Google I/O: ‘Universal Cart’ and ‘Agent Payments’ Unveiled for AI Shopping Google is betting that the future of online shopping will be managed by AI agents that don’t just recommend what to buy, but actually complete the purchase — turning the entire journey into a Google-mediated flow.

At Google I/O 2026, the company unveiled “Universal Cart,” described as an “agentic hub for managing shopping in one place” across Search, Gemini, YouTube, and Gmail. The cart, rolling out first in the U.S., lets users add items from multiple retailers into a single persistent basket, tracking price drops, surfacing price history, sending back‑in‑stock alerts, and running AI compatibility checks — such as warning if a PC processor won’t work with a chosen motherboard and suggesting alternatives.

The system builds on Google Wallet and the Universal Commerce Protocol (UCP), an open standard that enables checkout either directly through Google or via a seamless handoff to the merchant’s own site. Launch partners span major brands and platforms including Nike, Sephora, Target, Walmart, Wayfair, Ulta Beauty, and Shopify merchants like Fenty and Steve Madden.

In parallel, Google updated its Agent Payments Protocol (AP2), originally introduced in 2025, to let AI agents make purchases autonomously within user‑defined limits using cryptographically signed mandates. The goal is to turn AI assistants “from passive recommendation tools into active participants in online commerce,” positioning Google to “control more of the entire shopping journey, and potentially the relationship between consumers and the merchants competing for their attention.”

Commentators note that rivals in China and Amazon have already moved aggressively into AI shopping, and analysts estimate “a $5 trillion agentic commerce market by 2030,” heightening the stakes over who owns the default AI shopping layer. Some observers are struck by the sheer scope of Google’s ambition; DeepMind chief Demis Hassabis amplified a remark that “Google Omni might be too powerful,” hinting at unease about how far such integrated AI systems could go.

Continue reading https://foxvector.com