When a frontier model release changes what the mission can do, the people responsible for fielding, funding, and governing that capability need an answer they can act on — fast, technically credible, and independent of any commercial interest.
Three lines of delivery, producing critical work.
01 Briefings & Analysis
Direct engagement, classified-ready
Closed-door briefings for program offices, committee staff, and IC mission owners. Independent written analysis on technical questions where the answer benefits from an outside view.
02 Independent Model Evaluations
Built for acquisition and T&E
Capability assessments scoped to specific national security use cases. Oriented to existing federal T&E and acquisition processes. Results that travel from the bench to the program office without losing technical fidelity.
03 Neutral-Ground Convenings
Closed-door, on-the-record only by agreement
Facilitated dialogue between frontier labs and federal decision-makers. Editorial standards on both sides. Discretion on our partners' terms.
"AI, as a national security technology, now lives entirely inside the commercial sector — a condition without modern precedent."
A condition without modern precedent
Unlike past emerging technologies, the U.S. government does not have a monopoly on AI. Frontier capability is built in private labs, on private compute, by private workforces — and it is moving faster than any acquisition, oversight, or doctrine cycle can absorb on its own. Without a working channel between the labs producing the capability and the institutions responsible for fielding it, every model release widens the gap. The cost is paid in doctrine that falls behind, acquisition pathways that stall, and the opportunity for adversaries to co-opt American innovation.
The gap between frontier AI and the National Security Enterprise is not a coordination problem. It is a structural condition. We bridge the gap and create understanding.
Trusted interpretation, on the record
Decisions of this scale benefit from trusted interpretation. Our leadership has served at the senior levels of the Uniformed Services, Cabinet departments, the Intelligence Community, and on the Hill — the standing that allows findings to travel into doctrine, acquisition, and operations on their merits.
A protected channel between counterparties
Frontier labs cannot speak publicly about competitor capabilities, and they cannot easily exchange findings with one another. We provide a protected channel — a neutral venue where labs and government can share, evaluate, and act on what they know, with attribution on our partners' terms.
Federal reach at frontier-lab tempo
Matching the technical depth, velocity, and volume of frontier model output is more than any single institution can do alone. We add capacity to the work: aggregating across the labs, translating across the public–private divide, and synthesizing in the forms federal decision-makers can use on the timelines the mission actually runs on.
Operating model
Each program runs independently. Each is scoped to a decision the National Security Enterprise has to make. All four draw on the same engine: aggregation across the labs, translation across the public–private gap, synthesis for the field.
01
With the frontier labsSecuring American AI capability where it is built — the systems, the weights, and the people behind them. Partnership with frontier labs to harden against co-option, theft, and adversarial exploitation. The first line of defense for U.S. AI advantage.
02
Pentagon-facingA framework for evaluating frontier models against the specific mission decisions the NSE has to make. Oriented to existing federal acquisition and T&E vehicles; scoped to the tempo program offices operate on.
03
Foundational researchThe first sustained research program on how frontier AI is reshaping the strategic relationships between nuclear powers — and what stability and deterrence mean when capability cycles outpace policy cycles.
04
Federal cohortCurriculum, structured war games, and closed-door convenings that close the AI knowledge gap inside the National Security Enterprise. Includes a federal cohort program co-developed with higher education on AI policy, not vendor tooling.
"AI adoption and U.S. strategic advantage are mutually reinforcing — not competing. The hard part is doctrine, acquisition, and operations. That is where we work."
Briefings, evaluations, scenario work, congressional testimony, classified-context engagement — if your next decision needs an independent voice fluent in both the lab and the field, that is our work.
