Eternal Q2: Blinkit's 0.9% Margin and the Agent Question

Eternal's quick-commerce arm, Blinkit, grew its order value 79.3% in the September quarter and made an Adjusted EBITDA profit of ₹184 crore, against a ₹156 crore loss a year ago. That is 0.9% of what customers ordered. The company is now speeding up: it expects more than 3,500 Blinkit stores by March 2027 and about 5,000 a year later. And the shareholders' letter spends its longest section on a question investors are starting to ask: will AI shopping agents go around platforms like Zomato and Blinkit?

The quarter in one table

Source: [L]. The segments add up to the consolidated Adjusted EBITDA of ₹666 crore, up 197% (our arithmetic for the sum; the total is from [L]).

On the reported (Ind AS) numbers, consolidated revenue was ₹24,138 crore against ₹13,590 crore a year ago, up 77.6% (our arithmetic). Profit was ₹125 crore against ₹65 crore ([R]).

Revenue grew faster than orders for an accounting reason. Blinkit has moved to owning its inventory, so revenue now "includes the full monetary value of goods sold" rather than mostly a commission ([L]). Like-for-like, Adjusted Revenue grew 58%, not 76% ([L]).

Where the profit comes from

Food delivery still pays for most of it. Zomato's ₹627 crore is 94% of the group's ₹666 crore Adjusted EBITDA (our arithmetic). Its margin slipped "marginally" to 5.5% of NOV, while growth in order value (NOV) accelerated for a fifth straight quarter ([L]).

Blinkit has just turned the corner. Its margin has improved for six quarters in a row and is now positive. The letter gives two data points on what a mature Blinkit could look like:

• "Our mature cities already operate at over 5% Adjusted EBITDA margin." ([L])
• Cities with about 96% of NOV are positive at the contribution level, and cities with about 70% are positive at the Adjusted EBITDA level ([L]).

Tax is now a real cost. The parent company's carried-forward losses "will be fully utilized in FY27", so it now provides for tax on business income ([L]). Of ₹323 crore of pre-tax profit, ₹198 crore went to tax (our arithmetic from [R]: ₹323 crore less ₹125 crore profit). Blinkit's, Hyperpure's and District's companies still have past losses to set off ([L]).

The moat question: agents

The letter's longest answer is about AI agents that shop for consumers. Its argument, in short:

1. Ordering food is often browsing, not a chore. Over 65% of Zomato app opens are "exploratory", and reordered carts are "in low single digits" on both apps ([L]).
2. Agents can only bypass a platform when the product exists without it. A restaurant table or a non-exclusive concert ticket does; a meal delivered in minutes or a 10-minute grocery order does not, because the platform runs the kitchen integrations, riders, dark stores and inventory ([L]).
3. Someone has to own the outcome. When an order goes wrong, the letter argues, "An agent can place an order, but it cannot own the outcome" ([L]).

Its conclusion: "Blinkit and Zomato sit in the low-risk zone against the threat of agentic commerce" while "District sits in the medium-to-high risk zone." ([L])

That is a clear statement of where Eternal thinks its moat is: not in the app or the reviews, which the letter says agents will weaken ("Agents will weaken content moats and strengthen supply chain moats"), but in the physical network. The letter also concedes the obvious bias: "We are conscious that as incumbents, we may be predisposed to see resilience where others see vulnerability." ([L])

Spending to widen it

• Stores: 308 net new Blinkit stores in the quarter, 2,751 in all. The old target was 3,000 by March 2027; the new one is "to cross 3,500 stores by March 2027", and about 5,000 by March 2028 ([L]).
• Smaller cities: new stores there reach 1,000 orders a day about as fast as in metros, with order values at 85–90% of metro levels, the letter says ([L]).
• Cash: the cash balance fell ₹1,813 crore in the quarter, to ₹16,475 crore, mainly from festive inventory and the timing of restaurant payments ([L]).
• Restructuring: the board approved moving the Going-out business into a single subsidiary, Zomato Entertainment, and swapping its stake in Adonmo for shares in Mygate (about 4.4% of the merged company, valued at ₹156 crore against ₹112 crore invested in 2022) ([R]).

The letter also introduces dark.ai, a private AI assistant "built for power users", which it says "will launch globally, hopefully within 2026" ([L]). The letter does not say which company owns it or what it costs.

Moat and margin

Focus moat: cost advantage from fulfilment density. The filings support it at the edges: mature cities above 5% margin, contribution positive across almost all of NOV, and new stores ramping at metro speed in smaller cities. They don't yet prove it at the group level: Blinkit's margin is 0.9%.

Margin: rising in quick commerce, flat in food delivery. Blinkit is up for six quarters; Zomato slipped "marginally" to 5.5%. The company says it doesn't manage the business "for near-term margins", and the faster store rollout will weigh on new stores' margins in their early months ([L]).

MoatSCORE (Moat Screener, first pass): Eternal 5.12, AMBER+; for comparison, Swiggy 5.16, AMBER+ ([MS]).

The bear case

• One business carries the group. Food delivery supplies almost all of the Adjusted EBITDA. A slowdown there would show quickly.
• 0.9% is a thin margin to defend. The letter says new stores weigh on margins early, though it expects "more mature stores and cities to more than offset that"; and the plan now adds about 750 stores by March 2027, two quarters away (our arithmetic: 3,500 less 2,751).
• Tax now takes most of pre-tax profit at the group level, as the parent's past losses run out. The parent alone paid ₹214 crore on ₹855 crore, about 25% (our arithmetic from its standalone results); our reading is that the subsidiaries' losses cannot be set against the parent's taxable profit.
• The agent argument is the company's own. It is reasoned, but untested; the letter itself flags the incumbent's bias.
• A new bet with no numbers. dark.ai has no disclosed cost, entity or revenue model.

The case for

• Blinkit made money at the Adjusted EBITDA level while growing order value 79%.
• Food delivery growth is back above 20%, the company's long-term expectation.
• The balance sheet funds the build-out: ₹16,475 crore of cash at quarter end.

What to watch

1. Blinkit's margin as the store count rises from 2,751 to 3,500.
2. Whether Zomato's margin holds at about 5.5% of NOV.
3. District's path to break-even, guided for 4–6 quarters from January 2026 ([L]).
4. Any disclosure of where dark.ai sits and what it costs.

Sources

• [L] Eternal, shareholders' letter, Q2 FY27 (NSE)
• [R] Eternal, outcome of board meeting and unaudited results for the quarter ended 30 September 2026 (NSE)
• [MS] Moat & Margin, Moat Screener and How MoatSCORE works

Educational research, not investment advice. Every figure above is drawn from Eternal's own filings, read from the primary documents on 22 October 2026; derived figures are our arithmetic, labelled as such. MoatSCOREs read live from the Moat Screener on 22 October 2026. No buy/sell recommendations, no price targets, no share prices. Moat & Margin is not a SEBI-registered Research Analyst or Investment Adviser.