Storage Systems for AI Workloads

Argonne National Laboratory, as a DOE contractor, is seeking a subcontractor to supply and integrate high-performance, scalable storage systems specifically optimized for AI and machine learning data pipelines. The contract is open for responses until August 21, 2026, with systems required to support the demanding I/O throughput, low latency, and massive concurrent access needs typical of AI workloads. Performance requirements emphasize scalability to handle exponential growth in training data, seamless integration with existing compute infrastructure, and robust data management capabilities including metadata indexing, tiered storage, and high availability. The place of performance is fixed at Lemont, Illinois, 60439, and the North American Industry Classification System code is 518210, indicating this is for data processing, hosting, and related services. The contract type is a subcontract, suggesting the winning vendor will work under a prime contractor managing broader DOE objectives. There is no set-aside designation, meaning all eligible businesses may compete, and the solicitation is posted with no designated point of contact listed publicly.