Anduril, Palantir, and SpaceX Reshape US Military Strategy and Procurement

In December 2024, Palantir and Anduril announced the formation of a consortium — later joined by SpaceX and OpenAI — to compete against traditional defense contractors for a share of the $850 billion US military budget [1]. Four months later, President Trump signed an executive order making Other Transaction Authority agreements the Pentagon's preferred contracting method [2]. By March 2026, Anduril and Palantir were co-developing the software backbone for the $185 billion Golden Dome missile defense shield [3].

The speed of this transformation has no modern precedent in defense procurement. What began as a handful of Silicon Valley startups pitching the Pentagon on AI and autonomy has become a structural realignment of how the United States builds, buys, and deploys its weapons systems — with consequences that extend far beyond the balance sheets of Lockheed Martin and Raytheon.

The Money Trail: Contract Growth and Revenue Surge

Palantir's US government revenue illustrates the trajectory. The company earned $610 million from federal contracts in 2020. By 2024, that figure had roughly doubled to $1.2 billion, and Palantir is on track for approximately $1.8 billion in 2025 — a 50% year-over-year increase driven largely by the Maven Smart System contract [4].

Source: Palantir SEC Filings

Data as of Dec 31, 2025CSV

Anduril, which was founded in 2017, won $2.3 billion in contracts in 2024 and $2.6 billion in 2025, with $2.4 billion in active contracts at any given time [5]. SpaceX's federal portfolio is larger still: the company secured $3.8 billion in federal contracts in 2024 alone, with a cumulative total of approximately $15.4 billion, the majority through NASA [6]. The Space Force announced in April 2025 that SpaceX should expect $5.9 billion in launch contracts between 2027 and 2029 [7].

These figures remain small relative to the legacy primes. Lockheed Martin, the Pentagon's largest contractor, reported $71 billion in net sales in 2024. But the growth rate tells the story: Palantir's government revenue nearly tripled from 2020 to 2025, while Lockheed's revenue grew roughly 15% over the same period [4][8].

Bypassing the Rulebook: The OTA Explosion

Much of this growth has been enabled by a contracting mechanism most taxpayers have never heard of. Other Transaction Authority agreements allow the Department of Defense to award contracts outside the Federal Acquisition Regulation — the sprawling body of rules governing competitive bidding, cost accounting, and oversight that applies to traditional defense procurement.

OTA obligations have grown tenfold in eight years, from $1.8 billion in fiscal year 2016 to over $18 billion in fiscal year 2024, with prototype OTAs alone accounting for $16 billion [9].

Source: GAO Report GAO-25-107546

Data as of Mar 1, 2025CSV

A March 2025 Government Accountability Office report found that the DoD's contracting data for OTAs was often incomplete, making it difficult to assess whether these agreements were delivering better outcomes than traditional contracts [9]. The GAO recommended improved data collection — a recommendation the DoD concurred with but has not yet fully implemented.

Trump's April 2025 executive order went further, establishing OTAs as the preferred method for defense acquisition and directing a "comprehensive overhaul" of the contracting system [2]. Critics, including the Project on Government Oversight, argue that this effectively removes competitive bidding protections that exist to prevent favoritism and waste [10]. Defenders counter that FAR-based procurement takes an average of seven years from requirement to deployment — a timeline incompatible with the pace of software and AI development [11].

The Revolving Door: Personnel and Power

The personnel connections between Silicon Valley defense firms and the Trump administration are extensive and documented. The Project on Government Oversight identified multiple senior officials with direct financial or employment ties to the companies now receiving major contracts [10].

Mike Obadal, Trump's nominee for Army under secretary — who would oversee weapons and technology acquisition — was a senior director at Anduril from January 2023 to June 2025. His April 2025 ethics agreement states he will retain his Anduril stake, valued between $250,001 and $500,000 [12]. If confirmed, he would oversee decisions on programs in which Anduril is a major competitor, including Golden Dome.

Emil Michael, Trump's nominee for undersecretary of defense for research and engineering, has advocated for the Pentagon to rely less on custom-built systems and more on commercial technology — a position that directly benefits the business model of companies like Palantir and Anduril [11].

The military has also formalized tech-sector integration through Reserve Detachment 201, announced in mid-2025, which commissioned four tech executives as Army Reserve lieutenant colonels: Meta CTO Adam Bosworth, OpenAI product head Kevin Weil, Palantir CTO Shyam Sankar, and Palantir/OpenAI veteran Bob McGrew [13].

Whether this revolving door exceeds historical norms is debatable. Legacy prime contractors have long placed former generals and admirals on their boards — Raytheon, General Dynamics, and Lockheed Martin each typically have multiple retired flag officers in senior leadership. The difference is structural: legacy primes hire former government officials to navigate existing procurement systems, while the current cohort of appointees appears positioned to redesign those systems in ways that favor their former (and sometimes current) employers [10][11].

Combat Performance: What the Systems Have Actually Done

The operational track record of these companies' products is mixed but increasingly substantial.

Palantir's Maven Smart System began as Project Maven, the Pentagon's AI-enabled intelligence and targeting platform. The DoD raised Maven's contract ceiling to $1.3 billion through 2029 in May 2025, up from $480 million [14]. The Pentagon credits Maven with providing targeting support for US airstrikes in Iraq, Syria, and Yemen in 2024, and with locating hostile maritime assets in the Red Sea [14]. NATO finalized acquisition of the Maven Smart System in March 2025 and incorporated it into exercises by August [15]. By April 2026, DoD components faced what DefenseScoop described as an "aggressive" timeline for transitioning to Maven as a core system [16].

Anduril's Lattice command-and-control platform integrates sensor data from drones, cameras, and radar into a unified operating picture. In December 2024, Anduril and Palantir announced integration of Lattice Mesh with Maven, creating a pipeline from frontline sensors to AI-supported targeting workflows [17]. Specific combat performance data for Lattice remains limited in public reporting.

SpaceX's Starlink has the most extensive combat record of the three. Ukraine's military has used Starlink terminals extensively since 2022 for battlefield communications, drone operations, and artillery coordination. However, Starlink has also experienced documented disruptions: Russia has developed electronic warfare capabilities to jam or spoof Starlink signals in specific areas, and SpaceX has had to issue repeated software updates to counter these measures [18]. The dependency itself creates risk — Ukrainian units reported operational difficulties during periods when Starlink coverage was degraded or geofencing was applied near the Russian border.

Detailed military after-action reports on system failures remain largely classified. The available public evidence suggests these systems perform well in their intended roles but have not been tested against a peer adversary with full-spectrum electronic warfare capabilities.

The Single-Point-of-Failure Debate

The strongest case for concern about consolidation around a few venture-backed firms is straightforward: if two or three companies build the software layer that connects sensors, communications, and weapons across the entire US military, a vulnerability in any one of those platforms becomes a vulnerability in the force itself. The Anduril-Palantir consortium's stated goal — integrating Lattice and Maven into a unified kill chain — means that a software failure, cyberattack, or supply chain compromise could degrade multiple combatant commands simultaneously [17].

The consortium model raises additional competition concerns. When Palantir and Anduril formed their alliance, critics at the Quincy Institute described it as "colluding, rather than competing, for military contracts" — replacing the old oligopoly of five prime contractors with a new oligopoly of three tech firms [19].

The strongest counterargument is that the legacy primes have produced decades of documented failures. The F-35 program's Block 4 modernization is now $6 billion over budget and five years behind schedule [20]. Lockheed Martin delivered 110 aircraft in 2024, all late by an average of 238 days — up from 61 days in 2023 — while still collecting hundreds of millions in performance incentive fees [21]. Pratt & Whitney, an RTX subsidiary, delivered all 123 F-35 engines late in 2024 [20]. The Pentagon's own acquisition system has consistently rewarded incumbency over performance, creating what advocates of reform call a structural guarantee of cost overruns and technological stagnation.

Workforce and Regional Impact

The shift toward Silicon Valley defense firms concentrates economic benefits in regions that are already wealthy. Anduril is headquartered in Costa Mesa, California. Palantir is based in Denver. SpaceX operates primarily from Hawthorne, California, and Boca Chica, Texas. The venture capital funding surge — more than $28 billion in the first half of 2025 alone, surpassing most recent full-year totals — flows overwhelmingly through Sand Hill Road and its equivalents [22].

Legacy defense manufacturing, by contrast, is distributed across congressional districts in states like Connecticut (submarine and engine production), Texas (fighter jets and helicopters), Mississippi (shipbuilding), and Alabama (rocket propulsion). The F-35 program alone sustains an estimated 254,000 jobs across 45 states — a supply chain designed as much for political durability as industrial efficiency [20].

The Trump administration has simultaneously pursued reductions in the DoD civilian workforce. All employees approved for the DoD Deferred Resignation Program or early retirement through the Voluntary Early Retirement Authority were required to leave federal service by September 30, 2025 [23]. No comprehensive retraining or economic mitigation program has been announced for defense communities that may lose contracts as procurement shifts toward software-centric platforms built by smaller workforces.

The China Comparison

Proponents of Silicon Valley integration frequently cite China's Military-Civil Fusion strategy as justification. Beijing has established over 35 dedicated MCF funds since 2015, totaling approximately $68.5 billion in anticipated funding [24]. Analysis of 2,857 AI-related Chinese defense contract awards between January 2023 and December 2024 found that nearly three-quarters of suppliers had no self-reported state ownership ties — civilian companies and universities now win more AI defense contracts than state-owned enterprises [25].

But the comparison has significant limits. A CNAS report found that only about 2% of China's private high-tech enterprises were involved in defense work as of 2019, and progress from 2010 to 2019 was marginal [24]. Private companies in China face licensing costs of approximately $141,000 and multi-year approval cycles, and foreign investment disqualifies firms from sensitive contracts. The report concluded that "state coercion or direction cannot create the integrated ecosystem" necessary for genuine civil-military fusion — suggesting that China's top-down model may be structurally inferior to the US approach, which relies on market incentives and voluntary participation [24].

Russia's wartime industrial adaptation offers a different comparison. Moscow has dramatically increased weapons production since 2022, but largely by running Soviet-era factories on three shifts rather than through technological innovation. The Russian model has produced quantity but not quality — a pattern that reinforces the argument for the US emphasis on AI-enabled systems, but says little about whether those systems should be built by three companies or thirty.

The Musk Question

The most acute conflict-of-interest concerns center on Elon Musk. SpaceX held $15.4 billion in cumulative federal contracts as of 2024 [6]. Musk simultaneously served as head of the Department of Government Efficiency, a White House advisory body with access to agency budgets and personnel systems across the federal government. DOGE targeted agencies that collectively oversee the majority of SpaceX's government business.

Musk operated as a "special government employee" — a temporary appointment limited to 130 days per year [26]. Unlike AI and crypto czar David Sacks, there is no public evidence that Musk obtained a formal conflict-of-interest waiver [27]. The White House stated that Musk would "remove himself if any conflicts of interest arise," but the determination of what constitutes a conflict was left to Musk himself [27].

A Public Citizen report found that Musk had a direct business interest in at least 70% of the departments and agencies targeted by DOGE [28]. Representatives Lynch and Connolly launched a congressional investigation in April 2025, requesting documents on how DoD was ensuring Musk was not using his position to benefit his companies, citing $9.5 billion in defense contracts [6].

The applicable statutes include 18 U.S.C. § 208, which prohibits government employees from participating in matters affecting their financial interests, and 18 U.S.C. § 209, which bars compensation from outside sources for government work. The Office of Government Ethics has jurisdiction over executive branch ethics compliance, but enforcement depends on referrals to the Department of Justice — which is itself subject to the political dynamics of the administration [26].

SpaceX is also a frontrunner for the Golden Dome satellite constellation, with the Wall Street Journal reporting in December 2025 that the company was "set to receive" a $2 billion contract to build 600 satellites for missile targeting [29]. Whether Musk's DOGE role influenced this award is unknown, but the structural conditions for a conflict — simultaneous access to government decision-making and financial interest in government contracts — are present and documented.

What Comes Next

The structural transformation of US defense procurement is already underway and unlikely to reverse. The executive order prioritizing OTAs, the Golden Dome program's reliance on the Anduril-Palantir-SpaceX consortium, and the appointment of tech-sector executives to acquisition oversight roles all point in the same direction: a defense industrial base increasingly shaped by Silicon Valley's business model of rapid iteration, software-defined systems, and venture capital funding.

The open questions are about oversight. Will Congress impose conflict-of-interest requirements on the new class of defense appointees that match or exceed those applied to legacy contractor executives? Will the GAO's recommendations for OTA data transparency be implemented before tens of billions more in contracts are awarded outside traditional competitive bidding? And will the consolidation of AI-enabled battlefield systems around a handful of firms produce the resilience and redundancy that a military facing peer adversaries requires — or a fragility that only becomes apparent under fire?

The answers will shape not just the defense budget but the character of American military power for decades to come.