American defense contractor Palantir is at the center of U.S. and international initiatives and controversies.
Major Palantir-related Developments
(1) Palantir is the prime contractor for the U.S. Maven Smart System (MSS), which is now a centerpiece of U.S. CJADC2 efforts, having publicly broken through in FY26.
(2) The use of MSS in U.S. operations in Iran is highly controversial, associated with “AI-driven targeting” and the mistaken bombing of a school in Minab.
(3) NATO acquired its own variant of MSS in 2025, for which Palantir is also the prime contractor. The system achieved full technical operational capability in June 2026.
(4) Palantir is associated with a recent commercial push for access to a mixture of open- and closed-weight models.
Post-U.S. Midterm Takeaways
(1) Much hinges on whether the incoming U.S. Congress prioritizes oversight of Defense Department contractors rather than DoD leadership. Both are possible, but political capital and time are limited. One should bet on the MSS itself continuing.
(2) Whether the incoming Congress attempts to force a release of the DoD’s internal review of the Minab school bombing will in part shape whether Palantir’s role in the Department is substantively altered.
(3) The rapid development of the NATO MSS via Palantir generates some path-dependency, and the alliance’s system appears likely to continue with Palantir for the time being, with the caveat that political winds may shift.
(4) As the U.S. Congress changes composition, coinciding trends among U.S. allies with respect to American technology reliance are becoming more pronounced. France, Germany, and the United Kingdom each have – to varying degrees and for somewhat different if overlapping reasons – encountered either a governmental rejection of Palantir’s role in operations or public backlash.
These trends may limit the extent to which certain software-centric command-and-control (C2) functions are interoperable between the U.S. and some of its allies, which could trigger a re-evaluation of Palantir’s role. However, working in Palantir’s favor, the Japanese Self-Defense Forces (SDF) are reportedly open to deploying Palantir’s technology within their own operations. Japan has independent and particularly acute interests in maintaining interoperability with the U.S.
(5) Palantir officials practice considerable messaging discipline in public forums, potentially benefitting the company’s position in a new political environment. However, the company will find its relationship with U.S. Immigration and Customers Enforcement (ICE) particularly difficult to extricate from its political image.
Overview
2026 was the breakthrough year for the U.S. Maven Smart System (MSS). MSS is an outgrowth of the U.S. Department of Defense’s (DoD) Project Maven that seeks the use of rapid automated analyses of vast swaths of data, in part but not only by the use of artificial intelligence (AI) models. The system is, in short, the DoD’s means of consolidating, streamlining, and providing access to vast databases and real-time information, complete with subcomponents responsible for automated analyses over that data (e.g., geospatial intelligence analysis).
American technology firm Palantir is the prime contractor for the MSS (following Google’s protest-driven withdrawal from Project Maven in 2017). Per an internal DoD memo from Deputy Secretary of Defense Steven Feinberg earlier this year, MSS was to be designated a program of record by the end of FY26 in September.
Over the past year (give or take), Palantir and the MSS have been front-and-center in both an ongoing buildout of defense computing systems in the U.S. and abroad as well as a string of controversies over the use of AI in U.S. defense operations. This encompasses the company’s political positioning within the United States and the knock-on effects of the company’s influence in the governmental operations of states including France, Germany, and the United Kingdom.
At the same time, there are opportunities for Palantir in states including Japan, whose Self-Defense Forces are reportedly open to deploying the firm’s technology and the path-dependency generated by the company’s existing contracts and relationships with the U.S. DoD and NATO. (The U.S. Army indeed bundled existing contracts with Palantir last year into a $10 billion, 10-year award.)
Palantir’s role at the center of this broad spate of efforts is worth the focus of defense analysts and other contractors. The company presents something of a case study for the opportunities and risks associated with the defense computing revolution.
The results of the U.S. midterm elections puts Palantir into new focus, as the risks and opportunities for the company become more pronounced. This piece assesses Palantir’s role in U.S. DoD and international operations and lays out plausible outcomes.
MSS in Iran and the School Bombing
U.S. Central Command (CENTCOM) publicly provided figures on targeting related to the U.S.’s Operation Epic Fury in Iran that began in February of this year. CENTCOM indicated that over 1,000 targets were selected and struck by U.S. forces in the first 24 hours of the war. By April 6th, CENTCOM noted that number had risen to over 13,000 targets struck. The use of “AI” tools was alluded to in this respect, and it was later revealed that MSS usage surged during this period within the DoD, with personnel exhausting roughly twenty billion tokens per day at the peak of Epic Fury (tokens are digital representations of input and output data).
The initial salvo saw the mistaken bombing of an elementary school in Minab, a city in Southern Iran, killing at least 175 people. A preliminary internal DoD review of the strike, as reported initially by the New York Times, suggested that the Defense Intelligence Agency handed off outdated geospatial intelligence to CENTCOM that designated the school campus as a military base in error.
A string of Bloomberg reports paints a clearer picture: the school bombing was the result of an accumulation of errors beginning nearly a decade ago. Although an analyst input a correction in 2019 to the relevant intelligence database indicating that the military base belonging to the Islamic Revolutionary Guard Corps had been divided into a separate school campus in Minab, this correction did not make its way into data used to generate target options in the run-up to Operation Epic Fury. Additionally, amid the rush to generate this target list in February (after talks had collapsed between the U.S. and Iran), it was reported that civilian harm mitigation (CHM) staff – whose numbers had been previously cut in layoffs directed by Pentagon leadership – were deliberately boxed out of planning for Epic Fury. Interestingly, that same report indicated that some CENTCOM personnel believed the MSS would automatically flag outdated or inconsistent intelligence as targets were being compiled. It did not, though it is unclear it was designed for this purpose and why personnel believed it was.
The internal DoD review of the strike remains unpublished, as of writing.
Palantir and Wary U.S. Allies
Palantir has been a flashpoint in efforts by foreign governments to achieve (some conception of) digital sovereignty and reduce reliance on American technology firms.
In April, head of German cyber defense Thomas Daum publicly stated that the German armed forces would not award contracts to Palantir, noting that it would be “simply inconceivable at the moment to grant industry staff access to the national database.”
Shortly after, in June, French intelligence agency DGSI reportedly opted to excise Palantir’s services from its work, with France-based ChapsVision selected as their replacement.
The United Kingdom, for its part, contracted Palantir to streamline operations within its National Health Service (NHS), a decision which has led to significant public backlash. British publication The Economist convened perspectives on Palantir’s role in the U.K. government in August. This controversy is ongoing.
More conspicuously, the French armed forces are developing an indigenous version of the MSS that deliberately excludes participation by non-French firms. The system, dubbed “Arcadia,” is meant to be an equivalent of MSS with a distributed architecture.
To be sure, these developments do not occur in a vacuum. In the case of France especially, there are broader efforts to decouple sensitive government technology from reliance on American firms. For instance, the French government announced in January that it would cease its use of American-owned video conferencing software by year’s end, including Zoom and Microsoft Teams, in favor of the French government-owned platform Visio.
NATO MSS and “Kill Switch” Fears
In 2025, NATO acquired its own variant of the MSS, achieving full technical operational capability by June 2026. Palantir is the prime contractor for NATO MSS. NATO had also set up Task Force Maven by this point.
However, by October 2026, Palantir was at the center of a European fear: that NATO MSS may possess a built-in “kill switch” that could be flipped at the whim of American leadership. Palantir publicly pushed back against the idea, with company official Louis Mosley noting there is no such kill switch. The fear was not new, as private-sector organizations like the U.K.-based Future of Technology Institute released a report in April 2026 noting that the “threat of a ‘kill switch’ is no longer theoretical” given the Ministry of Defence’s reliance on American firms including Google, Oracle, Amazon Web Services (AWS), and Microsoft.
The next month, in May, Adm. Pierre Vandier, NATO Supreme Allied Commander Transformation, stated publicly: “As far as I know, today there is no real competitor for Palantir.”
Such fears are, in any event, a recent mainstay of U.S.-European relations, with the continent sporadically realizing its reliance on American technology firms (like Anthropic, for instance) comes with risks. Expect these fears to continue bubbling to the surface.
Palantir, Anthropic, and Open-Weight Models
Palantir’s leadership, including but not limited to CEO Alex Karp, has exhibited sensitivity to shifting political winds in public messaging over this period. For example, days after the Pentagon’s decision to slap a supply chain risk designation on American AI firm Anthropic – the first of the major Large Language Model labs to integrate its technology into classified DoD networks – Karp criticized (in no uncertain terms) the message that AI would eliminate white-collar jobs while also ‘screwing’ the U.S. military, a not-particularly-veiled response to comments by Anthropic CEO Dario Amodei.
Interestingly, Palantir likewise began positioning itself as a champion of open-weight AI models in June (which contrast with closed-weight, or proprietary models, provided by companies like Anthropic). The company released a statement on X promoting “AI sovereignty,” including the importance of data retention and a criticism of “tokenmaxxing,” or AI services that rely on high-token, and therefore high-cost, usage. Karp echoed this message in an interview noting the use of open-weight models to this end.
Palantir has a partnership with open-weight champion and chip designer NVIDIA that oversees the use of open-weight models for government and infrastructure customers, emphasizing many of the same themes about AI sovereignty on the product’s promotional page.
After the Midterms
How Palantir factors into ongoing U.S. DoD and international programs hinges in large part on political will, capital, priorities, and idiosyncratic state interests. The scale of the Democratic party’s victories makes the likelihood of pressure on the company in this respect, and a web of defense-related work that it touches, significantly greater.
Bet on the Continuation of U.S. MSS
Should newly elected leadership take aim at Palantir over the next two years, it is judged unlikely that the MSS itself will be deprioritized or decommissioned. The system is designed to alleviate a severe data-collection and processing bottleneck that has only grown since the inception of Project Maven.
It is unlikely that either the incoming Congress or the DoD itself will seek to reinvent the wheel in response to this bottleneck. It is similarly unlikely, should independent controversies surrounding Palantir come to the fore, that this will be coupled with the fate of the MSS. (Note, after all, that MSS was tested in real-world, operational conditions in Ukraine in 2022, under the Biden administration.) Bet on the MSS’s continuation.
Thus, any focus on Palantir specifically by the incoming Congress will find itself balancing the needs of continuing the development of MSS and the role of Palantir therein. Echoing a broader concern in U.S. defense literature about vendor lock-in, or the stickiness that a prime contractor in a particular program due to the use of their proprietary technology, there will doubtless be considerations about sustaining and expanding access to MSS in the event that Palantir were removed and replaced by another contractor (or, potentially, multiple contractors). Palantir, for its part, has taken pains to emphasize publicly that it opposes vendor lock-in.
Beyond vendor lock-in, a risk to Palantir’s role in the U.S. MSS hinges on how the incoming Democratic House members assess the role of Maven in the Iranian school bombing. If the bombing is seen as a failure generated directly by the system, members may seek to either substantively alter or remove Palantir from its role as the MSS prime contractor. Members could likewise push to have the internal DoD report on the bombing released publicly, building on calls to do so already.
In this author’s opinion, the Minab school bombing is the cumulative result of a series of interrelated errors tracing back nearly one decade ago within the DoD, beginning at least with the failure to migrate updated data in the Military Intelligence Integrated Database (MIDB) to the Machine-Assisted Analytic Rapid-Repository System (MARS). Where the MSS was involved appears to be a straightforward case of automation bias (i.e., personnel assuming, perhaps without good reason, that the MSS would automatically flag inconsistencies in underlying intelligence used to generate target options). Automation bias seems to have been worsened in the run-up to Operation Epic Fury for two reasons: first, the DoD rushed the targeting option generation. Second, the DoD cut civilian harm mitigation staff from the planning process for Epic Fury.
In any event, it is perhaps more likely that the new U.S. Congress will focus their pressure on DoD leadership, given the decisions made in the run-up to Epic Fury.
NATO MSS Generates Some Path-Dependency
The same bottleneck that drove the U.S.’s Project Maven confronts international defense organizations, too, including NATO. So intense is this bottleneck that U.S. Army Col. Arnel David – head of the NATO Task Force Maven – noted publicly in June 2026 that NATO’s acquisition of its own variant was motivated by the organization “drowning with data, but we’re starving for wisdom and strategic insight.”
The case for the persistence of NATO MSS along its current path is fairly bullish, particularly given that NATO and Palantir coordinated over the acquisition and achievement of full technical operational capability in a little over a single year.
The hesitancy of Germany and France in particular over the inclusion of Palantir in their sensitive defense operations will, to be sure, blunt its ultimate scope within NATO. Interoperability between American and European allies within NATO is a continuing and critical need, spanning far more than C2 functions handled by MSS-like systems. The French use of Arcadia and the German chilliness toward Palantir will present obstacles in this respect.
The United Kingdom, for its part, indicated recently that no decision has been made on Palantir’s involvement in the Ministry of Defence’s Digital Targeting Web program nor the country’s use of the NATO MSS.
That said, NATO MSS’s rapid transition from acquisition to certification generates some path-dependency. Indeed, in March 2026, NATO’s Joint Force Command in Norfolk, Virginia, reportedly planned on beginning its integration of MSS by May 2026, having already been installed at Allied Joint Force Command Brunssum in the Netherlands as well as Supreme Headquarters Allied Powers Europe (SHAPE) in Belgium. Palantir is deeply embedded in a process that is already well underway and having already gone some ways towards linking U.S. and NATO systems. This is likely to continue.
Kill Switch” Fears Depend on U.S. Alliances Strengths
The U.S. President exerts disproportionate control over American foreign policy. A Democratic House could, to be sure, exert significant pressure on the ultimate shape of this policy. U.S., particularly European, allies that are fearful of “kill switches” embedded in American technology will be influenced by just how much pressure the incoming Congress exerts on U.S. foreign policy decision-making across areas including tariffs and uses of force in or around Iran and the Strait of Hormuz.
So far as systems like NATO MSS is concerned, whether the incoming Congress finds ways to firm up American confidence in the alliance will matter for the continued integration of NATO MSS with member-states. “Kill switch” fears will likely abate in this scenario.
Japan’s Special Case
In August 2026, Japan’s Self-Defense Forces (SDF) were reportedly considering deploying the U.S.’s MSS (as well as Anduril’s Lattice platform), potentially mixed with Japanese AI models, for command-and-control (C2) functions. The deployment would serve needs over the next five years, including counter-attacks, missile defense, and drone control. A Japanese “command center” that would allow for the selection of specific AI models, American or Japanese, was likewise being considered by the SDF.
Japan represents a special case of relevance for Palantir. As Japan is not a member of NATO (albeit, with close NATO ties), the SDF must carve out relationships with allies like the United States on a bilateral or multilateral basis. Doing so is within the country’s interest given the uncertainties of a potential Chinese invasion of Taiwan in the coming years and the need for comprehensive defense coordination with U.S. forces operating in the region.
Thus, for Japan, a ‘kill switch’ fear may be diminished against this reality. The alternative risks a Japanese SDF operating with a fragmented, unclear operating picture in either the run-up to or during a conflict over Taiwan. Japan’s interests here indirectly aid Palantir, as the country is unlikely to insist on a substitution that grants it indigenous control over all relevant C2 systems.
Looking Ahead
In the wake of Google’s withdrawal from Project Maven in the 2010s, Palantir successfully positioned itself as the go-to firm for the ongoing defense technology boom primarily in the United States, but also increasingly – and rapidly – abroad. As the company took these steps, it became a major actor in the development and deployment of headline programs like the U.S. MSS and NATO MSS. At the same time, its ability to position itself in this way came with its involvement in both political- and defense-related controversies, not least of which is the association of the U.S. MSS with the Minab school bombing earlier this year.
Less noticed in defense circles, though somewhat overlapping, is Palantir’s relationship with Immigration and Customs Enforcement (ICE). The Electronic Frontier Foundation (EFF) accused Palantir in January 2026 of providing ICE with digital tools that aid in shady immigration practices. (The EFF followed up these claims in April 2026, whereas the ACLU and American Immigration Council made similar claims prior.) Palantir publicly and vigorously rejected the EFF’s claims in January 2026. Whatever the case, ICE is a flashpoint in U.S. national politics, such that it will find it difficult to maintain a positive image of its relationship with the agency in this political environment. Whether this bleeds into Palantir’s more central work with the U.S. DoD is unclear, but possible.
The incoming U.S. Congress will potentially put the company into new focus. How Palantir fares will be instructive for analysts and contractors alike: whether a single firm can retain a position like Palantir’s amid the sprawling defense technology boom and its pitfalls may in part shape contracting vehicles for relevant systems in the immediate years ahead.
The author’s opinions in this piece are his own.
Vincent Carchidi has a background in defense and policy analysis, specializing in critical and emerging technologies. He is currently a Defense Industry Analyst with Forecast International. He also maintains a background in cognitive science, with an interest in artificial intelligence.
