If you sell anything online, you've probably noticed the two biggest AI companies picked the same fight at almost the same time. In January 2026, Google announced that Walmart products would show up inside Gemini, complete with agentic checkout. Three months earlier, OpenAI had announced almost the identical thing: Walmart products inside ChatGPT, with a buy button attached. Same retailer, same basic pitch, two very different platforms racing to become the place people shop by typing instead of clicking.
That puts brands in an awkward spot. Neither company has published a shared standard, both are shipping new shopping features every few weeks, and most marketing teams don't have the headcount to optimize for both stacks at once. So which one earns your attention first?
I've dug through what's actually live on each platform, not just the announcements, and built this out as a straight comparison so you can make a call based on where the traffic and the technical requirements actually sit today, not on which press release sounded more exciting.
The two stacks, in plain terms
Google's approach is built on infrastructure it already owned. The Shopping Graph now holds more than 60 billion product listings, with roughly 2 billion refreshed every hour, and it feeds three different surfaces: classic Google Shopping results, AI Overviews, and AI Mode. AI Mode itself passed 1 billion monthly active users within about a year of launch, and it's the surface where Google is layering in conversational shopping, comparison tables, and now agentic checkout through a feature called Universal Cart. Since November 2025, agentic checkout has let a shopper set a price alert with Wayfair, Chewy, Quince or a handful of Shopify merchants, and have Google actually complete the purchase with Google Pay once the price drops.
OpenAI's approach is younger and has changed shape twice in under a year. ChatGPT Shopping launched in mid-2025 as an in-chat product discovery layer, pulling structured product data and showing carousels of options inside a normal conversation. Instant Checkout followed in September 2025, letting people buy directly inside ChatGPT starting with Etsy sellers, later extended toward Shopify merchants and Walmart. By March 2026, OpenAI had pulled back on that plan. Instant Checkout is being folded into ChatGPT's Apps framework instead of staying a native buy button, because product selection stayed thin and pricing wasn't always current six months after launch. OpenAI is now prioritizing discovery and search inside ChatGPT, and letting the actual transaction happen inside a connected retailer app rather than natively in the chat window.
Read the full mechanics of each launch in our breakdowns of how ChatGPT Shopping works and what Instant Checkout actually does today.
Usage share: who's actually doing the looking
This is the part most comparison posts skip, and it's the part that should decide your priority order.
By the numbers reported through mid-2026, ChatGPT still has the bigger audience in absolute terms. OpenAI put weekly active users at 900 million in February 2026, up from 400 million a year prior, and later reporting suggested that figure was closing in on 1 billion by summer. But OpenAI's share of the overall AI assistant market has actually been slipping, dropping below 50% for the first time as Gemini climbed to roughly 27.7% share and Claude picked up around 10%.
Google's shopping-specific numbers are separately strong because they're layered onto search behavior people already had. AI Overviews now show up on close to 14% of shopping queries, up sharply from about 2% in late 2025, and that's happening inside a search engine Google already owns over 90% of. In other words, Google doesn't need to convince anyone to switch apps. The AI shopping layer is arriving inside a habit that already exists. ChatGPT has to build shopping intent from scratch, one conversation at a time, inside an app people opened for something else.
If your priority is sheer reach today, Google's stack touches more shopping-intent queries because it's riding on existing search volume. If your priority is where the fastest-moving experimentation and the most generous merchant programs are happening right now, that's still ChatGPT, mostly because OpenAI is iterating faster and has fewer legacy constraints to protect.
How Walmart is playing both sides
Walmart is the cleanest case study here because it didn't pick a side. It signed with OpenAI first, announced in October 2025, then signed a comparable deal with Google three months later in January 2026. Reading both deals side by side tells you a lot about what each platform actually offers a retailer.
The OpenAI partnership put Walmart and Sam's Club products inside ChatGPT's shopping and Instant Checkout flow, built on the Agentic Commerce Protocol OpenAI developed with Stripe. It was framed as getting Walmart in front of people at the exact moment they're asking ChatGPT a shopping question, rather than waiting for them to open a retail app.
The Google deal does something similar but through a different door. It puts Walmart and Sam's Club products inside Gemini conversations, with the ability to link a Walmart account for personalization and to complete a purchase through Google's agentic checkout using Google Pay. Executives on both sides framed it as "agentic commerce," where the AI system does the comparing and, eventually, the buying, rather than a shopper clicking through ten tabs.
The lesson for smaller brands isn't "sign with both platforms." It's that Walmart, with a much bigger budget and integration team than most retailers will ever have, still treated this as two separate, non-overlapping technical builds. Don't assume optimizing for one automatically covers the other.
Google AI Shopping vs ChatGPT Shopping: side-by-side
| Google (AI Mode, Gemini, Shopping Graph) | OpenAI (ChatGPT Shopping, Apps/Instant Checkout) | |
|---|---|---|
| Core data source | Shopping Graph, 60B+ listings, refreshed hourly via Google Merchant Center feeds | Structured product feeds submitted via the Agentic Commerce Protocol (SFTP or API, 15-minute refresh) |
| Reach today | AI Mode passed 1B monthly active users; AI Overviews on ~14% of shopping queries | 900M+ weekly active users reported Feb 2026, shopping still a minority of total usage |
| Checkout model | Agentic checkout via Google Pay, rolling out with select merchants since Nov 2025 | Moving away from native Instant Checkout, shifting purchases into connected retailer Apps |
| Technical entry point | Google Merchant Center, existing Shopping feed infrastructure most brands already run | New submission process via OpenAI's commerce API, separate from Merchant Center entirely |
| What gets you cited | Structured data, feed accuracy, price/availability freshness, review signals already used in Shopping | Feed accuracy plus how well your product content answers conversational, comparison-style questions |
| Fastest-moving feature | Universal Cart and agentic checkout expansion | Apps framework replacing native checkout, still settling as of mid-2026 |
| Best fit for | Brands already running Google Shopping feeds who want to extend existing infrastructure | Brands wanting early visibility in a fast-iterating surface where the playbook isn't settled yet |
What each platform actually asks of your product data
Google's requirement is the one most ecommerce teams already have some version of. If you're running Google Shopping ads or even just Google Merchant Center for organic listings, you have most of the feed infrastructure AI Mode and Gemini shopping already read from. The incremental work is making sure prices, stock status and variant data stay current enough for an agent to transact against them without human review, since Google's agentic checkout won't complete a purchase against stale data.
OpenAI's requirement is closer to starting from zero for most brands. There's no existing "ChatGPT Merchant Center" you already have an account with. You're submitting a feed through the Agentic Commerce Protocol specifically, in CSV, TSV, XML or JSON, with a refresh cycle as tight as 15 minutes if you want to stay eligible for citation in live conversations. That's a genuinely separate technical project, not a checkbox on an existing one.
This is also where the two roadmaps diverge in risk. Google's stack is an extension of infrastructure that's been stable for over a decade. OpenAI's stack has already changed shape twice in less than a year, first launching Instant Checkout as a native buy button, then pulling it back into an Apps model within six months. If you build tightly around today's ChatGPT integration, budget for it changing again.
So where should you focus first
There's no universal answer, but there is a reasonable default based on where most brands already sit.
- If you already run a Google Shopping feed: extend it. The incremental cost of getting Merchant Center data ready for AI Mode and Gemini citation is small compared to standing up a brand-new integration, and the reach is already larger.
- If you're a smaller ecommerce brand without existing feed infrastructure: start wherever your customers already are in the research. If your audience skews toward using ChatGPT for research and comparison (common in categories like electronics, software-adjacent hardware, and niche DTC products), the ChatGPT Shopping and Apps ecosystem is worth the build even without an existing Google feed to extend.
- If you're weighing scarce optimization hours across both: put the first hour into Google, because the technical lift reuses infrastructure you likely already maintain, and put ongoing monitoring effort into ChatGPT, because that's the surface still changing shape and where an early, well-structured feed can outperform competitors who haven't bothered yet.
- If you sell through a major retailer partner (Walmart, Target, Best Buy): check what data your retail partner already requires from you for their own AI integrations. Walmart's dual deals mean suppliers may already be asked for feed data that satisfies both platforms, and duplicating that work internally wastes time.
None of this is permanent. Google's agentic checkout is still limited to a short list of named merchants, and OpenAI's Apps-based checkout model is still settling. Revisit this decision every quarter, not once a year, because the merchant programs on both sides are actively expanding their eligibility lists.
FAQ
Does optimizing for Google Shopping automatically help with AI Mode and Gemini?
Mostly yes. AI Mode and Gemini shopping both draw from the same Shopping Graph that powers regular Google Shopping results, so accurate, well-structured Merchant Center data benefits both. The gap is data freshness: agentic checkout won't transact against stale pricing or stock data, so if your feed updates slower than hourly, that's the first fix.
Is it worth building a feed specifically for ChatGPT if I don't already sell through Etsy, Shopify or Walmart?
It depends on your category. If your customers already research purchases conversationally (comparison-heavy or considered purchases), yes, because the technical bar to submit a compliant Agentic Commerce Protocol feed isn't especially high once you understand the format. If you sell low-consideration commodity goods where price and availability are everything, the reach argument favors Google first.
Which platform is safer to build around long-term?
Google's shopping infrastructure has changed less structurally, mostly because it's extending systems that already existed. OpenAI's shopping stack has already pivoted its checkout model once. That doesn't mean skip ChatGPT, it means build your integration in a way that isn't dependent on any one specific checkout mechanic staying exactly as it is today.
Where this fits into your broader AI visibility work
Neither platform replaces the other, and neither replaces a solid overall AI visibility strategy across the surfaces your customers actually use. The brands that come out ahead over the next year won't be the ones that picked the "right" platform in 2026. They'll be the ones that treated both as ongoing feed-and-content maintenance rather than a one-time integration project, then tracked which one is actually driving qualified traffic back to their store.
If you want a clearer read on where your own brand currently shows up (or doesn't) across ChatGPT, Gemini and the rest of the AI shopping landscape, get in touch or book a call and we'll walk through what we're seeing in the data.
