The New Silicon Triangle: Algorithmic Capital, Executive Power, and the Restructuring of U.S. Foreign Policy Decision-Making ()
1. Introduction
During the Trump administration’s second term, artificial intelligence (AI) technology has been embedded in the process of U.S. foreign affairs with unprecedented depth. Its impact extends beyond the technological level, fundamentally reshaping the underlying logic of U.S. national strategy, the security paradigms governing interstate interaction, and the structure of international political power. As an emerging geopolitical technology, AI is altering traditional power structures, strategic competition models, and the international relations landscape through its data analysis, pattern recognition, and predictive capabilities. Major powers are engaging in a new round of global technological strategic competition surrounding AI algorithms, commercial values, general-purpose technologies, and industrial supply chains.
The Trump administration reshaped U.S. foreign policy with “America First” and “great-power competition,” providing a unique political impetus for the deep integration of AI technologies into the U.S. foreign policy decision-making process. This policy orientation directly accelerated the application of AI technology in U.S. foreign policy and prompted Silicon Valley AI technology companies to develop a unique path to penetrate U.S. foreign policy formulation and implementation—shifting from the traditional lobbying approach of the military-industrial complex to a configuration in which AI technologies, the military-industrial complex, and the White House decision-making circle are deeply intertwined.
This paper addresses three core analytical questions: How is the New Silicon Triangle embedded within the U.S. foreign policy decision-making system? What impact does this embedding have on U.S. foreign policy formulation? And what are its strategic consequences?
Unlike the traditional path in American politics, in which specific interest-lobbying groups influenced individual decision-makers through lobbying channels, Silicon Valley AI companies represented by Palantir, Anthropic, OpenAI, Nvidia, and xAI have involved themselves in the formulation and implementation of foreign policies through technological binding with traditional military-industrial enterprises and through structural coupling with the decision-making mechanisms of the White House and the Department of War. Such a configuration has produced the “black-boxing” of U.S. external decision-making, the failure of the traditional power-accountability system, and consequent weakening of the ethical foundation of decision-making. The New Silicon Triangle formed by private technology companies, the military, and the White House has become a core participant in diplomatic decision-making and increasingly shapes the formulation and implementation of U.S. foreign policy through structural power, institutional power, and discursive power.
This paper contributes to three literatures. First, it extends the classical iron-triangle and military-industrial complex tradition (Eisenhower, 1961; Adams, 1981; Alic, 2021; Roland, 2021; Coyne & Goodman, 2022; Dillman, 2026) by identifying a structurally distinct successor configuration in which Congress is marginalized and a White House-Pentagon-technology-capital axis takes its place. Second, it brings the emerging algorithmic-politics and surveillance-capitalism literatures (Pasquale, 2015; Yeung, 2018; Zuboff, 2019; Crawford, 2021; Kreps & Kriner, 2023) into direct dialogue with Foreign Policy Analysis (Allison, 1971; Hudson, 2014), a connection rarely made. Third, it provides the first systematic mapping of Trump-administration-2.0 institutional innovations—the Department of Government Efficiency (DOGE), the reconstituted President’s Council of Advisors on Science and Technology (PCAST) under tech-capital co-chairmanship, Detachment 201, and the deployment of frontier AI in classified networks—as a coherent restructuring of U.S. foreign-policy decision-making rather than as a series of disconnected events.
The central analytical category of the paper is the embedding mechanism, defined here and developed more fully in Section 3 as a stable, reproducible institutional pathway through which a private actor’s preferences, technologies, or personnel become structurally constitutive of state decision-making, such that the actor’s withdrawal would generate disabling state dysfunction. An actor or event is classified as part of the New Silicon Triangle when three conditions are jointly satisfied: (i) a private technology firm has acquired institutional standing in foreign-policy or military decision-making that exceeds the conventional contractor or lobbying role; (ii) the relationship has been formalized through executive orders, statutory authority, military commissioning, or analogous instruments that confer ongoing decisional access rather than discrete transactional service provision; and (iii) the firm’s continued participation is operationally non-substitutable in the short term, such that an attempted state withdrawal would generate immediate functional dysfunction. These three criteria—institutional standing, formalized access, and operational non-substitutability—jointly distinguish New Silicon Triangle participants from the broader population of technology firms doing business with the U.S. government.
Methodologically, this paper is best understood as an interpretive case study of a configuration still in formation. It draws on public-source documentation—executive orders, departmental memoranda and strategy documents, congressional disclosures, court filings, mainstream news reporting, and corporate statements—supplemented by the relevant secondary scholarship in international relations, foreign policy analysis, and the political economy of technology. The paper is neither a quantitative test of a fully specified causal model nor a forward-looking forecast; it is a conceptual and empirical mapping of an emerging institutional configuration whose contours are visible but whose long-term trajectory remains contingent on subsequent political developments. The empirical claims advanced below should accordingly be read as descriptions of the configuration at a particular moment (mid-2026), subject to revision as additional documentary access becomes available.
A note on terminology. Throughout this paper, the Department of War refers to the executive-branch entity rebranded by Executive Order 14347 (2025); the underlying legal entity, the Department of Defense, retains that legal name pending congressional action. We use the Department of War form consistently throughout the analytical narrative, both because it is the form adopted by the executive itself during the period under analysis and because the rebranding is analytically relevant as a signal of the administration’s preferred self-presentation. Where the legal entity is at issue, the Department of Defense form is used.
The remainder of the paper is organized as follows. Section 2 reviews four relevant literatures, addresses three alternative explanations, and identifies the research gap. Section 3 develops a three-dimensional power framework and elaborates the concept of an “embedding mechanism.” Section 4 traces the historical evolution from the classical iron triangle to the New Silicon Triangle. Section 5 analyzes the four embedding mechanisms through which the New Silicon Triangle has been constituted, distinguishing three levels of analytical claim (policy influence, operational dependence, and co-decision) and indicating which evidence supports each. Section 6 examines its five strategic consequences, each tied to a concrete documented episode and an explicit causal step. Section 7 concludes with implications for theory, policy, and future research.
2. Literature Review: From the Iron Triangle to State-Tech Fusion
The argument advanced in this paper draws on, and seeks to integrate four bodies of scholarship that have so far developed largely in isolation: the iron-triangle and military-industrial complex tradition; the state-capture and regulatory-capture literatures; the public-private military partnership literature; and the recent scholarship on algorithmic politics and surveillance capitalism. This section reviews each in turn and concludes by identifying the conceptual gap that the New Silicon Triangle is offered to fill.
2.1. The Iron Triangle and Its Digital Evolution
The “iron triangle” concept—a closed coalition of congressional committees, federal agencies, and interest groups whose interlocking interests dominate sectoral policymaking—has been a staple of American political analysis since at least the 1960s (Cater, 1964; Heclo, 1978). In the defense sector, it found its most influential expression in Eisenhower’s (Eisenhower, 1961) farewell warning about the “military-industrial complex” and in subsequent scholarship on the political economy of defense procurement (Adams, 1981; Alic, 2021; Coyne & Goodman, 2022). Recent work has documented the iron triangle’s adaptation to the digital era: Roland (2021) argues that information technology has not dissolved but reconfigured the complex, while Dillman (2026) explicitly frames the alliance among Google, the Pentagon, and MITRE as a “new iron triangle” centered on digital surveillance.
This paper builds on but departs from this lineage in two ways. First, where the classical iron triangle places Congress as a constitutive vertex, the structure analyzed here is characterized by congressional marginalization: legislative oversight has been systematically displaced by executive-branch fusion with private technology capital. Second, where Dillman (2026) treats the new alliance as a continuation of surveillance-state expansion, this paper argues that the more consequential transformation is the embedding of generative AI systems into the operational core of foreign-policy decision-making itself—not merely the surveillance apparatus.
2.2. State Capture, Regulatory Capture, and Their Limits
The state-capture and regulatory-capture literatures (Stigler, 1971; Carpenter & Moss, 2014) provide useful vocabulary for analyzing private interests’ colonization of public decision-making but were developed primarily to analyze sectoral regulatory agencies in telecommunications, finance, and environmental protection. Two features of the New Silicon Triangle exceed this framework. First, the captured domain is not regulation but foreign and military policy—domains in which capture has historically been understood through different conceptual lenses (the military-industrial complex; the intelligence-industrial complex). Second, the captured technology—frontier AI—is not merely an instrument of policy implementation but an infrastructure that increasingly co-produces the decisional categories within which policy is formulated. This second feature—what Yeung (2018) terms “algorithmic governance”—requires conceptual resources beyond what classical capture theory provides.
2.3. Public-Private Military Partnerships
A third relevant literature concerns the privatization of state coercive capacity. Singer (2003) and Avant (2005) analyzed the late-twentieth-century rise of private military and security contractors—Blackwater, DynCorp, MPRI—as a return to historical patterns of mercenary warfare under contemporary state oversight. The New Silicon Triangle differs from this configuration in two crucial respects. The traditional private military contractor was a service provider hired by the state to execute the state’s strategic intent; the technology firm in the New Silicon Triangle is increasingly a co-formulator of that intent. And whereas private military contractors operated largely outside the formal decision-making architecture of the state, the actors examined here—through DOGE, the reconstituted PCAST, and Detachment 201—are embedded directly within executive institutions.
2.4. Algorithmic Politics and Surveillance Capitalism
The most rapidly developing relevant literature is the analysis of algorithmic and platform power. Zuboff (2019) describes a regime of “surveillance capitalism” in which the collection and computational processing of behavioral data have become the foundation of a new commercial-political order. Pasquale (2015) analyzes the “black-box society” in which algorithmic decision-making evades the transparency requirements that constitute democratic legitimacy. Crawford (2021) maps the material infrastructure of AI—data, energy, labor, geopolitics—on which any algorithmic system depends. Kreps and Kriner (2023) and Khan et al. (2024) extend this analysis specifically to the implications of AI for democratic governance and federal policymaking.
A parallel body of work on the military applications of AI—Scharre (2018), Boulanin and Verbruggen (2017), Horowitz (2018), Suchman (2020), and Schwarz (2018)—has examined the development and deployment of autonomous and algorithmic weapon systems, the international-security implications of AI for great-power balance, and the ethical and accountability problems raised by algorithmic violence. What remains comparatively underdeveloped, however, is sustained engagement with foreign-policy decision-making itself as a specific institutional site of algorithmic embedding. Most analyses of AI and the state focus on domestic regulation, surveillance, or service delivery; the military-AI literature focuses on weapons and warfighting; the decisional architecture that translates strategic intent into operational action has remained the connective tissue least examined. This paper addresses that gap by analyzing the institutional configuration through which algorithmic systems and the firms that produce them enter into the formulation of foreign policy itself.
2.5. The Research Gap
Synthesizing these four literatures yields the research gap this paper addresses. Existing scholarship has documented (a) the continuing relevance of iron-triangle-style alliances in defense policy; (b) the conceptual vocabulary of capture; (c) the privatization of coercive capacity; and (d) the algorithmic restructuring of governance. What has not yet been developed is an integrated account of how these four dynamics converge in a single institutional configuration that simultaneously transforms the actors, the technology, and the decisional architecture of U.S. foreign-policy formulation. The “New Silicon Triangle” concept advanced in this paper is offered as such an integrated account.
2.6. Alternative Explanations and Their Limits
The institutional patterns this paper attributes to the New Silicon Triangle could in principle be analyzed through three alternative frameworks, each of which captures part of the phenomenon. Considering these alternatives clarifies what the New Silicon Triangle framework adds.
First, executive-centralization accounts in the unitary-executive and presidential-studies tradition (Skowronek, 1993; Howell, 2003; Mayer, 2001) would interpret the configuration analyzed here as the latest iteration of a long-running tendency toward presidential aggrandizement, particularly in foreign and military policy. This framework correctly identifies the role of unilateral executive instruments such as Executive Order 14158 establishing DOGE, Executive Order 14347 rebranding the Department, and the use of Defense Production Act emergency authorities to accelerate procurement. But executive-centralization accounts cannot explain why the actors entering the executive’s expanded decisional space are specifically frontier-AI firms rather than the broader population of constituencies that prior centralization episodes have empowered (the unitary-executive Bush-era counterterrorism apparatus, for example, did not commission Silicon Valley executives into the Army Reserve). Executive centralization is a necessary but not sufficient condition for the configuration this paper analyzes.
Second, legacy military-industrial-complex analyses (Adams, 1981; Coyne & Goodman, 2022; Roland, 2021) would treat the new arrangements as a continuation of classical defense contracting under different vendor names. This framework correctly identifies continuities: the Defense Production Act of 1950 remains the legal substrate, Lockheed Martin and RTX retain dominant positions in conventional weapons procurement, and the revolving-door logic that Singer (2003) and Avant (2005) documented for private military contractors clearly persists. But classical military-industrial-complex frameworks cannot account for the qualitative transformation in which the embedded firms are no longer suppliers of weapon systems to be deployed at the state’s discretion but rather producers of the decisional infrastructure within which strategic choices are first formulated. When Anthropic’s Claude model writes the targeting recommendations that field commanders execute, the firm has not merely sold a product but has entered the constitutive process of military decision-making in a way that no twentieth-century defense contractor did.
Third, standard bureaucratic-politics accounts (Allison, 1971; Halperin, 1974; Hudson, 2014) would describe the configuration as a redistribution of influence among existing organizational actors—DOGE displacing OMB, the AI-and-Crypto Czar’s office displacing parts of OSTP and the NSC technology directorate, the reconstituted PCAST displacing the State Department’s traditional technology-diplomacy role. This framework correctly identifies organizational displacement and the politics of bureau-level turf battles. But the bureaucratic-politics model presupposes that the competing actors are themselves agencies of the state operating within a shared institutional environment; it cannot accommodate the appearance of private corporate actors—Anthropic, Palantir, xAI, Andreessen Horowitz partners—in decisional roles that the model does not contemplate as bureau-level players.
The New Silicon Triangle framework integrates what each of these alternatives captures while explaining what each misses. It identifies the specific actors (frontier-AI firms and their adjacent technology and venture-capital networks), the specific institutional vehicles (DOGE, PCAST under tech-capital co-chairmanship, Detachment 201, classified-network AI deployment, the May 2026 state-visit delegation), and the specific structural transformation (algorithmic systems becoming constitutive of the decisional environment) that the three alternatives leave unexplained. The relationship is one of complementary scope rather than replacement: where the alternatives explain antecedent or partial dynamics, the New Silicon Triangle explains the integrated configuration that those antecedents have produced.
3. Analytical Framework
3.1. Three Dimensions of Power
This paper analyzes the New Silicon Triangle through a three-dimensional power framework drawn from Strange (1988) and Lukes (2005). Strange distinguishes between relational power—the capacity of A to compel B to do something B would otherwise not do—and structural power: the capacity to shape the frameworks within which other actors operate, including the security, production, financial, and knowledge structures of the international political economy. Lukes (2005) adds a third dimension to the standard power debate: power operates not only through observable decisions and the suppression of alternatives but also through the shaping of preferences, perceptions, and the very conceptual horizons within which choices appear possible.
This paper operationalizes three corresponding dimensions of Silicon Triangle power:
Structural power: the capacity of technology firms to determine the material, computational, and infrastructural conditions on which foreign-policy decisions depend—for example, the classified-network deployment of frontier AI, satellite-imagery processing, and command-and-control platforms.
Institutional power: the capacity to occupy and reshape formal decision-making bodies—for example, the staffing of DOGE and PCAST, the appointment of venture-capital partners to advisory bodies, and the commissioning of technology executives into military ranks through Detachment 201.
Discursive power: the capacity to shape the conceptual vocabulary and ideological frame within which foreign-policy alternatives are perceived—for example, the algorithmic amplification of “innovation-over-regulation” framing and the use of Truth Social as a presidential communications platform owned by a corporate vehicle in which the president holds equity.
3.2. The Concept of an Embedding Mechanism
The term embedding mechanism is used here in a specific sense: a stable, reproducible institutional pathway through which a private actor’s preferences, technologies, or personnel become structurally constitutive of state decision-making, such that the private actor’s withdrawal would generate disabling state dysfunction. Embedding is thus distinct from lobbying (which influences but does not constitute), from contracting (which exchanges services without altering decisional architecture), and from regulatory capture (which neutralizes oversight without occupying decisional roles). An embedded actor is one whose continued participation has become a condition of state operation.
This definition has two analytical advantages. First, it sets a high empirical bar: for a configuration to qualify as embedded, withdrawal must produce demonstrable dysfunction, not merely inconvenience. The Anthropic case examined in Section 5.2 satisfies this criterion: even after the company was officially designated a supply chain risk in March, 2026, the Pentagon continued to use its Claude model in Operation Epic Fury because operational continuity required it. Second, the definition discriminates between configurations that look superficially similar (e.g., classical defense contracting versus contemporary algorithmic embedding) but differ fundamentally in the nature of the dependency they generate.
3.3. Mapping Mechanisms to Power Dimensions
The four embedding mechanisms identified in Section 5 below correspond to the three power dimensions as shown in Table 1.
Table 1. Embedding mechanisms mapped to power dimensions.
Mechanism (Section 5) |
Primary power dimension |
Secondary dimension |
5.1 Formal institutional restructuring (DOGE, PCAST
under tech-capital leadership) |
Institutional |
Structural |
5.2 Algorithmic monopoly over military power (Maven, classified-network frontier AI) |
Structural |
Institutional |
5.3 Algorithm-driven populist mobilization (Truth Social, Detachment 201, Sacks) |
Discursive |
Institutional |
5.4 Reshaping the decision-making architecture |
Structural |
Discursive |
This mapping demonstrates that the New Silicon Triangle is not a single-mechanism phenomenon but a coordinated multi-dimensional power configuration—which is precisely why classical accounts focused only on lobbying, contracting, or surveillance fail to capture its full character. A configuration that operates simultaneously through structural, institutional, and discursive channels can survive the failure of any single channel, and reform efforts targeting only one channel will be correspondingly insufficient.
4. From the Iron Triangle to the New Silicon Triangle: Historical Evolution
To understand the shift in power in contemporary U.S. foreign policy from the iron triangle to the New Silicon Triangle, it is necessary to trace the historical development of the military-industrial complex itself. President Eisenhower’s 1961 (Eisenhower, 1961) farewell address warned that the Cold War had fostered a symbiotic network of government, military, and private defense companies, which, through massive contracts, political lobbying, and donations, had formed a self-reinforcing cycle of interests. He warned that this combination of military institutions and the defense industry could pose a potential threat to democratic decision-making. The military-industrial complex itself originated with the Defense Production Act of 1950, authorized by President Truman for the supply of materials and armaments for the Korean War. One provision of the Act allowed the president to require companies to prioritize government contracts and orders for defense needs; other provisions granted the president the ability to use loans and additional incentives to increase the production of key commodities and authorized the government to establish voluntary agreements with private companies. This Act is considered one of the most powerful and adaptable industrial-policy tools available to the U.S. government.
During the Cold War, top defense contractors such as Lockheed Martin, RTX, Boeing, and General Dynamics dominated the U.S. national-security procurement system, covering the research and production of all types of military equipment, including fighter jets, missiles, nuclear submarines, and military electronic equipment. This created a stable iron triangle among the military-industrial complex, the Department of Defense, and Congress. The structure consolidated interests and shaped U.S. foreign-policy decisions through three mechanisms. First, the transfer of benefits primarily through political donations: military-industrial firms used substantial lobbying resources to influence government decisions, including lobbying the president, members of Congress, and various congressional committees related to defense. Second, the revolving-door mechanism between politics and business: former politicians and officials joined military-industrial, consulting, and lobbying firms after leaving office, while company directors entered the decision-making system, directly facilitating decisions favorable to military-industrial companies. Third, direct narrative manipulation: economic interests were domestically linked to voters’ politics and decision-making, fostering a symbiotic relationship between the U.S. military-industrial complex and electoral politics. Internationally, this involved selling war propaganda, using pretexts such as “great-power threat” or “humanitarian intervention” to manipulate public opinion and legitimize U.S. wartime actions, thereby achieving foreign-policy preferences with stable public support.
For decades, the shadow of the iron triangle has loomed behind every American foreign war and military operation. The protracted twenty-year war in Afghanistan, costing over $2.26 trillion, saw more than seventy percent of its revenue flow to military-industrial enterprises under the guise of “security construction” and “equipment maintenance.” The prioritization of personal, commercial, and factional interests over broader U.S. national-security objectives and global public interests became a fundamental driver of disorder and strategic miscalculation in U.S. foreign policy making. This dynamic not only reinforced hegemonic tendencies in U.S. foreign strategy but also intensified global geopolitical instability.
In the era of information technology, however, the foundation of national power has begun to shift. AI-oriented intelligence, algorithms, data, and computational energy are becoming the new foundation of national power, reshaping hegemonic advantage in ways that the classical iron-triangle framework cannot fully capture. The U.S. Department of War’s Artificial Intelligence Strategy—U.S. Military AI Dominance asserts that AI-enabled warfare will set the trajectory of military affairs in the next decade, and the Department aims to transform the military into an “AI-first” combat force. Traditional defense firms have responded by intensifying investment in AI research, using their manufacturing strengths to address gaps and leap forward in military decision-making, intelligence analysis, public opinion management, and weapon automation. For instance, Lockheed Martin’s Project Overwatch integrates AI into fighter-jet information systems to help pilots identify targets and respond to threats more rapidly. But the more consequential transformation is the entry of a new class of actor: the frontier AI firm, whose contribution to national power is not equipment but the algorithmic systems that increasingly constitute the decisional environment itself.
In traditional geopolitical discourse, the term “Silicon Triangle” typically refers to the trilateral interdependence of the United States, Chinese Mainland, and the Taiwan region within the semiconductor supply chain (Diamond et al., 2023). This paper uses the term in a deliberately different sense, and to avoid conceptual collision designates the configuration analyzed here as the New Silicon Triangle: a tripartite alliance among Silicon Valley AI firms, the White House, and the Department of War. This structure functions not merely as a production engine for technology but as a joint adjudicator of U.S. foreign policy and global strategy. Unlike the traditional military-industrial complex, in which private companies and the Pentagon form an axis subject to congressional oversight, the New Silicon Triangle exhibits an “administrative-corporate axis” structure: decision-making power is concentrated in the integration of executive departments and technological capital, marked by the process of embedding technology into administrative and military decision-making. Private AI companies are no longer merely defense contractors or external interest groups but enter the decision-making circle as providers of technology systems, embedded interlocutors, and co-formulators of rules.
The emergence of the New Silicon Triangle is not a sudden institutional change but a structural product gradually formed through the accumulation of factors—the needs of U.S. domestic national-security strategy, the convergence and development of technology, industry, capital, and talent—driven by great-power competition, alliance collaboration, global technology governance, and technological competition. Trump’s second term has overseen an unprecedented structural overhaul of the U.S. foreign-policy apparatus. Central to this transformation was the January 20, 2025, Executive Order establishing the Department of Government Efficiency (DOGE). This restructuring effectively institutionalized the New Silicon Triangle, systematically embedding tech-corporate interests into the core decision-making processes of the White House and the Pentagon.
5. Four Embedding Mechanisms of the New Silicon Triangle
Building on the analytical framework developed in Section 3, this section identifies four embedding mechanisms through which the New Silicon Triangle has been constituted as a structurally novel feature of the U.S. foreign-policy decision-making apparatus.
The analytical claims developed in this section are pitched at three distinct levels, and readers should be alert to which level each piece of evidence supports. The weakest claim is policy influence: the capacity of technology firms to shape policy outcomes through traditional channels such as lobbying, advisory participation, campaign contributions, and revolving-door personnel flows. Evidence for this level is widely available (the $226,000-per-day Big Tech lobbying figure, Anthropic’s $1.56 million in Q1 2026 federal lobbying, the dismantling of more than one hundred state-level AI regulations following industry pressure), is uncontroversial in the existing interest-group literature, and would not by itself justify a new analytical framework. The middle claim is operational dependence: the condition in which state functions can no longer be performed without ongoing private-firm participation, such that the state’s withdrawal of the firm’s services would generate immediate dysfunction. Evidence for this level requires procurement-volume data, operational case files, and—most decisively—instances of attempted state withdrawal that test substitutability; the Anthropic supply-chain-risk designation of February 27, 2026, and the Pentagon’s continued use of Claude during Operation Epic Fury even after that designation, provide exactly such a test, and the test confirms operational dependence. The strongest claim is co-decision (or co-authorship): the condition in which private-firm preferences, technologies, or personnel are constitutive of the decisional process itself, such that the formulation of foreign policy cannot meaningfully be characterized as occurring within the state alone. Evidence for this level requires demonstrating not only operational dependence but also active embedding of firm personnel and preferences in formal decision-making bodies; we develop this strongest claim primarily through the PCAST membership under Sacks’s co-chairmanship, the Detachment 201 commissioning of senior technology executives without recusal from their commercial roles, and the May 2026 state-visit delegation in which the U.S. president introduced technology CEOs to a foreign head of state on diplomatic terms. Readers should expect §5.1 to establish co-decision in domestic technology-policy formulation; §5.2 to establish operational dependence in military operations; §5.3 to combine co-decision in foreign-policy execution (the state-visit case) with policy influence at scale (the lobbying figures); and §5.4 to synthesize the three levels into the broader claim that the decisional architecture itself has been reshaped.
5.1. Formal Restructuring of Decision-Making Units: From Bureaucratic Pluralism to a Core Circle of Technological Oligarchs
The establishment of the Department of Government Efficiency (DOGE) by Executive Order 14158 on January 20, 2025 (Executive Order, 2025), marked a decisive transition for the Trump administration’s second term—moving from “revolving-door” lobbying toward institutional decision-making power in the technology field. Although DOGE was positioned as a temporary organization scheduled to terminate on July 4, 2026, it rapidly intervened in the budget-review processes of multiple federal departments, including the dismantling of the State Department’s Agency for International Development (USAID) and the restructuring of projects at the Defense Advanced Research Projects Agency (DARPA). A May 2025 Pentagon memorandum further empowered DOGE teams to review most unclassified Pentagon contracts above specified thresholds, granting the body unprecedented procurement-level influence over the Department of War.
Beyond DOGE, the Trump administration has restructured executive technology-advisory institutions to formalize technology-capital participation in foreign-policy and national-security decision-making. The President’s Council of Advisors on Science and Technology (PCAST), reconstituted in March, 2026, was placed under the co-chairmanship of David Sacks—the White House AI and Crypto Czar—and senior technology adviser Michael Kratsios. Its membership of up to twenty-four includes Marc Andreessen, Sergey Brin, Safra Catz, Michael Dell, Larry Ellison, Jensen Huang, Lisa Su, and Mark Zuckerberg—a roster heavily dominated by Silicon Valley founders, venture capitalists, and frontier-AI executives. Sacks himself retains, through his venture firm Craft Ventures, more than four hundred investments in technology firms with ties to AI; ethics scholars have characterized the ethics waivers issued to him as a “sham ethics waiver” lacking the rigorous objective analysis that would ensure public policy is made for public benefit (NPR, 2025). The reconstituted PCAST and the parallel advisory channels now operating under Sacks’s leadership concentrate decision-influencing power on foreign-policy technology issues in a small group sharing a “technological-liberalism” ideology, in ways that have visibly marginalized the State Department, the Department of Commerce’s traditional technology-policy offices, and relevant congressional committees. The administration’s signing of an executive order to preempt more than one hundred state-level AI regulations (Executive Order, 2025), and the May 2025 rescission by the Department of Commerce’s Bureau of Industry and Security of the Biden-era AI Diffusion Rule, illustrate how technology-policy outcomes are now being produced through tighter executive-tech-capital coordination rather than through the broader interagency processes typical of previous administrations.
Simultaneously, the Pentagon’s procurement and innovation structures have undergone reorganization under reshaped leadership. The Defense Innovation Unit (DIU), established in 2015, now reports through restructured channels under acting director Emil Michael, the Trump-administration Pentagon Chief Technology Officer. According to the Department of War’s “Artificial Intelligence Strategy Update: Operational Integration of Autonomous Systems,” released in May, 2025, frontier-AI procurement pathways have been streamlined to permit commercial products such as Anduril Industries’ “Lattice” autonomous system and Palantir’s “Meta Constellation” satellite-intelligence platform to enter Pentagon operations under accelerated procurement designations. The cumulative effect is a decision-making architecture in which the Pentagon’s institutional preferences and the technology firms’ product timelines have converged in ways unprecedented in post-Cold-War defense acquisition practice.
These structural changes have finalized the legitimacy of the integration of AI technology into the workflow of the White House and the Pentagon. Together with the executive orders, strategy documents, and shared-interest networks linking President Trump and oligarchs from both private technology and military-industrial businesses, they constitute the institutional foundation of the New Silicon Triangle.
5.2. Algorithmic Monopoly in the Reconfiguration of Military Power
The Department of War’s 2026 AI Strategy Memorandum does not merely describe a technological upgrade; it codifies a structural reconfiguration of military power in which algorithmic systems, developed and operated by a small oligopoly of private technology firms, have become the decisive substrate of American warfighting capacity across the entire spectrum from raw intelligence to lethal combat execution.
The clearest institutional expression of this transformation is Project Maven, officially designated the Algorithmic Warfare Cross-Functional Team, which since its 2017 inception has evolved from a narrow computer-vision tool for processing drone footage into what analysts now describe as the “technological backbone” of current Pentagon operations (Suchman, 2020; Scharre, 2018), streamlining the kill chain by reducing the time between detection and engagement. The scale of this algorithmic penetration is quantitatively striking: Maven’s platform ingests data from more than 150 sources—satellite imagery, drone video, radar, infrared sensors, signals intelligence, and geolocation data—with computer-vision algorithms automatically detecting and classifying battlefield objects, and an AI Asset Tasking Recommender proposing which weapons platforms and munitions should be assigned to each target (on the development of such autonomy in weapon systems more broadly, see Boulanin & Verbruggen, 2017; Horowitz, 2018), generating, as NGA Director Vice Admiral Frank Whitworth has publicly confirmed, approximately one thousand targeting recommendations per hour. The cognitive displacement of human judgment that this represents is made explicit by a striking operational benchmark: the 18th Airborne Corps reportedly achieved comparable targeting output to the two-thousand-person cell used during Operation Iraqi Freedom with roughly twenty people.
The DoW memorandum accelerates this trajectory by institutionalizing it as deliberate strategy, directing the Agent Network initiative to deploy AI agents for battle management and decision support spanning the full range from campaign planning to kill-chain execution, and directing Ender’s Foundry to tighten simulation-to-operations feedback loops—thereby embedding algorithmic systems not merely as analytical aids but as constitutive agents in the formulation and prosecution of military action. What makes this reconfiguration a question of monopoly rather than merely modernization is the degree to which this algorithmic power is concentrated in a single corporate actor. Palantir, with a market capitalization approaching $360 billion in early 2026, operates within a combined Pentagon and Army contracting framework that reaches into the tens of billions of dollars: the Maven Smart System contract ceiling was raised to approximately $1.3 billion through 2029, while a parallel Army enterprise agreement signed in 2025 is valued at up to $10 billion over a decade. When the operational effectiveness of U.S. military operations is structurally dependent on the continuity of these relationships, the traditional distinction between public military authority and private commercial interest is fundamentally compromised. The DoW memorandum tacitly acknowledges this dependency while seeking to manage it through competitive pressure, directing Modular Open System Architectures to prevent vendor lock-in and establishing the Swarm Forge as a competitive mechanism to iteratively test novel AI-enabled fighting approaches; yet the logic of the strategy simultaneously deepens reliance on frontier commercial models by mandating their deployment within thirty days of public release.
A second case is the infrastructure cooperation between Anthropic’s Claude model and the Department of War. Anthropic’s institutional relationship with the Department of War did not begin as a peripheral vendor arrangement but as a foundational one: Anthropic was reportedly the first frontier AI company to deploy its models in the U.S. government’s classified networks, the first to deploy them at the national laboratories, and the first to provide custom models for national-security customers, with Claude becoming extensively deployed across the Department of War and other national-security agencies for mission-critical applications, including intelligence analysis, modeling and simulation, operational planning, and cyber operations. This integration was formalized in July 2025 when Anthropic and the Pentagon entered into a contract under which Claude became the first frontier model approved for use on classified networks, part of a broader arrangement in which the Pentagon awarded contracts worth up to $200 million each to four leading AI laboratories—Anthropic, OpenAI, Google, and xAI—with the objective of embedding agentic AI into classified military networks to process intelligence faster, compress decision cycles, and give American commanders a cognitive edge over adversaries.
Through its deep integration with Palantir’s AI Platform and Amazon Web Services, Claude was woven into the Pentagon’s command-and-control architecture in a manner that defense planners described as forming a continuous loop across data collection, analysis, decision generation, and command issuance. The full operational weight of this integration was revealed starkly during Operation Epic Fury, launched against Iran on February 28, 2026 and conducted primarily in February-March, 2026, when Palantir’s Maven Smart System, powered by Anthropic’s Claude, generated approximately one thousand prioritized targets in the first twenty-four hours, with the system synthesizing classified feeds from satellites, surveillance drones, and archived intelligence data into prioritized target lists complete with precise GPS coordinates, weapons recommendations, and automated legal justifications for strikes—a targeting workflow that previously required an intelligence staff of approximately two thousand analysts during the 2003 Iraq invasion but now reportedly requires approximately twenty people.
What makes this case analytically significant beyond its operational scale is the constitutional and political crisis it precipitated. When Anthropic refused the Pentagon’s demand to remove safety guardrails against autonomous weapons and mass-domestic surveillance, the Trump administration designated the company a “supply-chain risk”—a label previously reserved for U.S. adversaries and never before applied to an American company—and directed all federal agencies to cease using its technology. Yet Pentagon Chief Information Officer Kirsten A. Davies confirmed to the Senate Armed Services Committee that the Pentagon was actively using Claude as part of Operation Epic Fury even after the designation, stating that the Department would not “in any way interfere with the success, the lethality, and the resilience of our warfighters.” This contradiction—simultaneously blacklisting a technology as a national-security threat while operationally depending on it to execute the largest American military campaign since the 2003 invasion of Iraq—exposes the structural logic that the DoW’s 2026 AI Strategy memorandum encodes but does not fully confront: in mandating thirty-day deployment cycles for the latest frontier models and directing the CDAO to establish “any lawful use” language as a standard procurement criterion, the Department has created conditions in which frontier AI models are no longer software that governments can pick up and put down—they are strategic infrastructure, and the corporations that build and constrain them have become, whether by design or by accumulation, co-authors of American military power.
The constitutional fragility of this arrangement was further exposed on March 26, 2026, when a U.S. district court in California indefinitely blocked the Pentagon’s supply-chain-risk designation against Anthropic, ruling the measure unconstitutional and characterizing it as approaching the Orwellian: in the court’s words, “nothing in the governing statute supports the Orwellian notion that an American company may be branded a potential adversary and saboteur of the U.S. for expressing disagreement with the government.” The court’s intervention, however, reached the executive’s punitive instrument while leaving the deeper structural dependency—the Department’s continuing operational reliance on Anthropic’s model during Operation Epic Fury—undisturbed. The pattern is diagnostic: the executive’s coercive instruments against a non-compliant technology supplier could be blocked by the judiciary, but the underlying co-dependence between the state’s warfighting capacity and the private firm’s frontier model continued operating regardless of either branch’s formal posture.
The result is a historically novel form of military power: one in which the algorithmic infrastructure of violence—from the classification of a satellite image as a threat to the assignment of a munition to a target—is no longer solely the sovereign prerogative of the state but is co-produced, co-owned, and co-operated by private corporations whose strategic alignment with the government is commercially incentivized yet constitutionally unaccountable.
5.3. Algorithm-Driven Populist Mobilization
Foreign-policy formulation involves the coordination of diverse interests, and decision theory focuses on how preferences are aggregated into collective choices. Traditionally, U.S. foreign-policy preferences have been formed through executive-legislative interaction and bureaucrat-interest-group games. The embedding of the New Silicon Triangle introduces a new preference-aggregation mechanism: algorithm-driven populist mobilization. David Sacks, a venture capitalist whose investment portfolio directly spans AI companies, was appointed White House AI and Crypto Czar in December, 2024; he maintained more than four hundred investments in tech firms with ties to AI, while critics described his ethics waivers as tantamount to a “presidential pardon in advance,” enabling him to shape federal AI policy—including the dismantling of over one hundred state-level AI regulations—in ways that directly served the commercial interests of OpenAI, Google, and Andreessen Horowitz, all of which had lobbied for precisely that outcome. Meanwhile, in June 2025 the U.S. Army formally commissioned four senior technology executives directly into the Army Reserve at the rank of Lieutenant Colonel through Detachment 201: The Army’s Executive Innovation Corps—Shyam Sankar (Palantir’s chief technology officer), Andrew Bosworth (Meta’s chief technology officer), Kevin Weil (OpenAI’s chief product officer), and Bob McGrew (advisor at Thinking Machines Lab and former chief research officer at OpenAI). The commissioning was explicitly framed by the Army as a vehicle for fusing senior commercial-AI leadership with operational defense planning, and reporting has confirmed that the commissioned officers will not be required to recuse themselves from the Department’s contracting relationships with their employer firms. This arrangement constitutes what may be characterized as a structural merger of corporate and military command authority and a fundamental reconfiguration of the public-private boundary in military decision-making (cf. Suchman, 2020; Schwarz, 2018).
The diplomatic counterpart to this military commissioning came into full public view in May, 2026, when President Trump conducted a state visit to Beijing accompanied by a delegation of more than a dozen U.S. business leaders, with a notably high-density technology and frontier-AI cohort: Elon Musk (Tesla and SpaceX), Tim Cook (Apple), Jensen Huang (Nvidia), Sanjay Mehrotra (Micron), Dina Powell McCormick (Meta), and senior executives from Qualcomm and Coherent, alongside financial-sector leaders Larry Fink (BlackRock), Jane Fraser (Citi), Stephen Schwarzman (Blackstone), David Solomon (Goldman Sachs), and Kelly Ortberg (Boeing). At the state banquet held at the Great Hall of the People on May 14, 2026, Trump personally introduced each chief executive to President Xi Jinping in turn, with state media coverage emphasizing the president’s role as the personal facilitator of his technology delegation’s commercial interests in the Chinese market. The substantive outputs of the visit confirmed this representational logic: shortly after the Trump-Xi summit, Washington cleared the export of Nvidia H200 AI chips to several major Chinese technology firms, and President Xi publicly told the assembled American CEOs that China’s door to U.S. business would “open wider.” (Xinhua News, 2026) What the trip made unmistakable is that the New Silicon Triangle is no longer an inside-Washington phenomenon. The public-diplomatic theater of U.S.-China relations now itself functions as a vehicle for the commercial agendas of frontier-AI and adjacent-technology firms, and the U.S. presidency operates, at its most ceremonious moments, as the diplomatic conduit of those agendas—a configuration in which the distinction between U.S. foreign-policy interests and U.S. technology-firm interests has been not merely blurred but performatively dissolved on the world stage.
Two additional mechanisms reinforce the New Silicon Triangle’s influence: the platform displacement of traditional diplomatic and policy deliberation, and the re-framing of complex foreign and technology policy through a venture-capital-funded discourse. The Trump second term has produced what Boston University diplomatic scholar Paul Webster Hare describes as a regime of “Truth Social public diplomacy,” in which the platform owned by Trump Media & Technology Group—of which Trump holds a controlling stake—functions simultaneously as the primary channel of foreign-policy announcement and as a financial asset whose value is directly augmented by presidential posting activity. A Washington Post analysis documented that, as of June, 2025, Trump had posted 2,262 times to Truth Social in the 132 days since his inauguration—more than three times the number of tweets he sent during the same period of his first presidency, with posts routinely preceding or substituting for formal diplomatic and institutional processes. When Trump announced new tariff rates on Japan, South Korea, and twelve additional countries in July, 2025, the U.S. president posted letters to social media outlining the intended new rates before any formal U.S. Trade Representative process was completed—a pattern the Congressional Research Service confirmed when it noted that many preliminary agreements were announced via social media before the official joint statements were finalized and released. During the Iran conflict in March, 2026, this displacement of institutional deliberation reached its most consequential form: Trump announced a claimed ceasefire agreement on Truth Social that Iranian officials immediately denied, yet when U.S. markets opened a few hours later, the Dow Jones, S&P 500, and Nasdaq all jumped, and global oil prices fell from more than $112 a barrel to less than $100, demonstrating how platform-mediated presidential signaling now directly moves global markets and shapes international expectations in advance of, and sometimes instead of, formal diplomatic outcomes.
The second mechanism is the tech oligarchs’ re-framing of policy discourse through venture-capital-funded ideological infrastructure. Unlike conventional lobbying, which targets discrete legislative outcomes, this mechanism functions upstream, shaping the very conceptual vocabulary through which policymakers, journalists, and the broader public understand artificial intelligence, market competition, and democratic governance itself. The material foundation of this discursive power is staggering in scale: major technology corporations collectively expend approximately $226,000 per day lobbying the United States Congress, a figure that structurally advantages incumbent platforms in regulatory negotiations while raising formidable barriers to meaningful legislative oversight. This expenditure is not static but accelerating, with AI-centered firms dramatically intensifying their Washington presence in direct proportion to the legislative threat posed by emerging regulatory frameworks. Meta Platforms leads this investment by a considerable margin, deploying resources that dwarf those of its competitors and effectively purchasing sustained agenda-setting influence across both chambers of Congress. Notably, even firms that publicly position themselves as safety-conscious or mission-driven are embedded within this same infrastructure of political expenditure: Anthropic recorded $1.56 million in lobbying expenditure in the first quarter of 2026 alone—a figure that reveals the structural compulsion all major AI actors face to participate in Washington’s influence economy regardless of their normative self-presentation. Under the strong alignment between the White House and Silicon Valley’s big-tech oligarchs, the ideological infrastructure is no longer confined to boardrooms and think tanks but is algorithmically amplified across digital publics, ensuring that industry-favorable framings—innovation over regulation, self-governance over state intervention, national competitiveness over civil liberties—circulate at a velocity and scale that civil-society actors, academic critics, and underfunded regulatory agencies are structurally ill-equipped to contest.
5.4. Reshaping the Architecture of American Foreign-Policy Formulation
American foreign policy is now undergoing a fundamental transformation. Foreign policy—historically understood as the domain in which state interest, moral obligation, and geopolitical calculation are negotiated through institutionalized diplomatic and legislative processes—is increasingly being subjected to an efficiency logic imported wholesale from the computational paradigm: one in which speed of implementation, algorithmic optimization, and data-driven resolution selection are elevated as terminal values, while the slower, messier, and irreducibly human processes of ethical deliberation, multilateral consultation, and national-interest adjudication are reframed as institutional friction to be engineered away. The New Silicon Triangle does not merely advise on foreign policy; it increasingly constructs the decision environment within which foreign-policy choices are perceived as available, feasible, or legitimate—supplying the intelligence platforms, communication infrastructure, satellite networks, and predictive analytics on which the State Department, Department of War, and intelligence community have grown structurally dependent. This dependency is most visible in the expanding role of firms such as Palantir, Microsoft, Google, and Amazon Web Services across classified and unclassified government operations.
The embedding of the New Silicon Triangle into the U.S. foreign-policy decision-making system is not merely a simple infiltration of interest groups; it is a profound restructuring of decision-making rationality. From the shrinking of decision-making units to the algorithmic nature of information, from the populist aggregation of preferences to the instrumental selection of solutions, this process systematically alters fundamental questions such as who makes the decisions, on what basis the decisions are made, and how the decisions are made. The impact will transcend Trump’s second term, shaping the long-term trajectory of U.S. foreign-policy formulation.
6. Strategic Consequences
The previous section identified four embedding mechanisms through which the New Silicon Triangle has been constituted. This section analyzes the strategic consequences of that constitution. Five consequences are particularly significant; together they constitute what this paper terms the “Silicon Triangle’s strategic dilemma.”
6.1. Cognitive Closure and Epistemic Narrowing
The first and most fundamental consequence is the narrowing of the cognitive space within which foreign-policy options are perceived. When the intelligence platforms, target-recommendation systems, simulation environments, and predictive analytics on which decision-makers rely are produced by a small oligopoly of firms operating under common commercial and ideological assumptions, the range of strategic alternatives that appear “available” or “feasible” is correspondingly constrained. Operation Epic Fury (February-March, 2026) provides a striking illustration: when Palantir’s Maven Smart System, powered by Anthropic’s Claude, generated approximately one thousand prioritized targets in twenty-four hours, each with weapons recommendations and automated legal justifications, the algorithmic system did not present the human decision-maker with the alternative of not prosecuting these targets or of contesting the underlying classification logic. The presentation format itself encoded a strike-oriented decisional grammar. Where Allison (1971) once analyzed bureaucratic politics in terms of competing organizational worldviews, the New Silicon Triangle produces something more pernicious: a single computationally generated worldview that crowds out alternatives without ever making the displacement visible.
6.2. Alliance Centrifugality
The second consequence is the accelerating erosion of allied trust. The documented episode that most clearly illustrates the causal mechanism is the July 2025 round of unilateral tariff announcements against U.S. treaty allies. On July 7-8, 2025, President Trump posted to Truth Social a series of letters specifying new tariff rates against Japan, South Korea, and twelve additional partners before the U.S. Trade Representative had completed any formal interagency consultation; allied governments learned of the changes from the social-media platform rather than from the State Department or USTR. The causal step from configuration to consequence runs as follows: because the New Silicon Triangle places decisional authority in a small circle of executive and tech-capital actors (DOGE, the reconstituted PCAST, the Sacks-led AI-and-Crypto policy directorate, and the president’s personal social-media account), the traditional alliance-management interfaces—the State Department, the interagency National Security Council process, regional bureaus, embassy-level consultation—are bypassed structurally and not merely stylistically. Allies that previously could lobby for outcomes through professional foreign-service channels now have no predictable interlocutor whose authority survives the next Truth Social post. Pew Research data show favorability ratings for the United States across the G7 declining to multi-decade lows in 2025-26, and Trans-Atlantic technology policy, Indo-Pacific minilateral defense arrangements (notably AUKUS and the Indo-Pacific Economic Framework reformulations), and Latin American economic realignments all show evidence of allies hedging away from U.S.-led coordination toward redundant or alternative arrangements. The centrifugal dynamic is not the result of occasional diplomatic missteps; it is the predictable downstream effect of a decisional architecture in which the institutional vehicles for sustained alliance management have been removed from the loop.
6.3. Policy Fragmentation and Inconsistency
The third consequence is the loss of policy coherence across domains and across time. The documented episode that most starkly demonstrates the pattern is the Anthropic case of February-March 2026: on February 27 the Trump administration directed all federal agencies to cease using Anthropic’s Claude technology and designated the company a “supply chain risk,” yet Pentagon Chief Information Officer Kirsten Davies confirmed to the Senate Armed Services Committee that the Department continued to use Claude operationally during Operation Epic Fury (begun February 28). On March 26, a U.S. district court in California indefinitely blocked the supply-chain-risk designation as unconstitutional, while the operational reliance continued throughout. Within a single thirty-day window the U.S. government thus simultaneously held three contradictory positions on the same firm: an executive blacklisting, an operational dependency, and a judicial repudiation. The causal step from configuration to consequence runs as follows: because the New Silicon Triangle distributes decisional authority across multiple bodies (DOGE, the reconstituted PCAST, the Department of War, individual cabinet secretaries, the president’s social-media account, and—through Detachment 201 and PCAST membership—the firms themselves) and each operates under different temporal logics, stakeholder pressures, and technology dependencies, the synthesis function performed in earlier decades by the National Security Council process is no longer institutionally located anywhere. Comparable contradictions appear across the trade, technology-export, immigration, and federal-personnel portfolios; what Allison called the “governmental politics” model presupposed that bureaucratic actors, however competitive, were operating within a shared institutional environment. The New Silicon Triangle dissolves that shared environment, and the Anthropic episode is its visible expression.
6.4. Soft-Power Depletion
The fourth consequence is the accelerating depletion of U.S. soft power. Nye’s concept of soft power located American international attractiveness in the credibility of institutional commitments and the appeal of values; the causal step from configuration to consequence runs through the destruction of those commitments. Specifically: because the New Silicon Triangle treats efficiency and short-term transactional yield as terminal values (the DOGE logic) and treats foreign-policy instruments as legitimate targets for that efficiency calculus, soft-power infrastructure—which by its nature produces returns only on multi-year time horizons—is structurally vulnerable to cuts that no NSC- or State-Department-led process would have authorized. Pew survey data show favorability ratings for the United States declining across most allied publics in 2025-26. Whereas previous administrations could repair such damage through diplomatic outreach and value-based discourse, the New Silicon Triangle’s institutional logic actively undermines the conditions of such repair: a foreign-policy apparatus that has dismantled USAID, marginalized the State Department, and outsourced strategic communications to a presidential social-media account cannot credibly project the institutional and normative coherence on which soft power depends.
6.5. Domestic Democratic Erosion
The fifth and longest-term consequence is the erosion of the domestic democratic conditions on which legitimate foreign-policy authority depends. The documented episode that integrates the three driving mechanisms is the Detachment 201 commissioning of June 2025 considered alongside the subsequent Military.com reporting (June 27, 2025) confirming that the four commissioned senior officers—Sankar (Palantir CTO), Bosworth (Meta CTO), Weil (OpenAI CPO), and McGrew (Thinking Machines Lab advisor)—would not be required to recuse themselves from their employer firms’ Pentagon business dealings. The causal step from configuration to consequence runs through the institutionalization of non-recusal: when private commercial decision-makers acquire formal military command authority while retaining their commercial portfolios, the line between private interest and public office is not merely blurred episodically by a revolving door but is structurally dissolved by design, and the constitutional principle that public officials act on behalf of the public—rather than as agents for private firms whose interests they continue to represent—loses its institutional anchoring. Three reinforcing mechanisms compound this dynamic. First, Congress is marginalized in defense and AI policy: the reconstituted PCAST and the Sacks-led advisory channels operate as instruments of the President, the Defense Production Act’s emergency procurement authorities are used to bypass appropriations debates, and major foreign-policy decisions are announced on a privately owned platform before they are discussed in any legislative venue. Second, platform-owning capital exerts asymmetric influence over public discourse: the $226,000-per-day lobbying figure understates the discursive dimension, since it captures formal disclosed expenditure but not the algorithmic amplification of industry-favorable framings across the platforms that those same firms own. Third, Detachment 201 and analogous arrangements institutionalize personnel pathways in which the line between private commercial interest and public military authority has been effectively dissolved. The cumulative effect is a foreign-policy apparatus that is technically more capable but politically less accountable—one in which the public on whose behalf decisions are ostensibly made has progressively less institutional capacity to contest those decisions.
6.6. Synthesis: The Silicon Triangle’s Strategic Dilemma
Taken together, these five consequences constitute what this paper terms the “Silicon Triangle’s strategic dilemma”: the same configuration that delivers tactical efficiency systematically erodes the conditions of strategic effectiveness. Faster targeting and faster decisions are real, but they are achieved through a decision-making architecture that progressively loses the capacity for the deliberative judgment, the alliance reciprocity, and the democratic legitimacy on which any sustained great-power posture ultimately depends. This is why the New Silicon Triangle is best understood not as a solution to the American strategic dilemma but as its most acute contemporary symptom. If the improvement of decision-making “efficiency” comes at the cost of cognitive openness, value pluralism, and long-term responsibility, it will lead to “precise errors” being executed efficiently in the wrong direction.
7. Conclusion
This paper has identified, theorized, and traced the consequences of a structurally novel configuration in U.S. foreign-policy decision-making: the New Silicon Triangle, composed of frontier AI and adjacent technology firms, the White House, and the Department of War. Drawing on the classical iron-triangle and military-industrial complex traditions, the state-capture and public-private partnership literatures, and the emerging algorithmic-politics and surveillance-capitalism scholarship, the paper has argued that the New Silicon Triangle constitutes a coordinated, multidimensional power configuration—operating simultaneously through structural, institutional, and discursive channels—that cannot be reduced to any single one of its antecedents.
Four embedding mechanisms have been documented: the formal institutional restructuring of executive decision-making through DOGE and the reconstituted tech-capital-led PCAST; the algorithmic monopolization of military power through Project Maven and the classified-network deployment of frontier AI; the algorithm-driven populist mobilization that displaces traditional diplomatic deliberation; and the broader reshaping of the architecture within which foreign-policy choices are even perceived as available. Five strategic consequences have been analyzed: cognitive closure, alliance centrifugality, policy fragmentation, soft-power depletion, and domestic democratic erosion.
The paper’s central argument is that these consequences are not contingent costs of an otherwise beneficial institutional innovation; they are structural properties of the configuration itself. The New Silicon Triangle is not a solution to the American strategic predicament but a profound manifestation of it: a foreign-policy apparatus that is more efficient by computational standards yet less capable of the deliberation, reciprocity, and accountability that sustained great-power effectiveness ultimately requires.
Three implications follow. For Foreign Policy Analysis as a field, the New Silicon Triangle suggests that the analytical vocabulary of bureaucratic politics requires substantial revision to accommodate non-state algorithmic actors as decisional protagonists rather than as mere instruments. For policymakers seeking to address the configuration’s pathologies, the implication is that incremental reforms to specific procurement or oversight mechanisms will be insufficient: the configuration is held in place by mutually reinforcing structural, institutional, and discursive mechanisms, and any serious response must address all three simultaneously. For the broader study of contemporary international politics, the New Silicon Triangle suggests that the most consequential transformations of state power in the twenty-first century may be occurring not in the visible domain of grand strategy but in the institutional infrastructure through which strategy is computed.
The argument advanced here is necessarily provisional. Many of the institutional innovations analyzed—DOGE, the reconstituted PCAST, Detachment 201, Operation Epic Fury, the Anthropic supply-chain-risk designation—are recent, ongoing, and incompletely documented in public sources, and several of the empirical claims advanced should be regarded as preliminary pending further documentary access. Future research should pursue three lines of investigation: comparative analysis of analogous configurations in other major powers (particularly China and the European Union); longitudinal tracking of whether the institutional innovations identified here outlast the Trump administration’s second term; and more granular empirical analysis of specific decisional episodes in which Silicon Triangle dynamics can be observed in operation. What the present paper claims to establish is not the final word on this configuration but the analytical vocabulary—New Silicon Triangle, embedding mechanism, the structural-institutional-discursive power triad—within which more granular investigations can proceed. Overcoming the Silicon Triangle’s strategic dilemma, finally, requires the rebuilding of the public nature of decision-making structures, the openness of cognitive frameworks, and democratic accountability for technological power—the very elements that the Silicon Triangle’s embedding has systematically weakened.
Funding Information
This paper is supported by “the Fundamental Research Funds for Beijing Municipal Universities” (QNTD25A002).