is close to meaningless. We value each segment on FY30E metrics and discount the total back four years at WACC.

The peer table is the weakest evidence in this note and we weight it accordingly. Eternal trades at 3.9x EV/revenue against a peer median of 5.1x, but as section 02 argues, that comparison sets a 1P retailer against 3P marketplaces and means little. The order-value view is cleaner: Eternal at 1.5x FY27E order value against Swiggy at 0.9x, a 77.7% premium that we think is justified by the margin gap and the funding position.

Each scenario is a live nine-year mini-DCF driven off the three variables that decide this stock — store count, throughput and margin. The assumption deltas are stated, not just the outputs.

The distribution is wide and asymmetric in a way worth stating: the bull case is 82% and the bear is -48%. This is not a low-variance holding. An investor who cannot tolerate a -48% drawdown on a competitive-intensity scenario should not own it at any target price.

1. Competitive intensity is the dominant risk and it is not diversifiable within the position. Reliance, Flipkart and Amazon can each fund a decade of losses. Our bear case at ₹125 assumes they do. Nothing in Eternal's control prevents it.
2. The valuation rests disproportionately on one method. The SOTP carries 50% weight and returns 33.7% more than the DCF. If a reader rejects the segment multiples, the target falls toward ₹255.
3. Throughput assumptions are analyst constructs, not disclosure. NOV per store per day is derived, not reported. If the true blended figure is materially below our ₹7.39 lakh FY26E starting point, the entire margin path shifts right.
4. Peer comparables are weak. Four of six peer rows in section 07 are single-source analyst estimates. The relative valuation carries only 20% weight partly for this reason.
5. Effective tax rate normalisation. We move from 50% in FY26E to 25.17% from FY31E on accumulated-loss absorption. A slower normalisation reduces FY28E–FY30E earnings materially.
6. Regulatory exposure on two fronts: gig-worker classification and social-security obligations, and FDI rules on inventory-led multi-brand retail. Either could alter the 1P model's cost base or structure.
7. Key-person and governance. No promoter holding, a founder-led executive team and a history of rapid capital reallocation into new formats — Bistro, Nugget, District — with limited disclosure on payback.

The model is a fully linked three-statement build with a balance check that reads zero in every period from FY25A to FY35E, a DuPont cross-check on return on equity, and an equity roll-forward that ties to the balance sheet. Reported historicals run FY21A–FY25A as one continuous income statement, balance sheet and cash flow. The revenue build is bottom-up from a dark-store roll-forward, not a growth rate applied to consolidated revenue. Cost of capital is built from CAPM with peer betas unlevered and relevered rather than assumed. The DCF is FCFF over nine explicit years with a Gordon-growth terminal value cross-checked against an exit multiple, plus a reverse DCF.

Reported financials FY21A–FY25A: company annual reports, cross-checked against the consolidated income statement, balance sheet and cash flow as filed. Q3 FY26 and 9M FY26 figures: company quarterly results and investor presentation. Segment order value, dark-store counts and category share: company disclosure and investor presentation for the December 2025 quarter. Reference price, shares outstanding and market capitalisation: exchange data as at 2 March 2026.

Stated limitation. Market data for four of the six peer comparables — Nykaa, Info Edge, FirstCry and Trent — is a single-source analyst estimate and has not been corroborated against a second independent source, contrary to the two-source standard applied to every other hard number in this note. Those rows are flagged in section 07 and the relative valuation they feed carries only 20% of the target weight. A reader relying on the peer median should verify it independently.