Military logistics is a critical combat multiplier that can be the decisive factor between success and failure on the battlefield. However, several challenges currently stand in the way of America’s defense logistics effectiveness at home and around the globe.

• Logistics data is siloed and distributed across cumbersome information systems, making data difficult to access and harness at scale for quick and accurate decision-making;
• Current military logistics systems lack essential predictive modeling and simulation capabilities critical for strategic planning; and
• Adversaries are rapidly embracing new technologies, including artificial intelligence.

A burdensome information architecture

The U.S. Department of Defense (DoD) is a highly complex enterprise consisting of hundreds of individual organizations. Each of these units have their own subsystems for logistics planning and execution, along with their own definitions of success. The absence of a single ERP means data is siloed across cumbersome information systems, handicapping logisticians’ ability to get the right stuff to the right place at the right time. 

These data limitations, coupled with insufficient visibility and interoperability across systems, have consistently led to a lack of confidence in military supply processes. Starting after the Gulf War, commanders adopted a “just in case” management philosophy to safeguard their units against unpreparedness. This narrowness of focus has negatively affected global force readiness on the whole. It has also resulted in excess stockpiling at DoD wholesale depots and military treatment facilities (MTFs) for more than thirty years. 

In 1990, Walter Reed Medical Center in Washington D.C. had 10 warehouses containing 6 months of pharmaceuticals. This was far more inventory than necessary, and up to $50,000 worth of expired materiel had to be destroyed each month (Duvall, 2003). More recently, from 2020-2022, excess on-hand potential reutilization stocks (PRS) of secondary items alone increased by $2.8 billion (DoD Supply Chain Fact Book FY 2022). 

The Pentagon has made efforts to address these issues, but an “ivory tower” architecture has hindered operational effectiveness in real-world scenarios. Consequently, inefficient resource and budget allocation, and suboptimal readiness remain.

Data as a strategic competitive advantage

The U.S. military collects 22 terabytes of operational data every day. However, since automated information systems (AIS) largely live in isolation across units and branches of service, there is currently no way to exploit this data. DoD prowess in machine learning (ML) technologies like forecasting, scenario wargaming, and automation have fallen significantly behind that of the private sector. 

This is a real and present vulnerability (and opportunity) that must be addressed to deter aggression. U.S. and NATO adversaries are investing heavily in these artificial intelligence (AI) driven capabilities. And while adversarial AI gets headlines when used in targeted cyberattacks, it’s also being used to fortify enemy logistics at machine speed. 

China understands that whoever wins the AI competition will tip the scales of global conflict. In a July 2023 testimony to the House Armed Services Subcommittee on Cyber, Information Technologies, and Innovation, Scale AI founder Alex Wang warned that the U.S. is losing ground to the People’s Liberation Army (PLA).

“If you compare as a percentage of their overall military investment, the PLA is spending somewhere between one to two percent of their overall budget into artificial intelligence whereas the DoD is spending somewhere between 0.1 and 0.2 of our budget on AI.”

In order to remain the world’s supreme fighting force, the warfighter needs to be sustained from headquarters to downrange and everywhere in between. AI is the most effective tooling to ensure operational superiority and overall mission success. The DoD and allied partners must outpace hostile forces in this military technology arms race lest there be a shift in the balance of global power.

Commercial trends influencing defense logistics 

Taking military logistics out of the 1990’s and into the 21st century will certainly be a challenge, but it’s not uncharted territory. Much of the work has already been put into practice by the private sector at exponentially greater scale.

For instance, as mentioned previously, the military ingests 22 terabytes of data each day. The best-in-class industry standard for logistics, Amazon, collects approximately 1 exabyte of purchase data from its customers. That’s 1 million terabytes.

Amazon continues to shape how data science, AI, and advanced robotics can improve logistics operations and decision making. For example, by analyzing consumer purchase patterns, its predictive analytics capabilities can forecast demand for more than 400 million products to ensure adequate inventory levels across all nodes of the supply chain. For comparison of scope, the DoD works with approximately 9,000 suppliers and sustains 900,000 consumable hardware items. Amazon is also able to leverage its data for product regionalization and automated order processing. This puts products closer to consumers, shortening fulfillment times.

Generative AI, a category of artificial intelligence technologies that are designed to create new and original content (i.e. ChatGPT), has been a hotly discussed topic in recent days. Amazon has put its weight behind this technology to create synthetic data for conducting simulations of real-world scenarios. Machine learning models can train on these scenarios to predict and create solutions to challenges that warehouse robots could encounter, for example. By training on thousands of potential simulations, Amazon ensures it can always meet customer expectations.

Applying AI at the tactical level

The example of Amazon is a glimpse into the potential of DoD logistics in the not-too-distant future. Imagine a forward operating base automatically receiving the medical supplies it needs based on forecasted demand and anticipated casualties from simulated combat scenarios. In a contested logistics environment, push-packs and strategically pre-positioned materiel can be the difference between life and death. This is not an unrealistic picture of AI-driven defense logistics capabilities.

Here are just a few ways AI can provide concrete value from the supply depot to headquarters and everywhere in between:

Inventory Right-Sizing

Machine learning models can precisely predict supply and demand based on historical consumption and real-time analytics. This reduces carrying costs and minimizes the risk of stockouts or wastage. For perishable goods like Class VIIIa pharmaceuticals or Class I Rations & Subsistence, AI-driven inventory optimization minimizes spoilage—improving budget efficiency without sacrificing combat readiness.

Dynamic Ordering

With a better understanding of future needs, AI can dynamically adjust ordering patterns with or without operator involvement. ML models can take in thousands of data points, like National Stock Number (NSN) lead times, anticipated demand, expiration dates, contested logistics simulations—among a series of other relevant factors—and automatically place orders for replenishment to maximize readiness across all nodes of supply and any scenario. This means artillery can arrive at the front line before ammunition runs out, advances are stalled, and warfighters need to take defensive positions. 

Mobilization Planning

Machine learning can also be trained to analyze supply blocks based on composition, missing materiel, due-in materiel, and other pertinent considerations. Then the system will select and ship the best blocks that maximize readiness for a deployment. This reduces the risk of misallocation of resources and streamlines collaboration between warfighters and logisticians, resulting in faster, more effective mobilization from the warehouse to the tip of the spear.

Modern warfare requires modern logistics

Defense logistics has historically been bogged down by a lack of effective systems for prediction and execution. This has been a significant barrier to efficiency and effectiveness, but times are changing.

The security of our nation and our position in the global power rankings requires we not be complacent or accept how things have been done in the past. Machine learning is the key to unlocking next-generation defense logistics characterized by agility, precision, and strategic foresight. A sense of urgency, partnering with small businesses who can adapt quickly, and learning from the paths paved by giants like Amazon are all essential to stay ahead of hostile rivals in the AI technology arms race. Only then can we ensure the United States military remains the supreme fighting force around the world.

‍