How Walmart Is Turning Advanced AI Infrastructure Into Every Day Value

Every week, 270 million customers interact with Walmart. They search for products. They browse recommendations. They check if an item is in stock at their local store.

Behind each of those moments is AI, helping customers find what they need, faster and more reliably.

As AI capabilities rapidly evolve, we’ve modernized our data centers while bringing in the latest high-performance computing infrastructure from leaders like NVIDIA.

This next-generation platform gives us the scale required to train larger, more connected AI systems — systems designed to learn from the full complexity of retail.

Customers won’t see this infrastructure directly. But they will experience the results.

From Siloed Models to Shared Intelligence

Retail generates an extraordinary amount of interconnected data. Customers, products, stores, suppliers and supply chain signals are constantly influencing one another.

Historically, many AI systems have been built to optimize specific use cases — search ranking, recommendations, demand forecasting — independently. But retail isn’t independent. A shift in demand affects inventory. Product attributes influence recommendations. Supply constraints shape what customers see online.

With expanded compute capabilities, we’re investing in unified models that learn from these connections simultaneously.

One example is a large Graph Neural Network (GNN) that maps relationships across hundreds of millions of products, customers and supply chain nodes. Rather than treating each system separately, this approach builds shared representations — foundational intelligence that multiple applications can use at once.

When that shared foundation improves, everything built on top of it improves too.

It’s a more connected way to build AI — and it requires significant computational power to do well.

Where Customers See the Impact

This foundational work supports critical experiences across Walmart, including:

• Search relevance: Helping customers quickly find the right product
• Personalized recommendations: Surfacing items that truly fit customer needs
• Inventory placement: Positioning products closer to demand
• Supply chain forecasting: Improving in-stock reliability

The goal isn’t complexity for its own sake. It’s simple: better decisions, made faster, at Walmart scale.

Compounding Value at Scale

Over the past decade, the industry has learned an important lesson: when you combine high-quality data with scaled compute, intelligence compounds.

By building shared AI foundations — and pairing them with advanced infrastructure — we enable improvements in one domain to strengthen others automatically. A smarter product understanding model improves recommendations. Better demand forecasting strengthens availability. Each advancement builds on the last.

Walmart has long invested in infrastructure that scales with our ambition — from pioneering supply chain systems to building one of the world’s largest ecommerce platforms. And because we operate thousands of stores and depots close to customers, we can turn better predictions into real-world availability and faster fulfillment. That reach powers a rapid delivery network, with a growing share of orders arriving in hours or minutes so customers see the benefits of AI almost immediately.

Today, advanced AI infrastructure represents the next evolution of that strategy: transforming powerful compute into everyday value for customers and associates alike.