Hardware Enabled AI Acceleration

Assessment Event (AE)

06-08 January 2026

Host Name: Sarah Marks

Location: SOFWERX

Share on Social Media

Submit NLT 09 December 2025 11:59 PM ET
U.S. Citizens Only

Purpose
SOFWERX, in collaboration with USSOCOM J24 Intelligence Data Science Team (IDST), will host an Assessment Event (AE) 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 system will be fully turnkey, including all GPUs, memory, storage, networking, cooling, and power infrastructure, ready for immediate operation. This configuration ensures fast, reliable Artificial Intelligence (AI) performance today and scalability for future growth. 

Background/Synopsis
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. 

Objective
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).

Why You Should Participate
USSOCOM seeks to enter into non-FAR or FAR-based agreements with Industry, Academic, and National Lab partners whose solutions are favorably evaluated by USSOCOM J24 IDST Subject Matter Experts (SMEs). As such, the follow-on Assessment Event (AE) is considered competitive and solutions will be evaluated independently of one another primarily for technical merit. This serves dually as notification of the intent to research the feasibility of an agreement under 10 U.S. Code, Section 4022 and/or Section 4022(f), and as notice of pre-solicitation activities IAW FAR 5.204.

Event Timeline
Phase 1 - 14 November 2025 to 09 December 2025 Submissions to the Assessment Event (AE) Open: Interested respondents who could potentially provide solutions that meet the needs of USSOCOM are encouraged to submit their capability for USSOCOM review. To submit your technology for potential Phase 3 attendance, please review the assessment criteria and follow the submission instructions at the bottom of this webpage.

Phase 1a - 03 December 2025 Q&A Telecon: Interested offerors may participate in a virtual Q&A session to better understand the PM's specific technology objectives. The telecon will take place on 03 December from 1:00–2:00 PM ET. RSVP at the bottom of this webpage.

Phase 2 - 10 December 2025 to 19 December 2025 Downselect: USSOCOM will downselect those respondents/submissions they feel have the highest potential to satisfy their technology needs. Favorably evaluated submissions will receive an invitation to attend the AE on/around 19 December.

Phase 3 - 06 January 2026 to 08 January 2026 Assessment Event (AE): During the AE, selected participants will be allotted a one-on-one session with the USSOCOM evaluation panel to pitch, demonstrate, and/or discuss their solutions. The forum will include a Q&A portion and discussions may continue outside of the event. Solution brief presentation guidelines will be outlined in the event invitation and solutions will be assessed according to the criteria in the link provided below. If the USSOCOM evaluation panel favorably evaluates a solution brief, negotiations for Phase 4 may immediately begin.

This event will be held in person at SOFWERX and will have a virtual component. 

Phase 4 - Path Forward: Successfully negotiated awards may fall under any combination of these categories:

NOTE: Awardees may need to be compliant with NIST SP 800-171, Protecting Controlled Unclassified Information (CUI) in Non-Federal Systems and Organizations

How You Can Participate
Government, Industry, Academia, National Labs, and Non-Traditional Attendees

  1. RSVP to the 03 December Virtual Q&A Telecon: Click Here to RSVP (NLT 02 December 11:59 PM ET)
  2. Review Assessment Criteria: Click Here to Review
  3. Download the White Paper Template: Click Here to Review
  4. Complete the Submission Form: Click Here to Submit (NLT 09 December 11:59 PM ET)

Questions?

For event-related questions, please contact Sarah Marks


© 2015 - 2025 SOFWERX, Inc. All rights reserved.

SOFWERX Logo
DEFENSEWERX Logo