What does a 17 percent revenue increase hide? Instacart's third-quarter report shows revenue of $1.1 billion for the quarter ended September, up from $939 million a year earlier, while adjusted earnings of $0.60 per share landed below the $0.64 consensusZacks. The gap between those two lines is where the interesting part sits: growth is intact, and the cost of holding it has moved. All figures below refer to the quarter ended September 2026, as of November 10, 2026.
What the quarter actually shows
The platform now reaches more than 2,200 retail banners representing nearly 100,000 stores, which makes its quarterly disclosures a reasonable proxy for on-demand grocery demand rather than a single company's storyRestaurant Hospitality News. When a network that size reports slower earnings growth, the cause is usually structural rather than seasonal.
Two lines in the quarter explain most of it. Advertising inside the marketplace and the enterprise software sold to retailers both carry far better margins than delivery itself, and both depend on retailers trusting the platform with their own customer data. That trust is the real asset, and it is also the constraint.
Why the cart hardware is being retired
The planned wind-down of Caper Cart is the clearest signal in the releaseTradingView. Smart carts were meant to bring the marketplace's data collection into the physical aisle, but the hardware carried installation, maintenance and training costs that a thin-margin grocery channel struggles to absorb. Retiring it moves the same job to software that runs on the retailer's existing systems.
That shift matters for brands. In-store data collected through proprietary hardware was hard to compare with online behaviour; software that integrates with a retailer's own point-of-sale produces a cleaner join between the aisle and the app, which is what retail media buyers have been asking for.
Three revenue lines, three different risks
| Revenue line | What drives it | Main risk |
|---|---|---|
| Marketplace delivery | Order frequency and basket size | Courier cost per order |
| Retail media | Number of active advertiser budgets | Retailer data-sharing limits |
| Enterprise software | Retailers replacing their own stack | Long sales cycles and switching costs |
Reading the three together is more useful than reading any one of them. Delivery volume tells you whether consumer demand is still there, advertising tells you whether brands still believe the audience is worth paying for, and enterprise software tells you whether retailers are willing to depend on the platform rather than compete with it. For a brand, the practical question sits one level lower: which stores can hold availability through the peak hours, because that subset is what produces repeat orders instead of refunds.
Store-level fulfilment is the missing number
A network of nearly 100,000 stores sounds like coverage, but the number that decides service quality is smaller: the stores that can pick, pack and hand off inside the promised window. Chains that join a marketplace quickly often find that part of their footprint cannot support the fastest promise without extra labour, and those stores generate the cancellations that pull down a whole category rating.
What a brand should watch next quarter
For a consumer goods team, the practical question is where the next unit of growth is cheaper to buy: a sponsored placement inside the marketplace, a promotion that lifts basket size, or better content on the retailer's own site. The quarterly numbers rarely answer that directly, but the trend in advertising revenue is a decent proxy for how crowded the first option is becoming.
- Watch advertising growth, not delivery growthMedia revenue growing faster than order volume means the shelf is getting more expensive, and organic placement less reliable.
- Track basket size against courier costA rising basket with flat delivery cost is the only combination that improves unit economics for the platform and the brand.
- Ask retailers about the software switchEnterprise adoption decides whether the platform keeps its data advantage or becomes one more delivery app.
- Recheck store-level availabilityWith 100,000 stores in the network, the average hides the stores that cannot fulfil a thirty-minute promise.
None of this makes the quarter a bad one. Revenue grew, the network kept expanding, and the company still expects the cost of capacity to normalize. What changed is that the easy growth from adding banners is behind it; the next phase depends on extracting more value per store, per order and per advertiser.
FAQ
Does a revenue beat with an earnings miss matter for brands?
Indirectly. It signals that the platform is buying growth, which usually means higher ad prices and more promotional pressure on brands to fund the difference.
Why does the Caper Cart wind-down matter?
It moves in-store data collection from proprietary hardware to software on the retailer's own systems, which changes who owns the shopper data and how it can be joined with online behaviour.
Is 100,000 stores a real coverage number?
It is a network number, not a fulfilment number. The stores that can meet a thirty-minute promise are a smaller subset, and that subset is what determines service quality.
What should a category team do differently next quarter?
Measure availability at store level rather than chain level, and compare the cost of sponsored placement with the return from fixing availability first.
