February 05, 2026

The Agentic Commerce War: Google vs OpenAI

Two rival ecosystems are forming around AI-powered shopping. Here’s what it means for merchants.

Five months ago, agentic commerce had one protocol.

Now it has two, and they’re backed by the two most powerful AI companies on Earth.

The timeline

September 2025 - OpenAI launches Instant Checkout inside ChatGPT , powered by ACP (Agentic Commerce Protocol) and Stripe. Merchants pay a 4% transaction fee. Etsy sellers go live first, with over a million Shopify merchants coming soon.

January 2026 - Google CEO Sundar Pichai announces the Universal Commerce Protocol (UCP) at NRF, enabling direct checkout inside Google AI Mode and Gemini. Google sits on 50 billion product listings, 2 billion updated every hour.

January 2026 - Microsoft launches Copilot Checkout, entering the race as a third player.

February 2026 - Klarna joins Google’s UCP , giving it a major payments partner. The protocol now connects AI agents, merchant systems, and payment providers under one standard.

In five months, we went from “interesting experiment” to a full-blown platform war for AI-driven checkout.

Update: Meta, Amazon, and Microsoft are also in the game.

Two ecosystems, two philosophies

	OpenAI (ACP)	Google (UCP)
AI surface	ChatGPT	AI Mode, Gemini
Payments	Stripe, Checkout.com	Visa, Mastercard, Stripe, PayPal, Amex, Klarna, Adyen
Data source	Merchant-hosted product feeds + checkout API	Google Shopping Graph (50B listings)
Merchant fee	4% per transaction	Not yet announced
First movers	Etsy, Shopify (1M+ merchants), Checkout.com	Shopify, Walmart, Target, Etsy, Wayfair + 20 endorsers
Philosophy	Open protocol, merchant-hosted feeds	Open standard, routed through Google Shopping Graph

The critical difference: ACP is an open protocol where merchants host their own product feeds and checkout APIs. Etsy was the first marketplace live with instant checkout. Shopify is rolling out to over a million merchants, including Glossier, Vuori, Spanx, and SKIMS. And Checkout.com recently adopted ACP to bring enterprise merchants on board.

UCP routes through Google’s existing product graph, the same data powering Google Shopping for years.

For merchants, ACP offers more control. UCP offers more reach.

The marketplace déjà vu

This isn’t the first time retailers face a “join or resist” moment.

The same pattern played out with Amazon:

First wave: resistance (“we won’t sell on Amazon”)
Second wave: selective presence (“we’ll list some products”)
Third wave: full embrace (Gap, Nike, Victoria’s Secret all selling on Amazon)

We’re watching the same cycle compress into months instead of years.

The biggest retailers are hedging, joining both ecosystems at once:

Walmart, Target, Etsy: co-developed Google’s UCP and partnered with OpenAI’s ChatGPT. Wayfair co-developed UCP.
Shopify: co-developed UCP and is rolling out ACP to its merchants
eBay: blocked third-party AI agents (though it still allows Google’s shopping bot)
Amazon: actively blocks OpenAI and Perplexity crawling while building its own AI ad integrations

The big players aren’t picking sides. They’re everywhere. Most smaller merchants, however, haven’t made a move yet. And that’s the real risk.

The fragmentation problem

It’s not just two protocols. The ad industry alone has spawned multiple competing standards : IAB Tech Lab’s ARTF, the Ad Context Protocol (AdCP) from a consortium including Yahoo and PubMatic, and Amazon opening its ad stack to AI agents via MCP.

As Simon Poulton from Tinuiti put it: “For any protocol to take off, you need a degree of network effect.”

The risk for merchants isn’t picking the wrong protocol. It’s waiting for a winner while both ecosystems grow without you. You don’t need to choose an ad protocol. But you do need complete, structured product data, because that’s the layer all of these systems read from.

What this means for Shopify merchants

If you’re on Shopify, you’re in a unique position: Shopify is rolling out ACP support to over a million merchants. Your products can already surface in ChatGPT conversations. And if your product data is already in Google Merchant Center, you’ll likely get UCP coverage too.

The good news: both protocols reward the same thing. Concrete steps you can take this week:

Check your ACP feed now. See how your Shopify products appear to AI agents today . This is the fastest way to find out what’s missing.
Fill in every product attribute Shopify offers. Materials, care instructions, dimensions, weight, variant names. AI agents rank products by data completeness, not SEO keywords. A product with “waterproof, fits true to size, 420g” beats one with “premium quality craftsmanship” every time.
Add GTINs and barcodes to every product. Both ACP and Google’s Shopping Graph use these as product identifiers. Without them, your product is harder to match, verify, and surface.
Rewrite product descriptions for facts, not vibes. AI agents extract concrete attributes: “dishwasher safe,” “fits laptops up to 15 inches,” “organic cotton.” Vague brand storytelling gets ignored. The question isn’t “does this sound good?” but “can a machine extract buying criteria from this?”
Keep your Google Merchant Center feed current. When UCP goes live, Google’s agents will use that data. Stale prices, out-of-stock items, or missing categories will cost you visibility in Gemini conversations.

What about European merchants?

Both protocols are US-only for now. ACP’s Instant Checkout works for US merchants selling to US buyers. UCP is rolling out to eligible US retailers first. Google has described UCP as an open standard designed for global adoption, and OpenAI has said European expansion is planned, but neither has announced a date.

The preparation is the same regardless of geography: complete product data, clean Google Merchant Center feeds, structured attributes. European merchants who do this work now will be ready when the protocols arrive in their markets.

The bigger picture

For twenty years, the question was: “Can Google find my store?”

Now the question is: “Can ChatGPT and Gemini find my products?”

McKinsey projects up to $1 trillion in U.S. agent-driven retail revenue by 2030. The war is on. The question isn’t which side wins. It’s whether your products are in the conversation at all.