Google Universal Cart arrived at I/O on May 19, 2026, positioning Search, Gemini, YouTube, and Gmail as a unified, agentic checkout channel. The pitch is to capture the more than one billion daily shopping interactions occurring across the Google ecosystem. As Vidhya Srinivasan, VP/GM Ads and Commerce at Google, noted:
We’re building the foundation for agentic commerce, with updates including a new intelligent cart that brings superpowers to your shopping.
The platform, which began its U.S. rollout in the summer of 2026, integrates a roster of retailers—Nike, Sephora, Target, Walmart, Ulta Beauty, Wayfair, Fenty, and Steve Madden. Google describes the utility plainly:
Universal Cart is an intelligent shopping cart and your new hub for shopping on Google.
Crucially, the company has preserved the merchant of record model. As Google stated:
No matter which way you buy, the brand stays the merchant of record.
This design choice keeps the retailer in control of the customer relationship—a departure from more experimental, decentralized models.
Regarding the mechanics, Google notes:
When you’re ready to buy, UCP makes checkout from your cart super smooth. You can check out with Google Pay in just a few taps with many of your favorite brands, or transfer your items to the merchant’s site to complete your purchase.
This is the first card-rails agentic checkout channel of its scale, and it stands in sharp contrast to the x402 protocol. While Google joined the x402 Foundation in April 2026 to align payment standards, the two approaches remain fundamentally different. x402 relies on crypto-native stablecoin settlement on the Base network, reporting roughly 165 million transactions totaling $50 million, with an average transaction value of $0.46. Universal Cart operates entirely on existing card infrastructure via Google Pay.
Despite the scale of these partnerships and the ubiquity of Google’s platforms, the company has remained silent on volume metrics. This absence of data is noteworthy. Google historically highlights product milestones with precision; the lack of public figures here creates a distinct gap between the readiness of the infrastructure and the reality of consumer adoption. The rails are built. The receipts are missing.
The friction is not technical—it is psychological. An April 2026 Product.ai survey of 1,463 U.S. online shoppers found that only 14% of AI users trust recommendations without verification. More pointedly, 42% of consumers will not trust an AI to complete a purchase exceeding $25. With 86% of users preferring to verify purchases before finalizing, the agentic checkout model hits a hard ceiling. The infrastructure may be ready to execute, but the consumer is not yet ready to delegate.
Google is attempting to bridge this divide through protocol convergence and feature expansion. The upcoming release of the Agent Payments Protocol (AP2) for Gemini Spark aims to enable autonomous agent purchases with user-defined guardrails—the oversight layer intended to address consumer hesitation around automated spending. The ecosystem is expanding too: Shopify has integrated native UCP support into its Admin interface, and the Google UCP Tech Council now includes over 60 partners—Amazon, Meta, Microsoft, Salesforce, and Stripe among them.
The platform is also pushing beyond standard retail, testing hotel bookings and local food delivery. A pilot program, Direct Offers, is testing advertiser-exclusive deals surfaced directly in AI Mode. These are the building blocks of a new commerce layer, yet they remain tethered to the same trust threshold that has anchored the market at the $25 mark.
The promise of autonomous shopping remains largely theoretical until that threshold shifts. Whether through improved agent reliability or more sophisticated user-defined guardrails, the industry must find a way to move the needle. For now, the gap between the infrastructure’s readiness and the actual volume of transactions is the defining metric. If the technology is as seamless as the partnerships suggest, why is the data still obscured?