FedEx

FedEx moves over 16 million packages a day across 220 countries. Every scan, route change, and failed delivery attempt generates data. To be precise, two petabytes of it. Daily.

For 50 years, the company’s advantage was physical and operational, with 700 aircraft, 200,000+ vehicles, and a network not so easy to replicate.

But the next decade is of data and intelligence.

CEO Raj Subramaniam mentioned, “FedEx is now entering a new era as we build the most flexible, efficient, and intelligent network in history.” The physical network will become an execution layer beneath a real-time intelligence system.

In today’s AI at the Top, we will learn how FedEx is building its “intelligent system.”

We will also explore its use cases, partnerships, and compare FedEx to its competitor, UPS.

Like many companies that generated data for decades, FedEx too had a fragmentation problem.

The company’s data lived across 600 separate analytics environments and 7,500 applications. Teams built their own tools, and there was no access to insights across the organization.

To scale FedEx’s AI applications, the company needed to unify its data.

So they’ve created Atlas, FedEx’s enterprise data platform. It consolidates FedEx’s data estate into a single governed layer, currently supporting 200+ AI use cases.

By the end of 2027, FedEx plans to consolidate 100% of its data into Atlas and cut its application footprint by 50% by 2029.

The company organised its transformation into four pillars running in parallel:

YOU DECIDE

At the enterprise level, even small unseen errors can cause large ripple effects.

That’s why enterprises invest heavily to predict bottlenecks and prevent them even before they occur.

FedEx’s MOBIUS (Maintenance Optimization by IOT Unified Systems) pulls real-time data from multiple points, like IoT sensors, PLCs, ultrasound tools, and imaging systems embedded in sortation equipment.

When ML models detect a failure pattern, the system automatically generates a work order and dispatches a technician.

No human decisions in this loop.

In the next phase, the company plans to expand to aircraft and vehicles.

Logistics is chaotic. Routes planned at 6 am are often not worth following blindly at 9 am.

You need your routes to update in real time.

FedEx optimizes 100,000 first-mile and last-mile routes daily using AI, with real-time traffic, weather, delivery density, and customer preferences as parameters.

Backstory… FedEx acquired RouteSmart Technologies in February 2025. They bought a proven routing engine and layered data that only FedEx has on top of it.

See what we mean when we say data is the moat?

Building an algorithm is step 1, and it becomes powerful when you integrate your data.

Network 2.0, the broader optimization program, has delivered roughly 10% lower pickup and delivery costs in integrated U.S. markets, with about 390 U.S. and Canadian locations integrated as of early 2026 and full completion targeted by end-2027.

FedEx Dataworks is a separate P&L unit that turns internal logistics intelligence into commercial products.

No competitor sees trillions of dollars in annual goods movement across 220 countries in real time. Dataworks provides these data insights to procurement teams, enterprise shippers, and retailers. 

It runs three lines:

As of today, two proof points are public.

The ServiceNow partnership embeds FedEx shipment intelligence into ServiceNow’s procurement workflows.

The Dun & Bradstreet Retail Momentum Index combines FedEx parcel data with D&B business intelligence to give retailers a real-time read on U.S. supply and demand.

We have curated a few more AI use cases across FedEx:

FedEx’s AI strategy is built on a combination of acquisitions, external partnerships, and platform vendors. FedEx treats partnerships as speed.

Here's who’s powering what.

FedEx and UPS are both turning logistics data into operational intelligence, but they’ve taken different approaches to where AI creates the most value.

UPS went deep on one tool, ORION, a route-optimization system built in-house since 2003, and compounded two decades of improvement into it. FedEx arrived later in consolidation but is moving faster with external data monetization.

UPS has more documented operational ROI today. FedEx has a more ambitious architecture.

If Atlas and Dataworks compound the way FedEx intends, the data → operational intelligence game gets interesting fast.

YOUR TURN