The next leg of retail media growth is likely to come from an unexpected corner: the food and grocery delivery apps. The pattern of recent moves suggests that DoorDash, Uber, and Instacart are about to push their advertising businesses off their own apps and into the open web, and that this “offsite” pivot will surface as a named growth driver on Q3 2026 earnings calls in late October and November. The prediction here is specific: before year-end 2026, at least one of the three US delivery platforms is likely to foreground off-platform, or offsite, commerce media as a distinct revenue lever, and the combined US delivery-app ad run rate is likely to cross roughly $5 billion annualized. The signals backing that call are recent, independent, and hard-dated.
This is not a forecast that retail media will grow, which is close to a tautology at this point. It is a narrower claim about where the incremental dollar goes, who captures it, and on what rails it travels. The tell is that the plumbing for offsite delivery-app advertising was assembled in June and July 2026, not promised for some indefinite future.
In short
- The prediction: food and grocery delivery platforms are likely to take their commerce media offsite at scale in H2 2026, with the shift visible as a named growth driver on Q3 2026 earnings calls (late October to November).
- Signal 1: on June 4, 2026, DoorDash rebranded its ad business as a global commerce media platform spanning DoorDash, Wolt, and Deliveroo, with more than 400,000 advertisers and an expanded offsite engine (Symbiosys) reaching search, social, and display.
- Signal 2: on July 15, 2026, The Trade Desk hired Uber’s global head of advertising, Kristi Argyilan, as chief commercial officer to run data partnerships, identity, measurement, and retail media, the exact rails offsite delivery ads would ride.
- Signal 3: the three platforms already clear more than $4 billion in combined annualized ad revenue (Uber past a $2 billion run rate; Instacart ads near $286 million in Q1 2026 alone), a base large enough to justify open-web distribution.
- Timeframe and caveat: the pattern points to first proof points before year-end 2026, though a soft H2 ad market, walled-garden gravity, and delivery-margin pressure could delay the crossing into early 2027.
Why this matters now
Retail media has been the fastest-growing advertising category of the decade, and by most industry estimates it cleared roughly $71 billion in US spend in 2026. Most of that money still lands on-site, meaning ads shown inside a retailer’s own app or website next to a product. That model is nearly saturated: on-site inventory is finite, and the shelf can only hold so many sponsored slots before the consumer experience degrades. The growth question, as we argued in our look at who actually gets to run retail media at $71 billion, is no longer whether the category expands but where the next dollar is monetized.
Delivery platforms sit on an unusually rich first-party dataset: what you buy, when, how often, where you are, and what you almost bought. That data is valuable precisely because it is purchase-grade and location-aware, the two attributes advertisers prize most as third-party cookies fade. The constraint has always been reach: a food-delivery app has tens of millions of users, not the billions of a walled garden. Offsite advertising resolves that constraint by letting the platform sell its data-driven targeting against inventory it does not own, on the open web, in social feeds, and increasingly on connected TV.
The timing matters because the enabling technology and the enabling talent both moved in the same eight-week window. When infrastructure, executive hiring, and revenue scale align inside a single quarter, the pattern usually precedes a public monetization push rather than following one. That is the window this piece is trying to read.
Signal 1: DoorDash rewires food delivery into a commerce media platform
On June 4, 2026, DoorDash consolidated its advertising operations across DoorDash in North America, Wolt in Europe, and Deliveroo in the UK and international markets into a single global commerce media platform. The company put the advertiser base at more than 400,000, a scale that puts it in the same conversation as far older retail media networks. The framing is deliberate: “commerce media,” not “delivery ads,” signals ambition beyond the restaurant tab.
The more important detail sits in the plumbing. DoorDash acquired the offsite adtech firm Symbiosys in 2025, and by the June announcement it said media dollars flowing through that engine had nearly doubled since the deal. Symbiosys lets a brand take DoorDash’s first-party signals and activate them across search, social, and display, that is, off DoorDash entirely. This is the mechanical definition of the offsite pivot.
The launch also bundled a LiveRamp clean room integration and a new homepage format called Spotlight. The clean room is the privacy-safe measurement layer that lets a CPG brand verify incremental reach without raw data changing hands; DoorDash cited tests showing more than 80% of consumers reached were new to the advertiser. Spotlight, meanwhile, reportedly drove roughly 2x the click-through of standard banners and sourced more than a third of sales from new customers for CPG brands. Read together, these are the components an advertiser needs before it will trust a delivery app with open-web budget.
One more tell: DoorDash disclosed that an auto-bidding tool tested in February 2026 saw more than 95% of campaigns beat their minimum return-on-ad-spend target. Performance guarantees of that kind are how a media platform graduates from experimental line item to always-on channel. The pattern suggests DoorDash is building for scale, not for a pilot.
Signal 2: The Trade Desk hires the delivery-ads playbook
On July 15, 2026, The Trade Desk, the largest independent demand-side platform on the open internet, announced it had appointed Kristi Argyilan as chief commercial officer, effective July 27. Argyilan’s resume reads like a map of the commerce media era: global head of advertising at Uber, senior roles at Albertsons Media Collective and at Roundel (Target’s retail media arm), a stint at IPG Mediabrands, and a board seat at LiveRamp. She will report directly to founder and CEO Jeff Green and lead the data partnerships division spanning identity, measurement, retail media, and governance.
Executive hires are noisy signals on their own, but the specificity here is unusually loud. The Trade Desk did not hire a generalist; it hired the person who built an ad business inside a delivery platform, and it pointed her at retail media data partnerships. That is a bet that commerce and delivery data will flow onto the open programmatic web, and that whoever brokers those partnerships captures the value.
The move also fits a broader reshuffle: the appointment was The Trade Desk’s third C-suite addition in recent months, following a new CFO and CMO. Companies rebuild the top team when they are repositioning, not when they are coasting. The prior precedent points to this kind of talent migration preceding a formal product or partnership announcement by one to two quarters.
Note the mirror image with Signal 1. DoorDash built the offsite pipe; The Trade Desk hired the person who knows how to fill it. The two moves are independent, from different companies with different incentives, yet they point at the same destination: delivery-grade purchase data monetized off-platform.
Signal 3: Uber and Instacart put the run rate on the board
The third signal is simply scale, because offsite distribution only makes sense once on-site demand is large enough to spill over. Uber Advertising has passed a $2 billion annualized run rate, and Instacart reported advertising and other revenue of roughly $286 million in Q1 2026 alone, which annualizes past $1 billion. Add DoorDash, and the three US delivery platforms together clear more than $4 billion in annualized ad revenue by mid-2026.
That base matters for two reasons. First, a $4 billion category with strong margins is worth building open-web infrastructure around; a $400 million one is not. Second, growth at this size can no longer come from on-site inventory alone, which is finite, so the marginal dollar has to find new surfaces. The arithmetic pushes these businesses offsite whether or not any single executive wants it to.
Instacart’s public posture reinforces the read: the company has consistently framed advertising, not delivery fees, as its margin engine, and it has partnered outward rather than staying walled. Uber has done the same, extending its ad formats across its rides and eats surfaces and into offsite deals. The pattern across all three is a deliberate widening of where the ad can appear, which is the definition of the shift this piece predicts. It connects to the same demand-side dynamics we traced in our analysis of how US grocery delivery is racing toward 15-minute fulfillment, where the delivery network and the data exhaust are two sides of one asset.
What the pattern suggests
Put the three signals side by side and a coherent thesis emerges. DoorDash built the offsite engine in June; The Trade Desk hired the offsite operator in July; and the underlying ad businesses are large enough that offsite is now the only place meaningful growth can come from. Any one of these on its own would be circumstantial. Together, from three independent actors inside eight weeks, they read as a category-level move rather than a company-level one.
| Signal | Source and date | What it implies |
|---|---|---|
| DoorDash global commerce media platform (400k advertisers, Symbiosys offsite, LiveRamp clean room) | Company announcement, June 4, 2026 | The offsite pipe is built and measurable |
| The Trade Desk hires Uber ad chief Kristi Argyilan as CCO for data partnerships | Company announcement, July 15, 2026 | The open-web DSP layer is staffing for commerce data |
| Uber Ads past $2bn run rate; Instacart ads near $286m in Q1 2026 | Company disclosures, Q1 2026 | The revenue base is large enough to justify open-web distribution |
The synthesis is that delivery commerce media is entering its “third wave.” The first wave was Amazon proving purchase data sells ads on-site; the second was every large retailer launching a media network of its own, a story we tracked as retail media and the independent DSP layer converge. The third wave, the one these signals point to, is high-frequency delivery data going offsite through independent DSPs. The pattern suggests the delivery apps become data and demand engines, while the open web becomes the distribution surface.
The prior precedents are instructive because each wave followed the same sequence: prove the data sells, build the network, then extend reach beyond the owned surface. Amazon spent years monetizing sponsored products on-site before it pushed audiences into its DSP and onto Prime Video. Walmart Connect and other grocer networks followed a compressed version of the same arc. The delivery platforms are now at the “extend reach” stage of that arc, which is the point at which growth reaccelerates.
| Wave | Who led it | Core move | Where it ended up |
|---|---|---|---|
| First (on-site proof) | Amazon | Sponsored products beside search results | A roughly $60bn ad business |
| Second (network build-out) | Walmart, Target, grocers | Every large retailer launches a media network | Roughly $71bn US category by 2026 |
| Third (offsite extension) | DoorDash, Uber, Instacart | Purchase data sold onto the open web via DSPs | The shift this piece predicts for H2 2026 |
What differs this time is the speed. The first wave took the better part of a decade and the second took about five years, but the offsite infrastructure for delivery apps was assembled in a single quarter because the adtech (clean rooms, identity graphs, DSP integrations) already exists off the shelf. That compression is itself a reason to expect the monetization narrative to arrive quickly rather than to build slowly, and it is why the prediction is scoped to quarters, not years.
If that read is right, the falsifiable markers are clear. A future observer in 90 to 180 days can check whether DoorDash, Uber, or Instacart names offsite or off-platform advertising as a growth driver on its Q3 2026 call, whether the combined run rate crosses roughly $5 billion, and whether an independent DSP announces a formal delivery-data partnership. If none of that happens by early 2027, the thesis is wrong.
Wider context: the walled-garden squeeze and the open web
The strategic backdrop is a squeeze between two giants. Amazon Ads is a roughly $60 billion business, an order of magnitude larger than the delivery platforms combined, and it keeps most spend inside its own walls. For a mid-sized commerce media network, matching Amazon on-site is hopeless; the only way to compete for a national brand’s budget is to offer reach the brand cannot get from Amazon, which means the open web.
This is why the independent DSP layer matters so much to the thesis. The Trade Desk’s entire pitch is that it monetizes the internet Amazon and Google do not own, from connected TV to open-web display. A delivery platform that plugs its purchase data into that layer instantly borrows scale it could never build alone. The Argyilan hire is best understood as The Trade Desk positioning to be the default broker for exactly this data.
There is a consolidation dimension too. As offsite monetization raises the value of delivery-app data, it also raises the strategic value of the platforms themselves, which feeds the M&A logic we explored in why another scaled delivery player is likely to exit independence. A subscale delivery app with a promising ad business is a more attractive acquisition target than one selling only logistics. The advertising flywheel and the consolidation endgame are, in that sense, the same story viewed from different ends.
Implications for retailers, brands, platforms, and investors
For brands and media buyers, the near-term implication is a new, cookieless, purchase-grade targeting source that reaches beyond a single app. That is genuinely useful for CPG and QSR advertisers who struggle to prove incrementality; a delivery platform can show it drove a first purchase, not just a click. The practical caution is measurement fragmentation: another network, another clean room, another set of metrics to reconcile.
For retailers with their own media networks, delivery-app offsite is both a competitor and a template. It validates the offsite model that grocers and big-box retailers have been slow to build, and it pressures them to follow, a dynamic adjacent to the one we described in why a second retail media network is likely to buy into connected TV. The retailers that move first to open-web and CTV distribution are likely to defend budget better than those that stay on-site.
For the delivery platforms, the prize is margin. Advertising carries gross margins many times those of the delivery core, so every offsite ad dollar disproportionately helps the path to profitability that all three have promised investors. This is the clearest reason to expect management to talk up offsite on earnings calls: it is the most investor-friendly line in the model.
There is also a channel-conflict wrinkle worth flagging. When a delivery app sells a CPG brand offsite reach, it competes for the same budget the brand might otherwise spend with a grocer’s own network, or with Amazon. That tension is likely to push the delivery platforms toward differentiated positioning, emphasizing purchase frequency and real-time intent rather than raw scale. The platforms that articulate a clear, non-overlapping value proposition are likely to onboard national brands faster than those pitching one more undifferentiated audience source.
For investors, the signal to watch is disclosure granularity. A platform that starts breaking out advertising, or offsite advertising specifically, as a reported segment is telling you it wants the market to value that revenue at software multiples rather than gig-economy multiples. The pattern suggests that reclassification is likely to begin around Q3 to Q4 2026 reporting.
| Scenario | What happens by year-end 2026 | Rough likelihood |
|---|---|---|
| Base case | At least one delivery platform names offsite ads as a growth driver on its Q3 call; combined run rate approaches $5bn | Most likely |
| Bull case | Two or more platforms disclose offsite segments; an independent DSP announces a formal delivery-data partnership; run rate clears $5bn | Plausible |
| Bear case | Soft ad market and margin focus keep spend on-site; offsite stays a footnote until 2027 | Less likely but real |
Caveats: what could go wrong
The most important counter-signal is walled-garden gravity. If national brands keep concentrating budget in Amazon and the large social platforms, the delivery apps may find that open-web reach is available but demand for their specific data is thin. Offsite infrastructure can exist and still underperform if buyers do not show up.
A second risk is timing driven by the ad cycle. H2 2026 could bring a softer advertising market, in which case platforms may delay loud offsite pushes to avoid launching into weak demand. The infrastructure would still be in place, but the earnings-call narrative could slip to early 2027, which would technically falsify the year-end version of this call even if the direction is right.
Third, delivery economics remain fragile. If core delivery margins deteriorate, management attention and capital could swing back to unit economics and away from ad-platform expansion. The prior precedent from 2022 to 2023 shows these companies can deprioritize growth bets fast when the core is under pressure.
Fourth, privacy and clean-room friction are real drag factors. Offsite activation depends on identity resolution and measurement that regulators and browser makers keep tightening. A material privacy ruling or platform policy change could slow the pipe even as the strategy stays sound. On balance, these caveats affect timing and magnitude more than direction, which is why the prediction is hedged to a window rather than a date. A disciplined reader should treat the year-end 2026 marker as the central case and early 2027 as the reasonable slip, rather than reading the thesis as broken if a single quarter passes quietly.
Frequently asked questions
What exactly is the prediction, in one sentence?
That US food and grocery delivery platforms are likely to take their commerce media offsite at scale in H2 2026, and that the shift will surface as a named growth driver on Q3 2026 earnings calls before year-end. It is a claim about where retail media’s next dollar goes and on what rails, not merely that the category grows.
What does “offsite” or “off-platform” commerce media actually mean?
It means a platform selling advertising powered by its own first-party purchase data against inventory it does not own, such as open-web display, search, social feeds, and connected TV. The opposite is “on-site,” meaning ads shown inside the platform’s own app. Offsite matters because on-site inventory is finite, so it is where incremental growth has to come from.
Why do delivery apps have an advertising advantage at all?
Because their data is purchase-grade, high-frequency, and location-aware, the attributes advertisers value most as third-party cookies fade. A delivery app knows what you bought, when, and where, which supports strong incrementality measurement. The historical constraint was reach, which offsite distribution is designed to solve.
Could this prediction be wrong?
Yes, and the piece flags several ways. Walled-garden gravity could keep budgets with Amazon and social platforms; a soft H2 ad market could delay the push into 2027; delivery-margin pressure could redirect attention to core economics; and privacy or clean-room friction could slow offsite activation. These affect timing and magnitude more than direction.
How would I verify or falsify this in 90 to 180 days?
Check whether DoorDash, Uber, or Instacart names offsite or off-platform advertising as a growth driver on its Q3 2026 earnings call, whether the combined US delivery-app ad run rate crosses roughly $5 billion annualized, and whether an independent DSP announces a formal delivery-data partnership. If none of these occurs by early 2027, the year-end version of the call is wrong.
Why is The Trade Desk hire treated as such an important signal?
Because of its specificity: The Trade Desk hired the executive who built Uber’s ad business and pointed her at retail media and data partnerships. That is a bet that commerce and delivery data will flow onto the open programmatic web through independent DSPs. Executive hires are noisy in general, but this one is unusually well-targeted at the exact thesis.
Does this threaten Amazon’s advertising dominance?
Not directly in the near term. Amazon Ads is roughly an order of magnitude larger and keeps most spend within its own walls. The delivery platforms are not trying to beat Amazon on-site; they are trying to offer reach and data Amazon does not, which is why the open web and CTV are central to the strategy.
What should brands and retailers do about it now?
Brands should test delivery-app offsite inventory for incrementality while budgeting for measurement fragmentation across multiple clean rooms. Retailers with their own media networks should treat the move as both a competitive threat and a template, accelerating their own offsite and connected-TV distribution. Investors should watch for platforms reclassifying advertising as a reported segment, a tell that they want software-style valuation.
Analysis based on company announcements and disclosures from June and July 2026. For the primary source on the executive move discussed above, see The Trade Desk’s investor relations announcement here. This piece is forward-looking commentary and not investment advice.