International media reports have revealed that Starbucks’ attempts to modernize its supply chain through artificial intelligence are being thwarted by persistent technical glitches and an antiquated digital backbone.
Despite the rapid rollout of “automated counting” tools under CEO Brian Niccol, store managers and baristas report that the AI-powered apps frequently misidentify inventory, confusing different types of milk or failing to recognize products on shelves entirely.
These digital errors are contributing to a frustrating cycle of “phantom” stockouts, where the system believes an item is available while the physical shelf remains empty.
The Ghost of 1997: Outdated Hardware Issues
A significant hurdle for the coffee giant is its continued reliance on the IBM AS/400 system—architecture the company first touted in 1997. While Starbucks has confirmed it is working to replace this “mission-critical” platform, current tech employees compare the upgrade process to “changing the engine of an airplane while it’s flying.”
This nearly 30-year-old hardware struggles to process real-time inventory data at the speed required for a modern global chain, leading to a massive disconnect between corporate forecasting and the actual needs of local cafes.
A Tightrope of Waste and Shortage
The breakdown in accuracy has led to extreme swings in inventory management. In some regions, the lack of storage space in older Starbucks cafes means that even a minor forecasting error leads to significant waste.
Baristas have reported throwing away “astronomical” amounts of unsold food—sometimes bags full of sandwiches—after misguided automated shipments flooded stores.
This balance of avoiding empty shelves while minimizing perishable waste remains the most pressing challenge for the company’s new leadership as they prepare to update investors on their turnaround strategy.