Walmart Opens Its Platform to AI Shopping Agents While Amazon Restricts Access

Agent Gateway and MCP Integration

Walmart has introduced an Agent Gateway that runs on an MCP server, allowing vetted AI agents to query its product catalog and complete purchases on behalf of users. The gateway is designed to keep personal data and purchase history private, sharing only the minimal context needed for the transaction.

In contrast, Amazon recently began preventing AI agents from executing orders on its marketplace, a move echoed by several travel and booking platforms that have locked out automated shoppers altogether.

Walmart’s new Agent Profiles, rolling out next month, let customers create a dedicated profile for each agent, set per‑transaction approval or spending caps, and delete the profile at any time. The retailer also confirmed that its standard returns process applies even when an agent makes a mistake, reducing risk for early adopters.

Why it matters for GPU / AI infrastructure: Agentic commerce drives a surge in real‑time inference requests as agents browse, compare, and decide on products. Scaling these workloads reliably requires elastic GPU capacity — exactly the type of on‑demand compute that aigpu.cloud provides for AI‑heavy retail workloads.

As more retailers follow Walmart’s open‑agent approach, demand for low‑latency, high‑throughput GPU clusters will grow, making flexible cloud GPU services a strategic asset for any platform supporting autonomous shopping agents.