AI Featured

The Pentagon Just Made Palantir the Operating System of the US Military

The March 9 memo from Deputy SecDef Steve Feinberg changed everything. Maven is now a program of record — permanent, fully funded, embedded in US warfighting doctrine. Here's what that means.

Alpha One

24 Mar 2026 • 10 min read

The March 9 memo from Deputy SecDef Steve Feinberg changed everything. Maven is now a program of record — permanent, fully funded, embedded in US warfighting doctrine. Here's what that means.

On March 9, 2026, Deputy Secretary of Defense Steve Feinberg signed a memo. It went to senior Pentagon leaders and US military commanders across every branch. Buried in the bureaucratic language was a phrase that carries enormous weight in the defence acquisition world: program of record. Six words. Everything changes.

What Feinberg's memo established — made public on March 20 — is that Palantir Technologies' Maven Smart System, the AI platform that has been processing battlefield intelligence and assisting in targeting operations across the US military, is no longer a pilot, a prototype, or an experimental initiative. It is now permanent infrastructure. And in Washington, that distinction is everything.

What Is a Program of Record?

If you don't live inside the Pentagon's acquisition bureaucracy, the phrase "program of record" might sound like paperwork. It isn't. It's a declaration of permanence.

In Department of Defense terms, a program of record (PoR) is a formally approved acquisition program with stable, multi-year funding written directly into DoD budgets. It has congressional visibility. It has dedicated programme managers. It has continuity that survives administration changes, budget battles, and the Washington political cycle. It is infrastructure — the same way GPS is infrastructure, the way the F-35 is infrastructure, the way AEGIS missile defence is infrastructure.

Once something becomes a program of record, it doesn't go away. It gets bigger.

The comparison to GPS is worth sitting with. When GPS became embedded in US military doctrine in the 1980s, it wasn't obvious that it would eventually reshape every sector of the global economy — agriculture, logistics, finance, navigation. What was obvious was that the military's commitment to it made it real in ways that civilian or ad-hoc systems could never be. Maven's PoR designation does the same thing for battlefield AI. It institutionalises it. It makes it doctrine.

What Maven Actually Does

The Maven Smart System is not a chatbot. It is not a search engine. It is a command-and-control AI platform that integrates data from over 179 sources — satellite imagery, drone feeds, radar returns, intercepted communications, signals intelligence, human intelligence reports — and synthesises it into actionable targeting intelligence in near real-time.

In practical terms: Maven identifies enemy vehicles, weapons stockpiles, military infrastructure, and personnel. It prioritises targets. It generates strike recommendations. It compresses what used to take weeks of analyst work into hours. Humans retain final authority over lethal decisions — but Maven does the cognitive heavy lifting.

By 2024, the system had more than 10,000 users across the Department of Defense. It was deployed across combatant commands including CENTCOM. During the 2026 US-Israeli military operations against Iran — Operation Epic Fury — Maven reportedly enabled US forces to strike approximately 1,000 targets in the first 24 hours of the campaign. Pentagon Chief Digital and AI Officer Cameron Stanley publicly credited the system with compressing target selection and strike timelines from weeks to hours.

The origin story matters. Project Maven launched in April 2017 under the Obama-era DoD as the Algorithmic Warfare Cross-Functional Team. Google was the original AI contractor — providing its TensorFlow framework for drone imagery analysis under a $9 million pilot. Then, in early 2018, more than 3,100 Google employees signed an open letter to CEO Sundar Pichai demanding the company cancel the contract. About a dozen engineers resigned outright. Google walked away in June 2018.

Palantir stepped into the void. By late 2019, it had assumed lead responsibility for Maven's AI and ML capabilities. It never looked back.

The Memo — What It Says

Feinberg's March 9 memo is worth reading closely, because it isn't just administrative restructuring. It's a strategic statement.

Key actions mandated by the memo:

Maven designated as DoD program of record by the end of FY2026 (September 2026)
Oversight transfers from the National Geospatial-Intelligence Agency (NGA) to the Pentagon's Chief Digital and Artificial Intelligence Office (CDAO) — within 30 days
Future contracting for Maven managed through the US Army
Scope: all military branches, all combatant commands

The most important language in the memo is Feinberg's direct statement of doctrine: "It is imperative that we invest now and with focus to deepen the integration of artificial intelligence across the Joint Force and establish AI-enabled decision-making as the cornerstone of our strategy."

Read that again. The cornerstone of our strategy. Not a tool. Not a capability. The cornerstone.

The memo also commits to embedding Maven to provide warfighters "with the latest tools necessary to detect, deter, and dominate our adversaries in all domains." That language — detect, deter, dominate — is not accident. It is a deliberate echo of the language US military doctrine uses to describe core warfighting objectives.

This is not a technology announcement. This is a declaration of how the US military intends to fight wars going forward.

The Money

Let's follow the dollars, because that's where the signal is strongest for investors.

The Maven Smart System contract timeline tells a clear escalation story:

2024: Initial $480M prototype contract awarded from the Army (Aberdeen Proving Ground)
May 2025: Expanded by $795M to a total ceiling of approximately $1.3 billion through 2029, reflecting surging operational demand
July 2025: US Army awards Palantir a separate 10-year Enterprise Service Agreement with a ceiling of up to $10 billion — consolidating 75 prior contracts into a single flexible framework
October 2025: CDAO awards Palantir a $483.5M task order under the new Army ESA

Program of record status doesn't announce a new contract number. What it does is far more valuable: it ensures stable, multi-year budget allocations that are written into DoD funding cycles and visible to Congress. It eliminates the constant contract renewal uncertainty. It transitions Maven from "big contract that could be cut" to "line item in the defence budget."

Palantir's government contract backlog now exceeds $11 billion. The company is guiding for roughly 61% revenue growth in 2026. The PoR designation doesn't change those numbers overnight — but it changes the quality of those numbers permanently.

What this means for $PLTR specifically:

Palantir currently trades more like a high-multiple SaaS company than a traditional defence contractor. That's unusual — and up to this point, it carried real risk. SaaS-style multiples require SaaS-style revenue visibility and recurring contract certainty. Government contracts have historically been lumpy, re-competed, and politically vulnerable.

Program of record status changes that calculus. It is the closest thing the Pentagon has to a long-term recurring revenue commitment. It is the moment $PLTR transitions from "controversial AI company with large government contracts" to infrastructure — a category that commands a different kind of institutional confidence.

The analogy that fits best: when Amazon Web Services secured the CIA's C2E cloud contract in 2013 and then the broader JEDI framework, it wasn't just a big revenue event. It was the moment the market understood that cloud infrastructure was foundational to US national security — and that AWS had won that lane. Palantir is the AWS of the kill chain. Today's memo is its JEDI moment.

The Competitive Landscape

A question worth asking: why couldn't anyone else take this?

The defence AI sector in 2026 is well-funded and crowded. Anduril Industries, now targeting a $60 billion valuation in a fresh funding round, is building a formidable hardware-software stack — autonomous drones, the Lattice AI platform, and a $20 billion Army contract of its own. Shield AI is developing AI-piloted aircraft, currently in talks to raise $1 billion at a $12 billion valuation. Legacy primes like L3Harris, Northrop Grumman, and Lockheed Martin are all running AI programs of their own, with L3Harris actively partnering with Palantir on AI-enabled production ramp-ups.

But here is the critical distinction: no one else has Maven's battle-tested deployment record at scale.

Anduril plays in a different lane — autonomous systems and hardware integration. Shield AI is pursuing autonomous flight. Neither is attempting to replace the kind of multi-domain intelligence fusion and targeting workflow that Maven provides. The legacy primes are being disrupted, not replaced — they remain essential for platforms, manufacturing, and systems integration, but they are not competing for Maven's role.

The moat Palantir has built is not just technological. It is operational. Maven has been tested in real combat against a real adversary under real conditions. It has processed real targeting data. It has assisted in real strikes. In the defence procurement world, there is no substitute for that kind of operational track record. It is extraordinarily difficult to replicate quickly.

The Anthropic Wrinkle

Here is the story within the story, and it deserves careful attention.

Maven integrates Anthropic's Claude AI for certain intelligence analysis and decision-support functions. Palantir partnered with Anthropic in 2024, alongside AWS, to bring Claude's language capabilities into military workflows.

In early March 2026 — just days before Feinberg signed the Maven PoR memo — Defense Secretary Pete Hegseth formally designated Anthropic as a supply chain risk to the Department of Defense. The designation, unprecedented for a US-based company, stems from a prolonged dispute in which the Pentagon demanded unfettered access to Claude for "all lawful purposes" while Anthropic insisted on AI safety guardrails prohibiting autonomous lethal systems and mass domestic surveillance.

Pentagon CTO Emil Michael called Claude a potential "pollutant" in the defence supply chain. A six-month phase-out has been ordered, though deep integration means full removal will take longer. Anthropic is reportedly planning legal action.

Palantir CEO Alex Karp addressed this directly on March 12: "The Department of War is planning to phase out Anthropic; currently, it's not phased out... Our products are integrated with Anthropic."

For investors, the Anthropic situation is a near-term operational wrinkle, not a structural threat. Palantir has modular integration architecture — Claude can be substituted with other LLMs as the transition progresses. What the episode reveals more broadly is a fault line that will define the defence tech sector for years: the tension between AI safety orthodoxy and military operational requirements. Palantir sits unambiguously on one side of that fault line. It does not moralize about its technology. It builds tools that work and lets its clients decide how to use them. That posture is an explicit competitive advantage in the Pentagon's current environment.

What America's Adversaries Should Know

The Maven PoR designation is not happening in a vacuum. It is happening in a world where AI-enabled warfare is actively being tested in live combat operations, and where adversaries are watching carefully.

China is the primary long-horizon threat. The People's Liberation Army has been running its own battlefield AI programs, and Chinese military doctrine has explicitly incorporated AI-assisted decision-making since at least 2017. PLA development of AI targeting systems — including reported programs in drone swarms and intelligence fusion — represents the most credible long-term challenge to US AI battlefield dominance. The gap matters, and it is not static.

Russia presents a different picture. The Ukraine war has severely degraded Russian conventional military capabilities, consumed significant material resources, and exposed deep weaknesses in command-and-control integration. Russia is far behind on operational AI battlefield deployment relative to its pre-war aspirations.

Iran is the most immediate case study. Iranian forces have already experienced the sharp end of Maven-assisted targeting. The scale and speed of the US-Israeli strike campaign against Iranian targets in early 2026 — 1,000 targets in 24 hours — represents a qualitative leap in operational tempo that Iranian air defences and dispersal strategies were not designed to counter.

The broader strategic message is straightforward: the United States has committed to AI as the core of its warfighting doctrine, it has a battle-tested platform deployed at scale, and it has just institutionalised that platform at the highest level of acquisition permanence. The gap between US AI-enabled warfare capability and its nearest adversaries is widening, not narrowing.

The Investment Angle

Let's be direct about what the Maven program of record designation means for $PLTR as an investment.

Palantir currently trades at approximately $150 per share with a market cap north of $350 billion — roughly 45x forward 2026 revenue. By any traditional defence contractor metric, that is a stratospheric multiple. RTX, Northrop, Lockheed — these names trade at single-digit revenue multiples. The market has been pricing Palantir as a software company, not a defence prime. Until recently, that carried genuine risk.

The risk is not gone. But it has fundamentally changed character.

The bull case in brief:

Program of record status means Palantir's US government revenue — which represents approximately 60% of total revenue — now has a different kind of permanence attached to it. It is institutionalised in DoD budget cycles. It has congressional visibility. It is structurally resistant to the kind of contract disruption that defence software companies typically face. The Army's $10 billion ESA, combined with the PoR designation for Maven, means Palantir has built something closer to a natural monopoly in AI battlefield management software than any competitor.

The moat is real: nine years of data integration, battle-tested operational credibility, 10,000+ DoD users already trained on the platform, and now the full weight of Pentagon doctrine behind it. Switching costs in this domain are extraordinary.

The risks worth taking seriously:

Political: A future administration more sceptical of AI in warfare could de-emphasise Maven's role, though the PoR designation makes outright cancellation substantially harder
Technical: Adversary AI capabilities — particularly Chinese battlefield AI — could erode the US tactical advantage over a multi-year horizon
Legal/Ethical: The Anthropic situation signals broader tensions; any high-profile Maven-related targeting incident involving civilian casualties (and the Iran girls' school incident already under Pentagon review) creates political exposure
Valuation: At 45x forward revenue, the good news is priced in. The upside from here is execution-dependent, not announcement-dependent

The honest framing for sophisticated investors: Maven's PoR status doesn't necessarily make $PLTR cheap. It makes the existing multiple more defensible than it was a week ago. Revenue visibility improved. Moat deepened. Downside risk on the government business meaningfully reduced. Whether the current valuation is attractive depends on your view of Palantir's commercial AI business trajectory and whether government remains the bedrock or becomes a smaller share of a much larger whole.

The Long View

Step back from the ticker and the memo and the contract numbers for a moment.

What happened on March 9, 2026 — formalised publicly on March 20 — is that the United States military made a decision about how it intends to fight future wars. Not a provisional decision. Not a pilot. A doctrinal commitment, written into acquisition law, backed by multi-year funding, spanning every branch of the armed forces.

AI-enabled decision-making is now, in Feinberg's own words, the cornerstone of Joint Force strategy. That is a statement of such significance that historians of warfare will likely mark it alongside the introduction of carrier aviation, stealth technology, or precision-guided munitions as a doctrinal turning point.

Maven is the platform. Palantir is the company that won.

That matters for investors calculating revenue multiples. It matters for geopolitical analysts modelling great-power competition. It matters for defence procurement professionals watching the industry restructure in real time. And it matters for anyone trying to understand what 21st-century warfare actually looks like — not in think-tank papers, but in operational reality.

The question now is not whether AI will define how America fights. That question has been answered. The question is how fast the advantage compounds — and whether anyone else in the world can close the gap before it becomes permanent.

Based on the trajectory of the past 90 days, the answer looks increasingly like: probably not.

AlphaBriefing covers the intersection of geopolitics and markets for investors and strategically-minded professionals. Follow @TheAlphaBrief on X.

Sources & Further Reading