Google launches Universal Cart across its shopping apps

Per Google's official blog post from Google I/O 2026, Google introduced Universal Cart, an intelligent shopping cart that operates across Google services. The blog states Universal Cart lets users add products while browsing Search, chatting with Gemini, watching YouTube, or reading Gmail, and that items added are monitored in the background for price drops, stock changes, and historical pricing insights. The blog further states the company draws on the Shopping Graph, which it describes as containing 60 billion product listings and supporting more than 1 billion shopping interactions per day.

The blog and contemporaneous coverage from Phandroid, TechCrunch, The Verge, Tom's Guide, and others report the cart can flag product compatibility issues when users aggregate items from multiple retailers, and that Universal Cart leverages Google Wallet data to surface payment perks and loyalty offers. Per Phandroid and the Google blog, checkout can occur through Google Pay with participating merchants including Nike, Target, Walmart, Wayfair, Sephora, and Shopify stores, or users can transfer items to merchant sites to complete purchases.

TechCrunch and Phandroid report Google discussed agentic commerce building blocks alongside Universal Cart. The announcements reference a common agent language UCP (Universal Commerce Protocol) and, according to TechCrunch, Google teased updates to the Agent Payments Protocol (AP2) and said it would bring AP2 to Google products in the coming months.

Companies building unified shopping experiences typically need three technical foundations: a canonical product graph, persistent cross-session user state, and background agents that monitor external retailer feeds and site changes. Implementing compatibility checks across disparate merchant catalogs requires normalized product attributes, variant mapping, and explicit compatibility rules or heuristics. For practitioners, those pieces translate into concrete engineering work: entity resolution at scale, price- and inventory-polling infrastructure, and a rules engine or trained model that emits compatibility signals.

Agentic payment flows such as those implied by AP2 raise a separate set of engineering and product-safety concerns that are industry-wide. Editorial analysis: Observers implementing agent-assisted checkout generally need to design explicit authorization surfaces, enforceable spending caps, idempotent payment flows, and detailed logging for reconciliation and dispute resolution. Those are not unique to Google; they are recurring engineering patterns when software agents gain the ability to transact on behalf of users.

Industry reporting frames Universal Cart as a move to centralize more of the shopping funnel inside Google's ecosystem. TechCrunch and The Verge highlight that consolidating cart state and enabling agentic buys could shift more user activity and transaction data toward a single service layer. Phandroid notes that competitors previously experimented with in-chat purchasing, with OpenAI reportedly rolling back similar in-chat buy flows in 2025 and routing users back to merchant sites. That history is relevant because it illustrates merchant, regulatory, and user-experience friction points that can slow agentic commerce adoption.

Editorial analysis: For the broader AI and e-commerce ecosystem, the Google announcements matter because they combine large-scale product graph data, multi-modal models like Gemini, and payments plumbing. Those components together lower friction for agent-assisted shopping, but they also concentrate sensitive signals about purchase intent, payment methods, and loyalty data within a single provider, which has implications for privacy engineering, data governance, and third-party developer models.

Observers should track merchant uptake and the list of participants for Google Pay checkout, the scope and developer availability of UCP and AP2, concrete privacy and authorization controls presented to users, and any merchant or regulator responses. For practitioners: monitor published APIs, data schemas for product canonicalization, and any Google guidance on agent authorization and audit logs, since those will shape integration patterns for third-party developers and marketplaces.