SOFWERX, in collaboration with USSOCOM J24 Intelligence Data Science Team (IDST), will host an Assessment Event (AE) 06-08 January 2026, to determine the best solution to upgrade a remote location with high-performance Graphics Processing Unit (GPU) servers to support large language model (LLM) workloads for up to 100+ concurrent users.
The U.S. Government’s data science portfolio is rapidly expanding its reliance on largescale AI workloads, especially LLMs and highspeed inference pipelines. To sustain this growth and to maintain a strategic edge, the program requires cutting-edge GPU acceleration, capable of delivering the throughput and memory bandwidth needed for state-of-the-art training, finetuning, and deployment. Advanced GPUs will provide a high-performance, energy efficient, and?future ready?foundation for advanced AI workloads, while ensuring low response times, reliability, and room for future growth. 
The goal of this assessment is to acquire, deliver, and install/deploy advanced GPU hardware that will add a high bandwidth, energy efficient GPU capability that is immediately ready to power LLM inference, finetuning, and Retrieval-Augmented Generation (RAG) workloads.  
The GPUs must be delivered as part of a complete, rack-mounted server solution suitable for immediate deployment in the data center at a remote site. The server must include all necessary hardware to operate the GPUs safely and efficiently, including redundant power supplies, appropriate power cables, connectors, and any transfer switches required to support high-availability operation. 
High-speed Peripheral Component Interconnect express-based (PCIe-based) network connectivity must be included, with all necessary cables and interconnects to support GPU-to-GPU communication within the server and connectivity to the site’s broader network. This includes GPU bridges (NVLink or equivalent), network cables, and any peripheral connections required for remote management and monitoring. Storage and memory subsystems must be pre-installed and connected, providing sufficient RAM and Non-Volatile Memory express (NVMe) SSD storage to support large AI models and high-throughput workloads. 
Servers must be pre-configured or delivered with the ability to quickly install necessary software, including firmware, drivers, and GPU libraries, so that the system can be operational for AI workloads immediately upon installation. All components, cabling, and connections must conform to enterprise data center standards and integrate seamlessly with the site’s existing power, cooling, and network infrastructure. Vendors must provide a turnkey solution that minimizes on-site assembly, configuration, and troubleshooting, ensuring the server is ready for immediate use with minimal IT intervention.  
The solution must support deployment to two networks, within air-gapped or otherwise strictly isolated environments. The server(s) and all GPUs shall be physically and logically isolated from each other, and any non-approved networks (no dual-homed network connections).
Submit NLT 09 December 2025 11:59 PM ET
U.S. Citizens Only
To learn more, visit events.sofwerx.org/hardware-enabled-ai-acceleration.