Four U.S. combatant commands have now confirmed on the record that they received Deputy Secretary of Defense Steve Feinberg's March 9 directive: make Palantir's Maven Smart System an official Department of Defense program of record by September 30, 2026. Northern Command, Transportation Command, Special Operations Command, and Space Command all told DefenseScoop they are moving. The seven remaining combatant commands declined to comment, citing operational security. That silence itself is revealing — it suggests the Pentagon is already treating Maven as classified infrastructure, not an experimental platform.
What this actually means is the Pentagon is converting a tool that started as a 2017 imagery-processing experiment into the backbone of joint military command and control. Maven has grown from a proof-of-concept at five combatant commands in May 2024 to active deployment across all eleven unified combatant commands plus the Joint Staff, with 20,000 active users as of March 2026 — a four-fold increase in two years. Feinberg's memo is not announcing a new capability. It is formally committing the Pentagon to fund and defend a platform that is already operational everywhere that matters. The geospatial portion of Maven was already designated a program of record under the National Geospatial-Intelligence Agency in 2023. This directive expands that protection across the entire military enterprise, shifting management from NGA to the Chief Digital and AI Officer, and shifting all contracting to the U.S. Army's Enterprise Agreement vehicle — a single contract that consolidated 75 separate Maven-related agreements into one procurement line.
The numbers tell you how big this has become. The initial Maven contract, signed in May 2024, was a $480 million, five-year IDIQ. By May 2025 — one year later — the Pentagon had raised the ceiling to $1.3 billion through 2029. The Army Enterprise Agreement backstop is up to $10 billion. All future Maven contracting flows through that Army vehicle, not through individual combatant command procurement channels or the NGA. Feinberg's memo orders the CDAO MSS Program Office to assume system administration, oversight, and support activities within 30 days. The transition to full program-of-record status — with Maven protected as a line item in the Future Years Defense Program — must be complete by September 30, 2026. One NORAD/NORTHCOM official told DefenseScoop: 'We are actively engaged in the transition planning and anticipate no challenges or disruptions to current operations.' That statement is either confidence or a prepared line. The timeline feels like the latter.
What created this moment is not a new technological breakthrough. It is a policy decision to operationalize something that was already working. Feinberg wrote that Maven can now generate 1,000 targeting recommendations per hour. By June 2026, according to NGA Director Vice Admiral Frank Whitworth, Maven will begin transmitting 100 percent machine-generated intelligence to combatant commanders — fully autonomous machine output, no human hands in the workflow. That is not a capability the Pentagon expects to have. That is a capability the Pentagon is already testing. The aggressive timeline exists because the second Trump administration wants to accelerate the deployment of AI across military operations. Katrina Manson, journalist and author, told DefenseScoop: 'In a sense, it's things that were already happening, but clearly this Pentagon wants to go a lot faster and is very aggressive about that.' The directive is bureaucratic formalization of operational reality, moving fast to lock it in place before institutional resistance or Congressional scrutiny can slow it down.
Who wins here is clear: Palantir. Program-of-record status means Maven enters the Pentagon's protected budget line. That means Congress votes to defend Maven's funding, not to approve it on a year-by-year basis. It means 10 years of revenue visibility, not five-year contract cycles. It means the Army owns the contract management, which simplifies procurement and removes some of the experimental flavor. A former senior defense official told DefenseScoop what everyone in the industry is already thinking: 'If [Maven] is brought under CJADC2 eventually, what will matter most will be whether Palantir tries to make this the platform of platforms. That would lock them in for a decade.' The answer is yes, Palantir is already trying to do exactly that. Maven is designed to fuse data from multiple sensors into a single C2 picture. Making it the official, formal program of record is how you make it architecture, not just a tool. Every competing command-and-control vendor — Anduril, Sarcos, Shield AI, and the traditional primes who have their own C2 offerings — just inherited a subordinate position in the military's data architecture. They will integrate into Maven or they will not be integrated at all.
What gets lost in the policy language is what Emelia Probasco, senior fellow at Georgetown's Center for Security and Emerging Technology, actually said: 'Maven is popular because people like it.' The military is not voting on Palantir because the Pentagon forced it. The combatant commands are adopting Maven because their commanders prefer it. That is strategically important because it means this is not a classic defense prime lock-in play where a incumbent vendor protects a contract through bureaucratic moat-building. This is a commercial software company whose product is genuinely useful in the military operational environment. The distinction matters because it means the Pentagon's move to formalize Maven is not reversible. Program-of-record status does not create dependency. It documents it. The dependency already exists. What the September 30 deadline actually does is prevent the Pentagon from accidentally fragmenting Maven adoption across different combatant commands with different contracting terms. The POR designation locks in a single contracting vehicle, single oversight structure, and single technical architecture. That is harder to change than an OTA contract or an IDIQ ceiling. But it is not harder to change because Palantir lobbied for it. It is harder to change because the military voted with its behavior first, and the Pentagon is now making that behavior official.
Watch three things before September 30. First, the CDAO MSS Program Office standup and the completion of transition of all MSS contracts to the Army EA vehicle. That is the structural moment that locks Maven into Army infrastructure. Second, watch whether the Pentagon's integration of Anthropic's Claude AI model survives the supply-chain scrutiny that has been building since early March 2026. Anthropic was flagged as a supply-chain risk after refusing to restrict domestic surveillance and autonomous lethal autonomous weapons system restrictions. If the Pentagon has to swap Claude out for a different LLM before June, that June milestone of 100 percent machine-generated intelligence becomes harder to hit. Third, watch the seven silent combatant commands. CENTCOM, EUCOM, and INDOPACOM are actively deployed in combat zones where Maven's targeting recommendations matter operationally. If any of them break silence to say the transition is disrupting operations, that becomes a Congressional lever for anyone who wants to slow down the timeline. None of those three look likely to do that. But silence about Maven is different from silence about other military programs. Maven is not a weapons system. It is invisible intelligence infrastructure. When combatant commands decline to confirm they received the transition memo, citing operational security, what they are saying is Maven is now classified. That is a meaningful step forward for Palantir's position.