Walmart Is Using AI to Cut Delivery Times to Under an Hour. Here’s How the System Actually Works.
Walmart wants to get faster at shipping, sorting, and delivery, and it's betting its AI investment can get grocery orders to doorsteps in well under an hour. Speaking during Walmart Associates Week, global CTO and CDO Suresh Kumar described the company's AI models as “this conductor of this orchestra of product movement,” handling everything from inventory forecasting to route optimization in real time.
The results are already showing up in the numbers. Walmart's global ecommerce business grew 26% in the first quarter of 2026, with US online sales growing at the same rate.
What Kumar Means by “Intelligence Layers”
Kumar's framing centers on what he calls intelligence layers sitting at the heart of Walmart's AI models, designed to predict what customers will order and when, then position the right product in the right part of the supply chain before the order even arrives. That's a meaningfully different approach than reactive fulfillment, where a system responds to an order after it's placed. Walmart is trying to have inventory already staged closer to where demand is about to spike.
The company has built out three specific algorithm layers to support this: a selection algorithm, a routing algorithm, and a scheduling algorithm, each handling a different piece of the real-time optimization problem. “Intelligence is where we are investing in all of these places,” Kumar said, “from forecasting to replenishment forecasting, ordering replenishment, scheduling, transportation, and route optimization.”
The Physical Infrastructure Behind the AI
The software layer only works because Walmart has spent heavily on the physical network underneath it. By March 2026, Walmart had retrofitted 23 of its 42 regional distribution centers with AI-powered robotics from Symbotic, and deployed 400 Accelerated Pickup and Delivery centers, automated mini-warehouses built into the backrooms of existing stores. Those centers can pick and pack a complex grocery order in under ten minutes.
Walmart is targeting 65% of stores served by automated fulfillment by the end of 2026, with a goal of cutting unit-level fulfillment costs by 20%. President and CEO John Furner has described the underlying ambition even more specifically: delivering physical solutions, often in under 40 minutes, once a customer's intent is understood.
Why LLMs Made This Possible Now
Walmart has invested in AI and automation for inventory orchestration for years, but the specific breakthrough enabling this current push is the maturity of large language models over the past year. Scot Wingo, CEO of AI startup Refibuy, told Retail Brew that retailers like Walmart essentially have “predictive analytics in a can” now, trained on data patterns that older machine learning systems simply couldn't process.
Andy Szanger, director of strategic industries at CDW, framed the practical shift in terms of what the models actually let Walmart calculate. AI gives Walmart a better understanding of which stores sit within delivery proximity of a given order, which of those stores can handle the staffing bandwidth needed to pick and pack quickly, and how external factors like weather and traffic affect real delivery time, not just distance on a map.
This Sits Inside a Broader AI Reorganization
The delivery speed push is one piece of a larger, more disruptive AI transformation happening across Walmart's technology organization. The company has built a companywide framework around four “super agents”: Sparky, a customer-facing shopping agent, Marty, a partner-facing agent built for suppliers, advertisers, and sellers, a store-centric associate agent, and a developer-facing agent for internal tech teams.
That reorganization has come with real workforce consequences. An internal review led by Kumar and fellow executive Daniel Danker resulted in roughly 1,000 corporate employees being cut or relocated, framed internally as eliminating duplicated effort across teams that had been working on overlapping problems. That followed a similar cut of 1,500 corporate employees roughly a year earlier. The delivery speed gains are real, but they're arriving alongside a genuinely significant internal restructuring, not as a standalone initiative layered on top of an unchanged organization.
What This Means If You Sell on Walmart Marketplace
Faster fulfillment infrastructure changes the baseline customers expect, and that expectation doesn't stay contained to items Walmart fulfills itself. As sub-hour and same-day delivery become normal for Walmart's own inventory, third-party sellers on Walmart Marketplace face rising pressure to match that speed on seller-fulfilled listings, the same dynamic that's played out on Amazon as Prime delivery windows have tightened year over year.
If you fulfill orders yourself rather than through Walmart Fulfillment Services, this is worth treating as an early warning rather than a distant trend. Review your current handling times and delivery windows against what Walmart's own AI-driven network is now capable of, and consider whether WFS enrollment makes more sense for your fastest-moving SKUs specifically, since those are the listings where a customer's expectation gap between your delivery speed and Walmart's own will be most visible, and most likely to cost you the sale.
