Lockheed Martin Wins HERMES Contract to Manage Massive Data Sets for Army AI Research — a federal contract (USAspending)

The performance of AI and machine learning (ML, a technique by which computers learn patterns from data) hinges largely on how good the data is and how efficiently it can be handled. Training and validation require enormous volumes of data, and without a solid foundation to organize, store, and retrieve it, even the most advanced model cannot reach its potential. The HERMES contract that Lockheed Martin won focuses on exactly this foundation: managing the massive data sets that underpin research. Its emphasis is less on a flashy deliverable and more on building the infrastructure that supports the research itself.

The awarding body, the U.S. Army Research Laboratory (ARL), conducts the Army's basic and applied research and serves as a hub for defense-related AI work. This contract was made through a Broad Agency Announcement (BAA), a mechanism in which the government, rather than specifying a finished product in detail, outlines a broad research theme, invites a wide range of proposals, and funds the promising ones. It suits research and development where the answers are not yet settled. The relatively long period of up to five years reflects the nature of a research contract meant to be pursued steadily rather than rushed.

Viewed across the broader landscape, this contract reflects how recent government procurement is investing not only in AI models themselves but in the data foundations that support them. Organizing and streamlining data is an unglamorous, behind-the-scenes domain, yet it is a decisive factor in making AI practical and is a shared challenge across defense, government, and research. Because the contract summary does not describe exactly what functions or outcomes HERMES will produce, this article confines itself to explaining the institutional and technical background.