Kroger launches ambitious AI transformation program

The Kroger Co. is building a state-of-the-art artificial intelligence (AI) lab and demonstration center with far-reaching enterprise goals.

America’s largest grocery retailer is partnering with AI technology provider Nvidia in an effort to expand Kroger’s freshness initiatives, improve shipping logistics, and create a better shopping experience in stores through digital “twin simulations,” which are virtual models designed to accurately reflect store layouts and other operations.

[Read more: Orangetheory scans studios for layout consistency]

“Our collaboration with Nvidia supports Kroger’s ‘Fresh for Everyone’ commitment,” said Wesley Rhodes, VP of technology transformation & research and development at Kroger. “We look forward to learning more about how AI and data analytics will further our journey to provide our customers with anything, anytime, anywhere.”

The lab will be located in Kroger’s Cincinnati-based headquarters office. It will use the Nvidia AI Enterprise software suite for retail, Nvidia Omniverse Enterprise to build digital twin simulations, and Nvidia cuOpt to developlogistics optimizations.

The initial installation will use nine Nvidia DGX A100 systems, Nvidia Infiniband networking and Nvidia RTX workstations to gather actionable insights on the grocer’s nearly 2,800 stores across the United States.

Utilizing the lab, Kroger will explore how it can leverage AI technology for purposes such as:

• Identifying early indicators of deteriorating freshness through computer vision and analytics.
• Using dynamic routing for last-mile delivery to ensure freshness from source to customer.
• Optimizing store efficiency and processes with digital twin store simulation.

Kroger bases new fulfillment model on AI
Kroger is no stranger to using AI technology in its enterprise. The grocer is expanding the efficiency and reach of its online delivery operation with a distribution network model based on high-tech customer fulfillment centers (CFC). Introduced in partnership with U.K.-based online grocer Ocado in May 2018, the CFC model combines vertical integration, machine learning, and robotics with affordable and fast delivery service for fresh food.

CFC facilities leverage proprietary technology solutions focused on artificial intelligence (AI) and advanced robotics and automation to create more seamless and efficient fulfillment, picking and delivery capabilities for enhanced digital commerce capabilities across the U.S. 

In CFCs, more than 1,000 robots traverse giant 3D grids, orchestrated by proprietary air-traffic control systems in the unlicensed spectrum. The grid, known as The Hive, contains totes with products and ready-to-deliver customer orders.

As customers' orders near their delivery times, the robots retrieve products from The Hive, which are presented at stations for items to be sorted for delivery via an algorithmic sorting process. For example, fragile items are placed on top, bags are evenly weighted, and each order is optimized to fit into the fewest number of bags, reducing plastic use.

Once completed, orders are loaded into a temperature-controlled Kroger delivery van, which can store up to 20 orders. Machine learning algorithms dynamically optimize delivery routes, considering factors like road conditions and optimal fuel efficiency. 

“Nvidia and Kroger’s collaboration on AI and digital twins will further expand Kroger’s digital technology leadership,” said Azita Martin, VP and general manager of retail at Nvidia. “Kroger already is delivering excellent customer experiences in stores and providing fast home delivery. With Nvidia AI, shoppers gain more personalized experiences in physical stores and online.”