Roundel is under 1% of Target’s sales and more than a fifth of its operating income. Chandhu Nair built the same machine at Lowe’s.
Target hired Chandhu Nair from Lowe’s as its first chief AI officer, effective August 24, and promoted Purvi Shah to senior vice president of user experience. Both report to chief information and product officer Prat Vemana. The coverage filed it under retail’s AI arms race. Target’s own income statement points somewhere narrower: a $915 million advertising line that grew 41% in a year when merchandise sales fell 2%, and now carries about 18% of company operating income.
Read Nair’s résumé backwards and the hire looks less like a lab hire and more like a monetization hire. At Lowe’s he ran stores, data, AI and innovation. Before that he was vice president of technology, omnichannel marketing and digital customer engagement, where Lowe’s credits him with launching MyLowe’s Rewards, MyLowe’s Pro Rewards and the company’s media network. Before Lowe’s he co-founded and ran Cognitive Retail, an analytics company selling to US retailers, after roughly a decade at Staples. That is a career spent turning shopper data into a billable product.
What Happened
Target announced the two appointments on Tuesday, August 11, eight days before it reports second-quarter results on August 19. Nair joins as senior vice president and chief AI officer after more than six years at Lowe’s, most recently as senior vice president of stores, data, AI and innovation. Shah, who has spent four years at Target and worked on the conversational shopping experience Target launched inside ChatGPT last November, moves up to senior vice president of user experience.
Nair said his first job is listening, and that “The most meaningful AI stories won’t be about what happens in a lab,” pointing instead at store teams, inventory and guest experience. Target named inventory management, employee tooling and faster decisions as candidate areas.
The title is newer than it sounds. IBM research cited alongside the announcement puts the share of large organizations with a chief AI officer at 76% this year, against 26% in 2025. Nair is not joining the leadership team as a peer to the CFO. He is a senior vice president reporting into the CIO, one layer below the executive committee, which is a reasonable place to put someone whose remit is coordination rather than a P&L.
The Backstory
Michael Fiddelke took over as CEO in February 2026 after three consecutive years of falling revenue, a stretch covered in detail in our Target SWOT analysis. On March 3 he laid out a multiyear turnaround: an incremental $1 billion operating investment on top of roughly $5 billion of capital spending, which is where the “about $6 billion this year” figure comes from. Store remodels, private-label rework, a food and beverage reset, and technology.
Then the first quarter landed well ahead of plan. Net sales of $25.44 billion rose 6.7%, comparable sales rose 5.6% after a 3.8% decline a year earlier, and comparable traffic rose 4.4%. Target raised full-year sales growth guidance from around 2% to around 4%. The stock has climbed more than 50% in 2026.
Underneath the recovery sits a second business growing on a different curve. Target reports an advertising revenue line inside net sales, and the footnote is where this gets interesting:
| Period | Advertising revenue | Change | Merchandise sales change |
|---|---|---|---|
| Fiscal 2024 | $649M | ||
| Fiscal 2025 | $915M | +41.0% | -2.0% |
| Q1 2025 | $163M | ||
| Q1 2026 | $246M | +50.9% | +6.4% |
Target’s own footnote says Roundel services land in one of three places depending on the deal: net sales, a reduction of cost of sales, or a reduction of SG&A. So $915 million is the visible portion. Target has separately described Roundel as a business worth about $2 billion a year across more than 2,000 vendors, and told investors back in 2023 that it had passed $1.5 billion. The gap between those characterizations and the reported line does not sit anywhere on the face of the income statement. It sits inside two cost lines, which is why Target keeps explaining gross margin movements by pointing at “advertising and other revenues.”
That accounting choice matters more than it reads. Advertising revenue arrives with no cost of goods. In fiscal 2025, Target’s gross margin dollars fell $795 million. Strip out the $266 million increase in the reported advertising line and the decline would have been $1.06 billion, a third worse. The ad business absorbed a quarter of the merchandise margin damage while nobody was writing about it.
The Plan
Target is not trying to out-scale Amazon on-site. It cannot. Amazon holds somewhere between 75% and 77% of US retail media spend and booked roughly $68 billion of advertising revenue in 2025. Walmart Connect sits around 6.9%. eMarketer expects US retail media to reach about $69.3 billion in 2026, with Amazon and Walmart taking 89% of the roughly $10.5 billion in incremental dollars. Roundel is fighting for a slice of the 11% that is left.
The only direction with room is off-site, and Target has moved fast on the newest off-site surface anyone has built. Three steps in five months:
- November 2025. Target launched a conversational shopping experience inside ChatGPT, with curated recommendations, basket building and checkout across drive up, order pickup and shipping.
- January 11, 2026. Google published the Universal Commerce Protocol, co-developed with Shopify, Etsy, Wayfair, Walmart and Target. Under UCP the retailer stays merchant of record and keeps the customer data, returns and loyalty relationship. Target products became buyable inside AI Mode in Search and the Gemini app.
- February 2026. Target became one of the first advertisers in OpenAI‘s ChatGPT ads pilot, alongside Adobe, Williams-Sonoma and Albertsons. The pilot carried ads from Target and from Roundel’s brand partners, served against keywords in a shopper’s prompt, on the Free and Go tiers.
That third step is the one nobody connected to Tuesday’s hire. Target did not merely put a storefront inside ChatGPT. It put its retail media network inside ChatGPT, and brought its CPG advertisers with it. Target says ChatGPT referral traffic has grown about 40% a month on average, which is a growth rate published without a base, the way companies disclose numbers when the base is small.
The Business Model Angle
Retail media prints money because the retailer owns three things at once: the audience, the surface the audience is looking at, and the transaction that proves the ad worked. Own all three and the marginal cost of an impression is close to zero, which is how a business under 1% of sales ends up carrying a fifth of operating income.
Take away the surface and the arithmetic changes shape. On Target.com, Roundel’s inventory costs Target nothing. Inside ChatGPT, that inventory belongs to OpenAI, and OpenAI prices it. Trade press covering the pilot reported CPMs around $60 and minimum commitments near $200,000, figures OpenAI has not confirmed. Whatever the real number, it is above zero, and it is set by the counterparty. Roundel’s margin on an off-site impression is the spread between what a CPG pays Roundel and what Roundel pays OpenAI. That is an agency business wearing a retail media badge, and agency businesses trade at agency multiples.
There is a second exposure. Roundel’s pitch to a CPG is that Target’s first-party data closes the loop from impression to receipt. Inside ChatGPT, the impression happens on OpenAI’s property and the receipt happens on Target’s, and the party that sees both sides of that ledger over time is the one that keeps the pricing power. OpenAI opened a self-serve ads manager on May 6 with no budget minimum and a $25 daily floor per campaign, having crossed $100 million in annualized advertising revenue and 600 advertisers within six weeks of launch. Every CPG currently buying ChatGPT reach through Roundel can eventually buy it directly.
Walmart already ran the transaction-layer version of this experiment and quit. We covered how its ad business grew to $6.4 billion in Walmart’s marketing strategy. It offered roughly 200,000 products through OpenAI’s Instant Checkout from November 2025, watched conversion come in at about a third of what Walmart.com delivers, and pulled out. Daniel Danker called Instant Checkout “a very temporary moment in time” at a March conference, and OpenAI retired the feature the same month in favor of merchant-run apps. Walmart then shipped Sparky, its own agent, into ChatGPT and Gemini, where reported conversion runs near 70% of its site rate. Walmart is not on the ChatGPT ads launch partner list.
Target and Walmart landed on the same answer for the transaction layer and opposite answers for the media layer.
The Risk
The strongest case against this reading is that Target has no better option and knows it. With Amazon and Walmart hoovering up 89% of incremental retail media dollars, a 1% share network that stays on-site does not grow. Renting a surface at a bad margin beats owning a surface nobody visits, and Roundel has sold off-site programmatic for years, so ChatGPT is a new address rather than a new model. Criteo data from around 500 US retailers showed traffic from LLM platforms converting at roughly 1.5 times other referral sources. If that intent premium holds, a $60 CPM can still clear.
The transaction side also looks well constructed. UCP keeps Target as merchant of record with the data, the returns and the loyalty file, which is exactly what Walmart went back for after Instant Checkout failed. Target did not repeat that mistake.
Three things would falsify the argument here. First, if Target starts breaking out off-site Roundel revenue with a margin figure attached and it holds near on-site levels, the landlord-to-broker concern is wrong. Second, if agentic referral volume gets large enough that Roundel’s ChatGPT inventory becomes a scale business rather than a pilot, the CPM cost stops mattering. Third, if Nair’s remit turns out to be supply chain and store labor rather than monetization, the résumé reading is overfitted.
The one number to watch on August 19 is not comparable sales. It is whether the advertising line keeps compounding above 40% while Target explains gross margin by pointing at it.
Quick Questions
Who is Chandhu Nair? Target’s first chief AI officer, joining August 24 as senior vice president. He spent more than six years at Lowe’s, ending as senior vice president of stores, data, AI and innovation, and earlier led omnichannel marketing technology there, launching MyLowe’s Rewards and the company’s media network. He co-founded retail analytics company Cognitive Retail and spent about a decade at Staples.
What is Roundel? Target’s retail media network. It sells advertising to brands using Target’s shopper data, including the Target Circle loyalty base of more than 100 million members. Target describes it as roughly a $2 billion annual business with more than 2,000 vendors.
Why doesn’t Roundel’s full size show up in Target’s revenue? Target’s earnings footnote states that Roundel services are classified as net sales, or as a reduction of cost of sales, or as a reduction of SG&A, depending on the arrangement. Only the first bucket appears as advertising revenue.
Is Target selling ads inside ChatGPT? Yes. Target joined OpenAI’s contextual advertising pilot in February 2026 as one of the first advertisers, running placements for itself and for Roundel brand partners against keywords in shoppers’ prompts.
How is that different from what Walmart did? Walmart tested letting OpenAI own the checkout, saw about a third of its normal conversion rate, and withdrew. It now runs its own Sparky agent inside ChatGPT and Gemini and keeps the transaction. Both retailers co-developed Google’s UCP standard, which keeps the retailer as merchant of record. The divergence is on advertising, where Target went in and Walmart did not.
Does this change Target’s turnaround story? Not the headline. Comparable sales, traffic and the $6 billion investment programme still drive the narrative. The advertising line changes how much of the profit recovery is merchandising and how much is media.
The Business Model Analyst Take
Titles are cheap in 2026. Three quarters of large companies have a chief AI officer now. What is not cheap is hiring the person who built a rival’s loyalty programme and media network, and putting him one desk away from the user experience lead, eight days before an earnings print where the fastest-growing line on the page is advertising.
Target’s turnaround has a merchandising story and a media story, and the market is watching the first one. The media story is that Roundel went from 11.7% of operating income in fiscal 2024 to 17.9% in fiscal 2025 to 21.7% in the first quarter of 2026, off a revenue line that never crossed 1% of sales. That is a small business doing outsized work on the bottom line, and Fiddelke is right to feed it.
The open question is where it gets fed. Retail media is a landlord business, and Target’s landlord position is weak: 1% of a market where two competitors take nine of every ten incremental dollars. So Target went looking for tenants on somebody else’s building. Inside ChatGPT, Roundel buys what it used to own, and the spread between the CPG’s cheque and OpenAI’s invoice is the whole business. Nair’s job, whatever the press release says about inventory and store teams, is to make that spread survive contact with a counterparty that just opened self-serve.
The same pressure shows up wherever platforms monetize attention they already own, which we broke down in take rate. Retailers spent a decade learning that owning the shelf was the asset. The next decade tests whether they can rent one.
