English Edition
Published: June 28, 2026
Publisher: Japan Institute for Crisis Management
A leaked invitation-only network linking NATO’s top commander, a U.S. Treasury Secretary, and Peter Thiel. A four-stage diagnostic framework traces how the Dark Enlightenment ideology moves from Silicon Valley to Buenos Aires — with early signs now emerging in Japan.
A Japanese-language version of this paper is available here.
Table of Contents
A Note on the Scope and Method of This Paper
Novelty and Aims of This Paper
Introduction: Framing the Question
Chapter One — The Ideological Infiltration Index: A Four-Stage Diagnostic Framework
Stage ①: Ideological Contact
Stage ②: Infrastructural Substitution
Stage ③: Policy Codification
Stage ④: Direct Political Engagement
Operational Policy of the Framework
Chapter Two — Anatomy of an Ideology: From Libertarianism to the Dark Enlightenment
2-1 Point of Origin: Thiel’s 2009 Essay
2-2 Systematization: Curtis Yarvin and Neoreaction
2-3 Translation into Political Power: The Vance Channel — A Reservation About Its Strength
2-4 Codification into Governance: Project 2025
Chapter Three — The Ideology’s First Practice: New Zealand as a Precedent
3-1 “No Country Aligns More with My View of the Future”: Citizenship as the Smallest Unit of Contact
3-2 Connections to Intelligence and Defense Agencies: An Approach Toward Stage ②
3-3 A Consistent Practice of Building Governing Units Outside National Borders
Chapter Four — An Institutionalized Network: Dialog
4-1 The Existence and Structure of Dialog
4-2 An Analysis of Participant Composition
4-3 A Working Hypothesis on Dialog’s Function
Chapter Five — Applying the Index (I): Argentina, a Case That Has Progressed Through Three Stages
5-1 Why Argentina Was Chosen
5-2 Stage ①: Ideological Contact
5-3 Stage ②: Infrastructural Substitution
5-4 Stage ③: Policy Codification
5-5 Approaching Stage ④: The Isaac Accords
5-6 Assessment
Chapter Six — Applying the Index (II): Japan, Early-Stage Signs
6-1 Why Japan Will Not Follow the “Argentine Pattern”
6-2 Stage ①: Ideological Contact
6-3 Stage ②: Infrastructural Substitution — A Deeper Mutual Dependence Than Initially Assessed
6-4 Stage ③: Policy Codification (Currently Unconfirmed)
6-5 A Comparison Table of the Two Paths
6-6 Assessment
Chapter Seven — Three Layers Where the Threat Appears
7-1 The Substitution Layer: Space Infrastructure
7-2 The Personal-Data Layer: Insurance and Financial Infrastructure
7-3 The Military-AI Layer: The Asymmetric Erosion of Alliances
Chapter Eight — Japan’s Structural Blind Spot: Why It Is Hard to See
8-1 Cultural Factors and the Gap Between Domestic and Foreign Awareness of Ideological Terminology
8-2 A Second Structural Vulnerability — the Effectiveness of Institutions: The Case of Broadcasting Foreign-Ownership Restrictions
Chapter Nine — Two Types of Autonomy: India and France as Two Precedents
9-1 The Indian Type: Leaving Operational Vulnerability Unaddressed Through Isolation
9-2 The French Type: Reconciling the Two Through Institutional Separation, Though Operational Challenges Remain
9-3 The Essential Difference Between the Two Types
9-4 A Third Risk Beyond the Two Types: The Supply-Chain Blind Spot of Foundation-Model Dependence
Chapter Ten — Policy Implications
Recommendation ① Institutionalizing the Operation of the Ideological Infiltration Index
Recommendation ② Institutional Separation of Information from Authority (Applying the French Model)
Recommendation ③ Expanding the Concept Underlying the Economic Security Promotion Act
Recommendation ④ Institutionalizing Ideological Literacy in Diplomacy and Defense
Recommendation ⑤ Making Visible the Foundation-Model Dependence Within AI Systems
Afterword: The Need for Continued Monitoring
A Note on the Scope and Method of This Paper
This paper is a risk analysis that draws a clear line between inference and fact. Throughout the text, statements of established fact are marked “Note: The following is an established fact,” while present-stage inferences or hypotheses are marked “Note: The following is a working hypothesis at the time of writing and requires further verification.” Statements without such a marker are facts that are clear from context or are widely known information.
Novelty and Aims of This Paper
● Analysis of the anti-democratic, anti-egalitarian governing ideology known as the “Dark Enlightenment” (neoreaction) has, to date, been confined to the context of U.S. domestic politics (the theorist Curtis Yarvin, Vice President Vance, the Trump administration, Project 2025). Critical commentary already exists pointing to an ideological affinity between the Dark Enlightenment and the Milei government in Argentina, and the international spread of China-style digital authoritarianism has likewise been addressed by existing analytical frameworks such as that of CNAS. The novelty of this paper is limited to the following: it structures this U.S.-originated anti-democratic governing ideology into an observable four-stage diagnostic index, and applies that index to Japan — a target country that has not previously been discussed in this context.
● Analysis of ideology within the United States, analysis of political upheaval in South America, and analysis of technology-company partnerships in Japan have, until now, existed as separate bodies of literature and reporting. This paper connects them as different expressions of a single underlying ideology and integrates them within one theoretical framework.
● As its own original contribution, this paper proposes the “Ideological Infiltration Index,” a four-stage diagnostic framework. This moves beyond the simple binary of “dependence versus autonomy” by decomposing autonomy itself into an “isolation type” and an “institutional-separation type,” thereby building an analytical tool that judges the progress of ideological infiltration against observable criteria.
● By using this index to position Argentina’s advanced case and Japan’s early-stage signs accurately — without either exaggeration or understatement — this paper presents a verifiable early-warning model.
Keywords
● Libertarianism and the Dark Enlightenment (neoreaction)
● The Ideological Infiltration Index
● The dual structure of technocracy and state dismantlement
● Dialog (the institutionalization of an ideological network)
● The Techno-Sovereign Actor
● Two types of autonomy (isolation type / institutional-separation type)
● Japan’s cultural blind spot in cyber security
Executive Summary
In plain terms: The idea that “the state is inefficient and should be reduced” and the movement to “expand private infrastructure that uses data and AI to replace state functions” appear contradictory, but are in fact a paired structure. Wherever democratic control by the state withdraws, something inevitably moves in. When that “something” is a specific group of unaccountable private actors, the state retains its formal independence while steadily losing its substantive freedom to choose. This paper shows, on the basis of verifiable facts, that this structure was born in the United States, has been implemented in South America, and has begun to show signs of emerging in Japan as well.
● An ideology systematized in U.S. Silicon Valley, known as the “Dark Enlightenment,” rejects democracy as an inefficient system and seeks to replace it with governance by data, AI, and algorithms. This is no longer mere blog commentary: it has directly influenced the thinking of a figure within the U.S. administration (Vice President Vance) and has been translated into a policy document (Project 2025).
● The private, invitation-only organization “Dialog,” co-founded by Peter Thiel — an embodiment of this ideology — had its internal records leaked in June 2026, and its substance was verified by news organizations. A NATO supreme commander, the U.S. Treasury Secretary, and 222 individuals including several U.S. senators have been confirmed as included on its participant list.
● In Argentina, this ideology is being concretely implemented, starting from an ideological resonance with the Milei government and proceeding through the privatization of state telecommunications infrastructure, the entry of a foreign private company (Starlink), and a plan to consolidate citizen data.
● The same pattern’s early signs have begun to be observed in Japan as well, evidenced by the deepening partnership between a major private insurer and Palantir, and by the fact of a courtesy visit by Thiel to Japan’s prime minister.
● This paper presents an “Ideological Infiltration Index” that decomposes this process of infiltration into four stages, and evaluates Argentina and Japan according to their respective stage of progress, without exaggeration.
● In conclusion, the policy Japan should pursue is neither dependence on techno-sovereign actors nor international isolation through self-reliance. A French-style “institutional-separation model” — which maintains international sharing of threat information while never ceding ultimate authority over core infrastructure — is the most realistic guide for Japan.
Introduction: Framing the Question
In June 2026, the internal records of “Dialog” — a private organization co-founded in 2006 by the American investor Peter Thiel — were leaked and verified and reported by WIRED.[8] The leaked participant list included the Supreme Allied Commander Europe of NATO, the U.S. Treasury Secretary, several U.S. senators, and major figures of the “PayPal Mafia,” including Elon Musk.
This news demonstrated, more directly than ever before, a structure in which the private sector, the military, and politics intersect within a single non-public space. But this discovery alone cannot answer one crucial question. Does this ideological network remain confined within the United States, or does it cross borders and spread to other nations? And if it does spread, which nations are chosen, and in what sequence does the process proceed?
The answer to this question is already partially observable. The anti-democratic governing ideology born in the United States — known as the “Dark Enlightenment” or “neoreaction” — has, since 2023, accumulated verifiable ideological and practical contacts with Javier Milei’s government in Argentina. At the same time, in Japan, contact with companies and individuals connected to the same ideological sphere has been progressing in stages since 2019.
Two important gaps exist in the existing scholarship. First, analysis of the Dark Enlightenment has been almost entirely confined to the context of U.S. domestic politics (Yarvin, Vance, Project 2025), and no empirical analysis exists that traces how this ideology spreads across national borders. Second, analysis of South America’s political upheaval and analysis of Japan’s technology-company partnerships have existed as separate bodies of literature, with no analysis connecting the two as a single ideological phenomenon.
This paper aims to fill these two gaps. To do so, it constructs its own original diagnostic framework — the Ideological Infiltration Index — and analyzes Argentina and Japan on that basis. The working hypothesis of this paper is as follows.
Ideology selects nations in which the conditions are ripe, and is implemented by grafting itself onto a political consensus the nation already holds (distrust of the state, demand for efficiency). This implementation is not coercive; it is an observable process that proceeds through multiple stages.
Chapter One — The Ideological Infiltration Index: A Four-Stage Diagnostic Framework
This chapter presents the central contribution of this paper: a diagnostic framework that decomposes the process by which ideology infiltrates a nation into four observable stages.
Stage ①: Ideological Contact
Definition: The stage at which an embodiment of the ideology (a person or organization) builds a personal, informal relationship with a political leader or business figure of the target nation. The ideology has not yet appeared in institutions or policy.
Criteria:
● The frequency and form of bidirectional contact (visiting, or being visited) between the embodiment of the ideology, or its affiliated organization, and the political leaders or business figures of the target nation
● The use, by the target nation’s leadership, of the ideology’s terminology or logic (e.g., adoption of statements such as “the state is inefficient” or “the private sector is faster”)
● A record of participation by figures connected to the target nation in a non-public network
Disqualifying criteria (conditions under which this stage should not be deemed to apply):
● Where the subject matter of the contact is, so far as can be confirmed, limited to technical cooperation or commercial dealings, and no statement concerning the nature of state governance is found at all
● Where the meeting was conducted within the bounds of formal diplomatic protocol, and no statement from the target nation’s leadership expressing ideological sympathy or resonance has been recorded
● Where the contact was a single occurrence, and it is clear that it took place as part of the routine business of the embodiment’s affiliated company rather than as a personal act of the embodiment of the ideology
Operational note: satisfying any single criterion alone does not warrant a finding of Stage ①. A finding of Stage ① requires that multiple criteria be satisfied simultaneously, and that none of the disqualifying criteria above apply. A single visit, in itself, is a routine occurrence between many major powers, and applying any one criterion in isolation has no discriminating power.
Stage ②: Infrastructural Substitution
Definition: The stage at which a function previously carried out by the state (telecommunications, data management, intelligence-gathering) is replaced — through privatization, deregulation, or outsourcing — by a private company connected to the ideological network.
Criteria:
● Whether the privatization of state infrastructure and the market entry of a specific private company are proceeding concurrently in time
● Whether, after the substitution, technical substitutability (the ease of switching to another provider) remains, or has been lost
Disqualifying criteria (conditions under which this stage should not be deemed to apply):
● Where the contract was awarded through a competitive bidding process among multiple candidate companies, and no preferential treatment toward a specific company can be confirmed — an ordinary commercial procurement process does not, in itself, constitute infrastructural substitution
● Where the subject of the contract is not a state function (sovereignty over communications, intelligence-gathering, management of core data, etc.) but is limited to the internal operational efficiency of a single private company (e.g., partial renewal of an internal system)
● Where it can be confirmed that termination of the contract, or switching to another provider, is practically feasible under both the contract’s terms and the relevant technology — the formal existence of substitutability should not be read, on its own, as evidence of irreplaceability
Stage ③: Policy Codification
Definition: The stage at which the ideology, through think tanks or government documents, is translated into policy language that appears politically neutral — “efficiency,” “administrative reform,” “national strength” — and institutionalization proceeds.
Criteria:
● The degree to which concepts of ideological origin are reflected in official government documents or draft legislation
● Whether there is a documented record of individuals or organizations connected to the ideological network being involved in the drafting process of the relevant policy
Disqualifying criteria (conditions under which this stage should not be deemed to apply):
● Where the drafting process of the policy or legislation can be explained on the basis of an existing domestic political consensus independent of any particular ideological current (cross-party agreement, a long-standing matter within the bureaucracy, etc.)
● Where, even though words such as “efficiency” or “administrative reform” are used, their substance does not match the concrete institutional design advocated by the embodiment of the ideology (authoritarian governance, centralized data control, etc.) — superficial similarity of language is not evidence of ideological influence
● Where no involvement whatsoever of individuals or organizations connected to the ideological network can be confirmed in any public record (note that this paper currently applies this disqualifying criterion to Japan’s proposed National Intelligence Agency)
Stage ④: Direct Political Engagement
Definition: The stage at which a corporate-level relationship develops into direct dialogue and policy consultation with the core of the government itself. From this point on, the ideological network has standing access to the state’s own decision-making process.
Criteria:
● What is discussed, and at what frequency, in meetings between the embodiment of the ideology and the government’s top leadership
● Whether the meetings are a one-off act of courtesy, or have developed into a framework of continuing policy consultation
Disqualifying criteria (conditions under which this stage should not be deemed to apply):
● Where the meeting was a single occurrence, and no record of continuing contact can be confirmed over a subsequent period (this paper uses twelve months as a benchmark)
● Where the government side has clearly denied or limited the content of the meeting in a public forum (e.g., parliamentary testimony), and no subsequent evidence emerges of that company’s or ideology’s influence on policy decisions — this is treated as positive grounds for a finding against Stage ④
● Where the contact has taken place in parallel with multiple competing companies or holders of multiple ideological positions, and no single actor has obtained exclusive access
Operational Policy of the Framework
Note: The following is an established fact.
These four stages do not necessarily proceed in a single linear sequence. Nor does a nation’s reaching Stage ④ mean that “infiltration is complete”; rather, it is a warning indicator showing that the range of options available to the state’s decision-making is, in an observable way, beginning to narrow.
This paper secures falsifiability by presenting, for every stage, both a set of criteria and a paired set of disqualifying criteria. The mere fact that “contact occurred” is not, by itself, evidence of any stage. A finding that a given stage has been reached is made only when the criteria are satisfied and none of the disqualifying criteria apply. For example, the SOMPO–Palantir partnership discussed in Chapter Six of this paper was initially assessed as remaining within the bounds of a single company’s outsourcing arrangement; later verification, however, revealed that the relationship involved an equity stake, leading this paper to revise that assessment — this is not a mechanical finding of “Stage ② because a contract exists.”
Note: The following is a working hypothesis at the time of writing and requires further verification.
This paper uses this index to evaluate both Argentina and Japan, but does not claim that the two nations will necessarily follow the same path. The speed and sequence of progression may differ according to each nation’s political and cultural conditions. With this in mind, the index should be understood not as a “prophecy” but as an “observational benchmark for early warning.”
Figure 1. The four-stage flow of the Ideological Infiltration Index, and the stage reached by each nation discussed in this paper (Source: prepared by the author)
Chapter Two — Anatomy of an Ideology: From Libertarianism to the Dark Enlightenment
2-1 Point of Origin: Thiel’s 2009 Essay
Note: The following is an established fact.
In 2009, Peter Thiel contributed an essay titled “The Education of a Libertarian” to Cato Unbound, in which he stated plainly that he “no longer believe[d] that freedom and democracy are compatible.”[1] This marked an ideological turning point that went beyond simple libertarianism (the maximization of individual freedom) and positioned democracy itself as an obstacle to freedom.
2-2 Systematization: Curtis Yarvin and Neoreaction
The intellectual movement known as the “Dark Enlightenment,” or “neoreaction,” is an anti-democratic, anti-egalitarian, anti-progressive current of thought that emerged online in the mid-2000s.[2] The software engineer Curtis Yarvin wrote a blog titled “Unqualified Reservations” from 2007 to 2014 under the pen name “Mencius Moldbug,” positioning American democracy as a “failed experiment” and arguing for its replacement with a “responsible monarchy” resembling corporate governance structures.[3] Multiple academic analyses note that Yarvin and Nick Land drew inspiration from Thiel’s essay declaring that “freedom and democracy are no longer compatible.”[2]
2-3 Translation into Political Power: The Vance Channel — A Reservation About Its Strength
Note: The following is an established fact.
U.S. Vice President J.D. Vance is a former protégé of Thiel’s. In a 2021 podcast appearance, Vance cited Yarvin by name and, as advice to Trump, stated that the administration should “fire every single mid-level bureaucrat in the administrative state and replace them with our people.” The New Republic and other outlets have named Thiel and Vance as the foremost figures influenced by Yarvin’s thought.[4]
However, important reservations have been voiced by the parties themselves regarding the strength of this chain of influence. Yarvin himself told The New York Times in 2025 that his relationship with Vance has been “definitely overstated.” Joshua Tait, a journalist who studies Yarvin directly, likewise assesses that “Yarvin is not the grand theorist of DOGE, much less of the Trump administration. I don’t think he has actual, concrete influence.” Similarly, Thiel himself told The Atlantic in 2023 that, while he does not think Yarvin’s ideas would “work,” he finds him an “interesting” and “powerful” historian — hardly the language of an unconditional supporter.
In light of this reservation, this paper avoids depicting the relationship among Thiel, Yarvin, and Vance as a simple “chain of mentorship.” The reality is a more complex structure in which a financial tie — an investment in Yarvin’s company, Tlon — coexists with a partial and selective sharing of ideas. The “export of ideology” discussed in this paper should be understood not as direct ideological transmission between specific individuals, but as a looser ideological proximity mediated through a capital network, namely Founders Fund.
2-4 Codification into Governance: Project 2025
Note: The following is an established fact.
“Project 2025” (released in April 2023 under the title “Mandate for Leadership: The Conservative Promise”), led by the Heritage Foundation, is a comprehensive policy document aimed at reorganizing the U.S. federal government and concentrating executive power in the service of right-wing policy. Multiple analyses note that, in both its philosophical outlook and its policy prescriptions, it bears a striking ideological parallel to Yarvin’s neoreactionary vision — albeit packaged in the language of think-tank policy.[5]
This ideological parallel has a concrete, practical connection that is more verifiable than Vance’s personal influence. Russell Vought, who served as Director of the Office of Management and Budget in the first Trump administration, is the central drafter of Project 2025. The Center for Renewing America (CRA), the think tank Vought leads, produced many of Project 2025’s policies as well as draft executive orders ready for signature immediately upon the administration taking office. Vought’s posture of confrontation with a “corrupt bureaucracy” is substantively consistent with what Yarvin calls the dismantling of “the Cathedral” — the entrenched institutional power complex. Political-research organizations have also noted that several junior staff members of the Department of Government Efficiency (DOGE) have prior experience working under Thiel, or have publicly expressed neoreactionary views.
Once an idea is translated from a blog into a policy document, this paper regards it as having acquired an “exportable form.” A personal expression of belief on a blog and the practical language of governance found in a policy document differ fundamentally in how readily they can spread. Project 2025 is an example of the Dark Enlightenment — an extremist ideology — being converted into the internationally palatable language of “administrative reform,” and its path of institutionalization can be observed more clearly through the organizational, practical channel of Vought and the CRA than through the uncertain line of Vance’s personal influence.
Chapter Three — The Ideology’s First Practice: New Zealand as a Precedent
Before the Dark Enlightenment was institutionalized into a policy document within the United States, Thiel’s personal ideology had already been put into practice, in relation to a single nation, nearly two decades earlier. This chapter confirms this precedent in order to show that the four-stage index proposed in this paper reflects a longer-running, recurring structure that is not confined to Argentina and Japan.
3-1 “No Country Aligns More with My View of the Future”: Citizenship as the Smallest Unit of Contact
Note: The following is an established fact.
In June 2011, Thiel acquired New Zealand citizenship after the country’s Internal Affairs Minister, Nathan Guy, applied an “exceptional circumstances” provision. Ordinarily, a permanent resident applying for citizenship must reside in the country for 1,350 days over five years; Thiel had made only four visits over five years, totaling a mere 12 days of residence. This fact became public in 2017 through investigative reporting by the New Zealand Herald. On his application, Thiel wrote that “no other country aligns more with my view of the future than New Zealand.”[6]
This is the purest example showing that Stage ① (ideological contact) in this paper’s index can be constituted by the legal status of citizenship itself, without substantive residence or institutional involvement. Ideological contact is not measured solely by the number of in-person meetings or visits; it can also take the form of “selecting a particular nation as the place most closely aligned with one’s own ideals, and obtaining legal status from that nation.”
3-2 Connections to Intelligence and Defense Agencies: An Approach Toward Stage ②
Note: The following is an established fact.
Investigative reporting by the New Zealand Herald has revealed that Palantir, Thiel’s core business, held a long-standing relationship with the New Zealand Defence Force, the Security Intelligence Service (SIS), and the Government Communications Security Bureau (GCSB).[6] This serves as a precedent illustrating a pathway by which Stage ① contact — the granting of citizenship to an individual — subsequently connects to a Stage ②-type structure: a relationship between Palantir and the nation’s intelligence and defense agencies.
3-3 A Consistent Practice of Building Governing Units Outside National Borders
Note: The following is an established fact.
Thiel was an early backer of the Seasteading Institute (a concept for building autonomous communities on the high seas), and later turned his interest to charter cities (autonomous cities on land). He is a leading investor in Pronomos Capital, an investment fund aimed at establishing experimental, semi-autonomous cities on undeveloped land; projects backed by the fund include “Próspera” in Honduras. Praxis, a related company, has sought to establish a new city-state and considered Greenland as one candidate site. In 2025, when the Danish government moved to integrate Palantir into its own military, police, and intelligence agencies, one argument raised by opponents was that the Praxis project (which Thiel backs via Pronomos) posed a threat to Greenland’s sovereignty.[7]
Taken together, these facts show that Thiel’s ideology — skepticism of democratic control by the state, and an ideal of freedom of movement unbound by national borders — was not first put into words in his 2009 essay, but had already been accumulating, even before that, as a consistent practice across multiple nations and regions. New Zealand stands as its first and most mature example, with Argentina, roughly fifteen years later, representing a more comprehensive implementation.
Chapter Four — An Institutionalized Network: Dialog
4-1 The Existence and Structure of Dialog
Note: The following is an established fact.
“Dialog” is a private, invitation-only organization co-founded in 2006 by Peter Thiel and Silicon Valley angel investor Auren Hoffman. It kept its membership undisclosed for twenty years until, in 2026, the Swiss hacktivist maia arson crimew discovered a public directory embedded in the source code of dialog.org, after which WIRED independently verified and reported its contents.[8]
4-2 An Analysis of Participant Composition
WIRED’s verification confirmed that the registrant list for a leaked retreat scheduled for August 2026 (to be held near Dublin, Ireland) named 222 individuals, including NATO’s Supreme Allied Commander Europe and Commander of U.S. European Command (attending since 2021), the U.S. Treasury Secretary, several senators (including Ted Cruz), and Palantir co-founder Joe Lonsdale.[9] A separately leaked directory of 113 affiliated individuals was confirmed by multiple news organizations to include Elon Musk, Senator Cory Booker, a Stanford University president, and the writer Ezra Klein. Session topics on the agenda included “Navigating WWIII” and “Build-a-Cult.”
This paper avoids characterizing this gathering as a “cult.” Even if some session titles contained ironic language, treating that language as proof of the organization’s true nature would be an interpretation that exceeds verifiable fact. What this paper emphasizes is the objectively confirmable structure: that the private sector, the military, and politics are intersecting continuously within one and the same non-public space.
4-3 A Working Hypothesis on Dialog’s Function
Note: The following is a working hypothesis at the time of writing and requires further verification.
No direct evidence currently available in the public record confirms that Dialog is a venue where concrete policy decisions are made. However, this paper positions Dialog as an organization that may function as a human-network foundation for ideology as it crosses national borders. This is a hypothesis, not a settled conclusion.
Figure 2. The genealogy of an idea — the process of translation from blog to policy document (Source: prepared by the author)
Chapter Five — Applying the Index (I): Argentina, a Case That Has Progressed Through Three Stages
5-1 Why Argentina Was Chosen
Note: The following is a working hypothesis at the time of writing and requires further verification.
This paper proposes the hypothesis that Argentina was not selected by “geographic coincidence,” but because its existing political conditions had lowered the barriers to ideological infiltration more than almost any other nation. Specifically, three conditions converged: (1) an existing political climate of “distrust of the state” produced by decades of economic crisis and hyperinflation; (2) an existing international political consensus favoring “shrinking the state,” arising from IMF demands for structural reform; and (3) a personal ideological resonance between Milei and Thiel.
5-2 Stage ①: Ideological Contact
Note: The following is an established fact.
Peter Thiel, himself a fellow libertarian, has had repeated contact with Argentine President Javier Milei, beginning with a meeting at the Casa Rosada (presidential palace) in January 2024. They met again at the Milken Institute forum in Los Angeles in May 2024, and once more at the Casa Rosada on April 23, 2026.[10] After their April 2026 meeting, Milei called it “a wonderful meeting” and said of Thiel, “He is, like me, philosophically an anarcho-capitalist.” The two are reported to have discussed wealth taxation, the culture war, and the future of liberalism. From his campaign through his time in office, Milei has consistently used the phrase “the state is the enemy,” and has repeatedly staged a performance of dismantling government agencies, including a campaign speech in which he wielded a chainsaw.
Note: The following is an established fact.
Here, too, the application of this paper’s disqualifying criteria becomes important. Milei himself has stated that no specific investment opportunities were discussed in the April 2026 meeting, explaining that the conversation centered on how the two might sustain the political philosophy they share. Opposition legislator Juan Marino, however, demanded an explanation from the Foreign Ministry over the visit, asking whether the introduction or purchase of Palantir’s surveillance and defense technologies had been on the agenda. As of this writing, no official government response to that question has been confirmed.
5-3 Stage ②: Infrastructural Substitution
Immediately after Milei’s 2023 election victory, Elon Musk posted on X that “prosperity is coming to Argentina.”[11] It should be noted that Starlink had already obtained an operating license from Argentina’s national telecommunications regulator (ENACOM) in December 2021, and did not newly enter the market with the launch of the Milei government. Nonetheless, Starlink is one of the few private companies Milei has named in his speeches, and the launch of its commercial service in March 2024 should be understood as a case in which an already-licensed company was politically deployed as a symbol of the Milei government’s deregulatory agenda.
Note: The following is an established fact.
A more direct instance meeting the criteria for Stage ② can be found in the state’s intelligence infrastructure itself. On December 31, 2025, while Congress was in recess, the Milei government issued Emergency Decree (DNU) 941, amending the National Intelligence Law. This decree obligated more than fifteen state agencies — including ANSES (Social Security Administration), AFIP (the tax authority), RENAPER (the national identity registry), and the public health system — to share personal data with the State Intelligence Secretariat (SIDE) without a court order and without notifying citizens.[12]
5-4 Stage ③: Policy Codification
In April 2026, Peter Thiel visited Buenos Aires and met with the President and the head of SIDE. The Argentine legal-professionals association (AGC) has noted that, while no public contract exists, DNU 941/2025 has created a regulatory environment that aligns almost perfectly with Palantir’s business model — large-scale data integration, AI-driven predictive analysis, and automated targeting decisions.[13]
This observation matters for the theory of this paper. The progression of Stages ② and ③ need not take the form of a public contract. The same effect can arise when the regulatory environment itself is designed to fit a particular private actor’s business model “like a key fits a lock.”
Note: The following is an established fact.
On May 22, 2026, the Milei government announced the launch of “Gemelo Digital Social” (Social Digital Twin), an AI platform that consolidates information dispersed across state agencies in order to design and forecast social policy. The announcement took the form of a promotional video on X, and no reference to a legal basis, a public tender, or data-protection protocols could be found within either the promotional video or the public information this paper was able to confirm (the possibility that some other official procedural document exists separately from the promotional video cannot be ruled out). The announcement coincided with Thiel’s presence in Argentina.[14]
5-5 Approaching Stage ④: The Isaac Accords
Note: The following is an established fact.
On April 19, 2026, President Milei and Israeli Prime Minister Netanyahu formally announced the “Isaac Accords” in Jerusalem. This trilateral framework among the United States, Israel, and Latin America governs cooperation on trade, technology, and security, and includes a memorandum on security cooperation and artificial intelligence.[15] This shows that Argentina’s structural transformation has expanded beyond a bilateral personal relationship into a multilateral institutional framework.
Note: The following is a working hypothesis at the time of writing and requires further verification.
The Isaac Accords are themselves a diplomatic treaty between states, and no primary source shows any direct connection to Thiel’s ideological network. The reason this paper nonetheless treats the Accords as an indicator of Stage ④ lies in the soil from which they grew. The Palantir-style regulatory groundwork the Milei government has steadily built — DNU 941 and Gemelo Digital Social (Stages ② and ③) — has already established, as a fait accompli, a governing posture of “opening state functions outward,” and the Isaac Accords can be positioned as the result of that same posture becoming multilateral in the security and AI domain as well. In other words, this paper does not claim that the Isaac Accords are themselves a product of Thiel’s ideological sphere; rather, it treats them as a parallel institutionalization arising from the same political soil — a corroborating indicator of an approach toward Stage ④.
5-6 Assessment
Argentina can be assessed, under this paper’s index, as a case in which Stages ① through ③ have clearly progressed, and which has reached Stage ④ — direct, continuing policy consultation with the government (note that part of this connection to Stage ④ uses the Isaac Accords as its indicator, which rests on the working hypothesis discussed in section 5-5). This is, at present, the only observable case in which the “ideological infiltration” discussed in this paper has progressed to its most advanced, near-complete form.
Figure 3. A combined timeline — the accumulation of contacts in Argentina and Japan (Source: prepared by the author)
Chapter Six — Applying the Index (II): Japan, Early-Stage Signs
6-1 Why Japan Will Not Follow the “Argentine Pattern”
Note: The following is an established fact.
Japan has no political consensus, of the kind seen in Argentina, favoring the “dismantling of the state.” There is no external pressure comparable to IMF-mandated structural reform, nor is there a broadly shared political culture of fundamental distrust toward state institutions. Stating this clearly is essential to ensuring that the analysis in this paper is not a simple prophecy that “Japan, too, will follow Argentina’s path.”
6-2 Stage ①: Ideological Contact
On March 5, 2026, Peter Thiel paid a courtesy visit to Prime Minister Sanae Takaichi at the Prime Minister’s Office. For Thiel, who had previously visited Japan several times for business or speaking engagements before 2019, this was his first formal contact with a government official.[16] The meeting reportedly lasted about 25 minutes, and is reported to have involved an exchange of views on Japan-U.S. scientific and technological prospects, including small modular reactors, ahead of the Prime Minister’s planned visit to Washington.
This fact alone cannot determine whether the “ideological resonance” required for Stage ① is present. The reported content of the meeting was scientific and technological cooperation, and no clear adoption of ideological language — of the kind seen in Milei’s case — has so far been confirmed. This case falls within the disqualifying criterion for Stage ① set out in Chapter One: “the subject matter of the contact is limited to technical cooperation or commercial dealings, and no statement concerning the nature of state governance is found.” The Japanese case therefore amounts to no more than the most preliminary sign of Stage ① — the mere occurrence of contact — and, measured against the disqualifying criteria, should be judged not yet to have reached the stage of ideological resonance.
Note: The following is an established fact.
This meeting has already become an open political issue within Japan’s Diet. At a House of Councillors budget committee session on March 17, Prime Minister Takaichi, responding to a question from Sohei Kamiya, leader of the Sanseito party, stated clearly that “there was no discussion whatsoever of using [Palantir’s] services.” The Prime Minister also revealed that she had been briefed on Thiel as someone who “is on good terms with Vice President Vance” and “was one of the architects of the Trump administration.” This is an important fact, showing that concern over the possible adoption of Palantir is already a subject of open political debate in Japan.
Note: The following is an established fact.
This contact has not run only in one direction, from Thiel toward Japan. On January 16, 2026 (local time), Defense Minister Shinjiro Koizumi, while visiting the United States, toured Palantir Technologies’ Washington, D.C. headquarters and exchanged views with company executives on the use of AI in the security field. It has been confirmed, both through Minister Koizumi’s own social-media posts and through an official announcement by the Ministry of Defense, that he was received by the company’s head of defense — a former U.S. congressman — Mike Gallagher.[33]
If the criteria for Stage ① set out in Chapter One of this paper are redefined to include not only visits by an embodiment of the ideology, but also bidirectional contact that includes visits from the target nation, then this visit, too, can be regarded as an additional indicator for Stage ①. It must be stressed, however, that what occurred here was the fact that “Japan’s defense minister visited the headquarters of a company Thiel co-founded,” not the fact that “Japan’s defense minister met with Thiel.” No record of Thiel himself having been present can be found in the Ministry of Defense’s official announcement, in Kyodo News’ reporting, or in Minister Koizumi’s own social-media post, and blurring this distinction would run contrary to the principle of falsifiability that this paper holds central.
6-3 Stage ②: Infrastructural Substitution — A Deeper Mutual Dependence Than Initially Assessed
Note: The following is an established fact.
On November 18, 2019, SOMPO Holdings and Palantir Technologies agreed to jointly establish a venture, Palantir Technologies Japan, to operate a data-analytics platform business. The two companies each contributed roughly half of approximately ¥10.8 billion in capital. The press conference was attended by Thiel himself, along with SOMPO Group CEO Kengo Sakurada and Group CDO Koichi Narasaki. In the question-and-answer session that followed, Thiel addressed concerns about the handling of personal information drawn from SOMPO’s real-world data, stating that he “respect[s] privacy as a philosophical principle.”[17]
This relationship runs deeper than this paper initially assumed. In 2020, SOMPO invested $500 million (roughly ¥54 billion) directly in Palantir itself — an equity relationship that goes beyond a mere business partnership. Palantir, for its part, expanded its partnership with SOMPO in 2023, and in August 2025 announced a multi-year contract as a “second expansion.”[18] As of August 2025, the number of users within SOMPO Group, Sompo Japan Insurance, SOMPO Care, and Fujitsu in Japan had reached approximately 8,000.
The SOMPO case should therefore be reassessed not as a simple case of “outsourcing by a single company,” but as a case that has progressed substantially within Stage ②, combining equity investment, management participation, and large-scale operational integration. It is nonetheless important to note a difference: whereas Argentina’s case (DNU 941 and Gemelo Digital Social) involves a presidential decree that legally mandates data-sharing among state agencies — a substitution of state function itself — the SOMPO case remains a voluntary management decision by a private company, without the exercise of state coercive power. Regarding the disqualifying criterion for Stage ② set out in Chapter One — “where termination of the contract, or switching to another provider, is confirmed to be practically feasible” — whether SOMPO could unilaterally dissolve its relationship with Palantir, given the scale of the capital relationship involved, may in practice be more complex than the formal existence of substitutability would suggest. This paper does not reach a firm conclusion on this point and flags it as a matter for future verification.
Note: The following is an established fact.
Following SOMPO, other major Japanese companies have also deepened their relationship with Palantir. Fujitsu began collaborating with Palantir in 2020, signed a global partnership in December 2023, and on August 19, 2025, signed a license agreement for the generative-AI platform Palantir AIP (targeting roughly $100 million in revenue by the end of fiscal year 2029). Fujitsu has further been awarded a commissioned research contract from the Acquisition, Technology & Logistics Agency’s Defense Innovation Science and Technology Institute, titled “Research on Acquiring AI Staff Capability through Defense Multi-Agent AI.”[19] It should be noted that, so far as is publicly known at this time, this commissioned defense research and the Palantir AIP license agreement are separate contracts, and no primary source confirms any technical connection between the two. Sumitomo Corporation, too, signed a license agreement for AIP with Palantir Technologies Japan through a group company on March 19, 2026, and is proceeding to introduce it into supply-chain management for its energy-related (oil-pipe) business.[20]
Note: The following is information that could not be confirmed by an independent primary source at the time of this paper’s writing.
A number of web-based media outlets have published essentially identical claims that Japan’s Ministry of Defense is considering introducing Palantir’s “Gotham” platform under the guise of “integrated analysis of disaster-related information.”[21] Tracing these claims, however, did not lead to any official statement or budget document from the Ministry of Defense, nor to any independent investigative reporting by a major national newspaper. The fact that multiple outlets repeat the same wording does not constitute multiple independent confirmations; rather, the possibility cannot be ruled out that a single, unverified claim is simply being repeated across web media (so-called “citogenesis”). This paper does not treat this information as a confirmed instance of Stage ③, and records it only as an observation that warrants further verification.
6-4 Stage ③: Policy Codification (Currently Unconfirmed)
In 2026, under the Takaichi government, the creation of a “National Intelligence Agency” is moving forward. The National Intelligence Council Establishment Act passed the House of Councillors plenary session on May 27, 2026, with support from the Liberal Democratic Party, the Japan Innovation Party, and the Democratic Party for the People, and the government plans to launch the agency as early as the summer of that year.[22] At a press conference, Prime Minister Takaichi stated that this had been a policy she had repeatedly pledged during her party’s presidential election, calling the law “a first step,” and indicating that she would consider, going forward, enacting an anti-espionage law and establishing a foreign-intelligence service. Some opposition parties raised concerns about privacy infringement. This initiative had already been a campaign pledge by the Prime Minister herself before the Takaichi cabinet took office (October 2025), and falls clearly within the disqualifying criterion for Stage ③ set out in Chapter One: “where the drafting process of the policy or legislation can be explained on the basis of an existing domestic political consensus independent of any particular ideological current.” No primary source whatsoever currently confirms the involvement of any specific private actor or ideological network.
According to reporting, a dynamic separate from the theoretical framework of this paper — a domestic turf war among ministries — is at play, with the Public Security Intelligence Agency, the National Police Agency, the Ministry of Foreign Affairs, and the Ministry of Defense all actively maneuvering to take the leading role in the new agency.[23]
Note: The following is a working hypothesis at the time of writing and requires further verification.
Which company will, in the future, be responsible for building the National Intelligence Agency’s systems remains undisclosed at this time. In keeping with the index this paper proposes, monitoring future public information on procurement and personnel constitutes the concrete practice of early warning. However, there is currently no basis for finding that this meets the criteria for Stage ③.
6-5 A Comparison Table of the Two Paths
| Element | Argentina | Japan |
| Stage ① (Ideological Contact) | Confirmed (3 visits, statements of ideological resonance) | Contact confirmed (1 courtesy visit by Thiel + 1 visit to Palantir HQ by Defense Minister Koizumi); ideological resonance unconfirmed |
| Stage ② (Infrastructural Substitution) | Confirmed (privatization of state telecoms + Starlink entry) | Limited (outsourcing by a single company; not a substitution of state function) |
| Stage ③ (Policy Codification) | Confirmed (Digital Twin plan) | Unconfirmed (no confirmed involvement of a private actor in the National Intelligence Agency concept) |
| Stage ④ (Direct Political Engagement) | Confirmed (continuing policy consultation + the Isaac Accords; the latter rests on a working hypothesis) | Unconfirmed (a single courtesy visit only) |
| Overall Assessment | An advanced case (Stages ③–④) | An early-stage sign (part of Stage ①; a limited instance of Stage ②) |
6-6 Assessment
Under this paper’s index, Japan shows only a limited sign of Stage ① and an extremely limited instance of Stage ② (outsourcing by a single company). This is not grounds for concluding that “a crisis is underway,” but it does indicate a focal point that warrants continued monitoring as a matter of early warning.
Chapter Seven — Three Layers Where the Threat Appears
Building on the analysis so far, this chapter organizes the three layers in which the export of ideology is technically implemented. These three layers can be understood as the concrete forms that Stage ② (infrastructural substitution) in the Chapter One framework takes within each different domain. Across the differing domains of telecommunications, data, and military AI, the underlying structure is common: state functions are being replaced by a specific group of private actors. This chapter shows the shape that structure takes within each domain.
7-1 The Substitution Layer: Space Infrastructure
The substitution of communications infrastructure extends not only to the physical layer of undersea cables, but also to the substitute layer of satellite communications. NATO announced “Project HEIST” in 2024, which would reroute high-priority data to satellites in the event of an undersea-cable failure; however, concerns have been raised that Russia could simultaneously escalate in both the undersea and space domains, leaving this strategy vulnerable to a “Catch-22” — a bind in which either choice leads to a dead end.
7-2 The Personal-Data Layer: Insurance and Financial Infrastructure
As the SOMPO–Palantir case shows, infiltration of the personal-data layer is the least conspicuous pathway, requiring no political consensus. The SOMPO case, moreover, involves not merely outsourcing but an equity relationship, showing that corporate-level dependence can reach a depth that cannot be unwound by a simple contract termination. This shows that Stage ② — “infrastructural substitution” — can proceed at the corporate level, ahead of, and in a more complex form than, the substitution of state functions themselves.
7-3 The Military-AI Layer: The Asymmetric Erosion of Alliances
Japan, South Korea, and Australia already have contractual relationships with Palantir in the military and security domain. Should this dependence deepen, constraints on action may arise on a dimension separate from treaty-based alliance obligations — a phenomenon this paper terms “the asymmetric erosion of alliances through technical constraint.”
Chapter Eight — Japan’s Structural Blind Spot: Why It Is Hard to See
8-1 Cultural Factors and the Gap Between Domestic and Foreign Awareness of Ideological Terminology
Note: The following is an established fact.
The terms “Dark Enlightenment” and “neoreaction” are not entirely unknown within the Japanese-language sphere. An independent entry, “Ankoku Keimo” (Dark Enlightenment), exists on Japanese Wikipedia,[24] Satoshi Kizawa has published an explanatory book, Nick Land and Neoreaction: The Dark Ideas Covering the Contemporary World (Seikaisha Shinsho, 2019),[25] and several introductory articles can be found on individual blogs such as note.
Within the scope of this paper’s research, however, all of these references remain confined to an intellectual-historical or subcultural context (philosophical works, individual commentary). No reference connecting this ideology to Palantir’s expansion in Japan, or to the National Intelligence Agency concept, could be found in the political or security sections of Japan’s major newspapers, or in the analyses of think tanks affiliated with the Ministry of Defense or the Ministry of Foreign Affairs. This paper’s conclusion is based on a limited web-search investigation, and reflects findings as of the time of writing rather than a permanent judgment.
Note: The following is a working hypothesis at the time of writing and requires further verification.
Japan’s political culture, which places weight on harmony, the collective, and consensus, has a fundamental incompatibility with an explicitly anti-democratic ideological language such as the Dark Enlightenment. This cultural distance may, paradoxically, produce a “lack of ideological immunity.” An ideology with which one has no affinity is less easily recognized as a threat, creating a risk of misjudgment in dealings with an actor driven by that ideological motive.
In other words, the “blind spot” in Japan does not lie in any absence of introduction to the ideology itself, but in the lack of a bridge between intellectual-historical knowledge and vigilance in the practice of security policy. It is precisely this missing bridge that motivates this paper’s proposal to build an early-warning model.
8-2 A Second Structural Vulnerability — the Effectiveness of Institutions: The Case of Broadcasting Foreign-Ownership Restrictions
Note: The following is an established fact.
Distinct from, but related to, the dependence on a specific actor connected to an ideological network that this paper discusses, a structural vulnerability can also be found in the operation of Japan’s own institutions. The Broadcast Act and the Radio Act require terrestrial and satellite broadcasters to keep foreign voting rights below 20%, in order to prevent foreign capital from controlling the press. In 2021, however, it came to light that Fuji Media Holdings had exceeded this ceiling for roughly two years, from September 2012.[26] Tohokushinsha was likewise found that same year to have violated the foreign-ownership restriction in a separate case, resulting in the partial revocation of its satellite-broadcasting authorization.
What matters is that even after Fuji’s violation came to light, the institutional penalty of license revocation was never actually applied. This is a concrete example, within the information and media domain, of a structural pattern this paper has repeatedly noted: a gap between the formal existence of regulation and its actual enforcement. In 2022, the Ministry of Internal Affairs and Communications revised the relevant law to require periodic reporting of foreign-ownership ratios, but this strengthened only after-the-fact reporting, not the immediate application of penalties.
This case belongs to a different context from the discussion of techno-sovereign actors, but it reinforces, through an existing domestic Japanese example, the central question of this paper: “Does the institution exist, and does it actually function?” The scope of this paper’s investigation has been concentrated in the domains of information, data, and telecommunications; whether similar structural vulnerabilities exist in other critical-infrastructure domains — transportation, water, energy, and so on — lies outside the scope of this paper and is left as a matter for future verification.
Chapter Nine — Two Types of Autonomy: India and France as Two Precedents
This paper’s research has found that “autonomy” — the avoidance of dependence on a techno-sovereign actor — can take two entirely different forms.
9-1 The Indian Type: Leaving Operational Vulnerability Unaddressed Through Isolation
Note: The following is an established fact.
India is known as a nation with a strong commitment to domestically developed platforms (such as the UPI payment system). In April 2022, however, the U.S. cybersecurity firm Recorded Future (through its Insikt Group threat-research unit) disclosed that a hacking group suspected of links to the Chinese state — which the firm designated “TAG-38” — had been conducting sustained intrusions into seven State Load Despatch Centres (SLDCs, critical facilities responsible for power-grid control and distribution) in the Ladakh border region of India. India’s power minister, R.K. Singh, officially stated that there had been two attempts by Chinese hackers, but that they had “not succeeded.”[27] Recorded Future assessed that this was not ordinary espionage aimed at economic gain, but was instead intended to build capability and pre-position for a future contingency. This is a case showing that even a strongly self-reliant nation cannot escape an organized cyberattack from abroad. Furthermore, according to a 2026 threat report, 62% of all detections within cloud environments were attributable to misconfiguration,[28] showing that a commitment to self-reliance does not necessarily translate directly into improved operational security.
What India’s case shows is that the premise “if we build it ourselves, it is safe” does not hold. Where self-reliance entails isolation from the international sharing of threat information, it can leave operational vulnerabilities unaddressed.
Note: The following is an established fact.
India’s self-reliance was tested again in 2026, from a different direction. In June 2026, following an export-control directive from the U.S. government, Anthropic cut off all access to its top-tier models (Fable 5 and Mythos 5) from outside the United States. As a result, Indian developers and enterprises — including those involved in a large-scale training partnership with Tata Consultancy Services (TCS), which had integrated Claude into its operations — lost access immediately after signing their agreements.[34] Mohandas Pai, former CFO of Infosys, called on Prime Minister Modi to create an annual “sovereign AI fund” worth roughly ₹500 billion, and Zoho co-founder Sridhar Vembu stated that “technology has become the ultimate weapon of sovereignty.”[35]
This is a case running in a different direction from this paper’s argument that “self-reliance leaves operational vulnerability unaddressed,” and must be carefully positioned. The TAG-38 case shows that “building it yourself does not guarantee safety”; the Anthropic case, conversely, shows that “depending on advanced foreign technology can mean losing access at the stroke of a single geopolitical decision” — a fact that works in favor of self-reliance. In other words, self-reliance carries the cost of operational vulnerability, while dependence carries the cost of geopolitical access denial. The “third path” this paper proposes — institutional separation of information from authority — is an attempt to avoid both costs simultaneously.
9-2 The French Type: Reconciling the Two Through Institutional Separation, Though Operational Challenges Remain
Note: The following is an established fact.
In its 2026–2030 National Cybersecurity Strategy, France frames cooperation with the EU as a “fight for sovereignty,” and places weight on threat-information sharing at the European level through mechanisms such as the CSIRT network.[29] At the same time, under the “Cloud au Centre” doctrine launched in 2021, when state administrative bodies and public-interest organizations handle sensitive data — legally protected secrets, or data necessary to the essential functions of the state — they must choose a cloud service that has implemented measures effectively preventing unauthorized access by non-EU authorities. In practice, this means using a provider that holds the “SecNumCloud” certification (established in 2017; the current version, 3.2, includes roughly 1,200 requirements) issued by ANSSI, France’s national cybersecurity agency. This constitutes a strict sovereignty requirement intended to prevent extraterritorial interference under instruments such as the U.S. CLOUD Act and FISA.[30] As of 2026, only nine providers hold SecNumCloud certification, with twelve more under review; the market remains small, at roughly €18 million within France, against a wider European sovereign-cloud market of approximately €12.4 billion.
It must, however, be stated accurately that France’s operational record has both a bright and a dark side. In the first quarter of 2024, France Travail, the country’s public employment-services agency, suffered the largest data breach in the history of the French public sector, through a social-engineering technique that deceived CAP EMPLOI staff into surrendering their credentials, exposing the personal data — social-security numbers, names, dates of birth, email addresses, and phone numbers — of roughly 43 million people. France’s data-protection authority, CNIL, concluded on January 29, 2026, following its investigation, that there had been a “disregard for basic security principles,” citing deficiencies in authentication processes and a lack of anomaly-detection logging, and imposed a fine of €5 million.[31] This is a fact showing that even an institutional-separation model cannot fully prevent operational failure — particularly failures of human and organizational management, as opposed to technical vulnerability.
By contrast, at the Paris Summer Olympics held the same year, two ransomware attacks were detected, but ANSSI assessed that, through the segmentation of information systems and a rapid response, these did not disrupt critical systems or interfere with the operation of the Games. Comparing the two cases shows that France’s strength lies not in “an ability to prevent attacks entirely,” but in “a design that separates critical infrastructure from general systems and limits the spread of any damage.”
9-3 The Essential Difference Between the Two Types
The Indian type is “simply the inverse of dependence.” It turns the exclusion of foreign private companies into an end in itself, and in doing so shrinks even the cooperative dimension of international threat-information sharing. The result is that operational vulnerability is left unaddressed.
The French type is “the institutional separation of information from authority.” It participates actively in international cooperation on threat-information sharing, while keeping ultimate authority over access to core data in the hands of the state, governed by clear institutional standards such as SecNumCloud.
Note: The following is a working hypothesis at the time of writing and requires further verification.
The French model, too, is not complete. According to an April 2025 study commissioned by the think tank Cigref from the firm Asterès, 83% of European corporate spending on cloud and software goes to U.S. companies (chiefly Amazon Web Services, Google, and Microsoft Azure), and 80% of the resulting value is generated within the United States — amounting to an estimated €26.4 billion flowing annually into the U.S. economy (equivalent to roughly 1.5% of the EU’s GDP).[32] This shows that, even in France, with its strict sovereignty requirements such as SecNumCloud, a deep dependence on the United States remains within the actual structure of the market. An institutional-separation model does not amount to “complete autonomy.” This paper positions France not as a “model to be emulated,” but as one realistic direction Japan may draw upon.
9-4 A Third Risk Beyond the Two Types: The Supply-Chain Blind Spot of Foundation-Model Dependence
Note: The following is an established fact.
In addition to the two failure modes shown in sections 9-1 and 9-2 — leaving vulnerability unaddressed through isolation, and economic erosion through dependence — a case emerged in 2026 illustrating a third failure mode, distinct from either. As described in section 9-1, in June 2026 a U.S. government export-control directive caused Anthropic to cut off all access to its top-tier models (Fable 5 and Mythos 5) from outside the United States, and users in non-U.S. countries, including India, lost access. This is neither a failure of self-reliance nor a deficiency of institutional separation; it is an independent, third failure mode, in which dependence on a foreign AI service is severed by a unilateral policy decision on the part of the providing country.
More importantly, this layer connects directly to this paper’s discussion, in Chapter Six, of dependence on Palantir. Palantir’s AI Platform (AIP) is not limited to in-house models; it adopts a “multi-model” design that selectively incorporates multiple external foundation models, including those of Anthropic, OpenAI, and Google.[36] Indeed, in February 2026, after contract negotiations between Anthropic and the U.S. Department of Defense broke down, the Department designated the company a “supply chain risk,” demanding that every defense-related contractor doing business with Anthropic cease that business (it should be noted that this measure was halted in March 2026 by a federal court, which found it “retaliatory and unconstitutional”; this rests on a different legal basis than the export-control suspension of Fable 5 and Mythos 5 for foreign users, and falls outside the scope of that ruling).[37]
What this sequence of events reveals is a “nested” structure of dependence: however carefully Japan manages its relationship with a single company, Palantir, that management cannot reach the geopolitical position of the foundation-model providers that lie beneath it.
Note: The following is an inference and requires further verification.
It cannot be confirmed from public information which foundation models Japan’s Palantir-related systems (at SOMPO, Fujitsu, and elsewhere) currently use. However, given that AIP is a multi-model design, the structural possibility always exists that, whichever foundation model is eventually adopted, access could change — independent of Japan’s own intentions — according to a policy decision made in that provider’s home country (chiefly the United States). Neither of the two types this paper presents (the isolation type and the institutional-separation type) directly addresses this third risk.
Figure 4. Two types of autonomy and dependence — three positions, and a proposal for Japan (Source: prepared by the author)
Chapter Ten — Policy Implications
Recommendation ① Institutionalizing the Operation of the Ideological Infiltration Index
This paper recommends institutionalizing the four-stage index it presents as a benchmark for continuous early-warning monitoring, to be jointly operated by the Ministry of Foreign Affairs, the Ministry of Defense, the Ministry of Economy, Trade and Industry, and other relevant bodies. This would make it possible to grasp the stage of progression in advance — rather than “noticing only after dependence has deepened” — and to secure the time needed to respond.
Recommendation ② Institutional Separation of Information from Authority (Applying the French Model)
Threat-information-sharing agreements (cooperation with bodies such as the FBI, NSA, and CISA) should be subject to a different approval process and a different oversight regime than contracts for building or operating core systems. Establishing a clear institutional standard that distinguishes “gathering information” from “ceding authority” is the most realistic policy for Japan.
Recommendation ③ Expanding the Concept Underlying the Economic Security Promotion Act
The concept of “specified critical technologies” should be expanded to cover military-AI and data-integration platforms, and Japan should consider institutionalizing a “geopolitical autonomy impact assessment” applicable to the entry of foreign private companies.
Recommendation ④ Institutionalizing Ideological Literacy in Diplomacy and Defense
Assessing the risk of technological dependence cannot rely on legal and technical standards alone. The ability to read the ideological motives of a counterpart must be built into the training of diplomats, defense officials, and politicians. Japan’s political culture has a low affinity with aggressive ideological language such as the Dark Enlightenment, and consequently has a structural tendency to underestimate the intentions of an actor driven by such ideological motives. Recognizing this “cultural blind spot,” and establishing ideological literacy as a formal pillar of security education, is an urgent priority.
Recommendation ⑤ Making Visible the Foundation-Model Dependence Within AI Systems
For any AI system the government or a public institution introduces — including multi-model platforms such as Palantir AIP — the providing company and providing country of the foundation model actually used as its base should be identified, and the risk that access to that model could be cut off by a geopolitical decision on the part of the providing country (a change in export controls, defense-procurement regulations, etc.) should be explicitly written into the evaluation criteria used in selecting and continuing to operate such systems. Managing the relationship with the vendor (Palantir, etc.) alone cannot capture the dependence risk that lies in the foundation-model layer beneath it (see section 9-4).
Conclusion
This paper has empirically traced, using a four-stage diagnostic framework, which nations an anti-democratic governing ideology — systematized in U.S. Silicon Valley — selects as it crosses national borders, and in what order it spreads within them.
Under this index, Argentina is, at present, the only case in which nearly all four stages — from ideological contact, through infrastructural substitution and policy codification, to continuing policy consultation with the government — have progressed, making it the closest thing to a complete realization of the process (part of its connection to continuing policy consultation rests on the Isaac Accords as an indicator, which is itself a working hypothesis). Japan shows only the early stage of ideological contact and a limited instance of infrastructural substitution at the level of a single company; it cannot be said that a crisis is under way. This does not mean, however, that Japan is “safe” — it means that Japan remains at a stage where early warning still leaves room for a range of responses.
The policy this paper ultimately proposes is neither dependence nor isolation. Between the two failure modes shown by India — leaving vulnerability unaddressed through isolation — and by Argentina — the loss of options through dependence — lies a third path, demonstrated by France: the institutional separation of information from authority. Japan should consciously design itself along this path. Gather information, without being controlled. This is the most realistic strategic policy this paper proposes for Japan’s cyber security.
Finally, a note on the limits of the word “export,” as used in this paper’s title. Many of the facts this paper has presented come down to a passive structure, in which a structural condition already present in the receiving nation — distrust of the state, demand for efficiency — precedes any active effort by Thiel’s side, and ideology is then “grafted” onto it. “Export,” accordingly, is not a term that demonstrates that a particular actor has deliberately targeted other nations; it is recorded here as a metaphor referring to an observed pattern of infiltration.
Afterword: The Need for Continued Monitoring
The Ideological Infiltration Index this paper presents is not a fixed verdict, but an analytical tool premised on continued observation. Whether Stages ② and ③ progress within Japan — in particular, which entity will be responsible for building the National Intelligence Agency’s systems, and whether contact between the Takaichi government and Thiel’s circle develops from a one-off courtesy visit into continuing policy consultation — remains undetermined as of this writing. Continuing to monitor public information going forward is the concrete practice of early warning.
In particular, given that Fujitsu is separately deepening its relationship with Palantir in both the defense and commercial domains — through its commissioned research for the Acquisition, Technology & Logistics Agency (defense multi-agent AI research) and its Palantir AIP license agreement (with no primary source showing any technical connection between the two) — the possibility that Fujitsu could, in the future, become involved in building the National Intelligence Agency’s systems is one concrete object this paper believes warrants continued attention. However, no primary source confirms this at the time of writing, and no firm conclusion can be drawn.
References
[1] Thiel, P. (2009). The Education of a Libertarian. Cato Unbound. https://www.cato-unbound.org/2009/04/13/peter-thiel/education-libertarian/
[2] Wikipedia, “Dark Enlightenment.” https://en.wikipedia.org/wiki/Dark_Enlightenment / Miranda, A. (2025). “Understanding Neoreaction: A Focus on Curtis Yarvin.” illiberalism.org.
[3] Wikipedia, “Curtis Yarvin.” https://en.wikipedia.org/wiki/Curtis_Yarvin
[4] The New Republic (2024). “Where J.D. Vance Gets His Weird, Terrifying Techno-Authoritarian Ideas.” https://newrepublic.com/article/183971/jd-vance-weird-terrifying-techno-authoritarian-ideas / Newsweek (January 18, 2025). “Who is Curtis Yarvin? Conservative linked to JD Vance wants ‘monarchy’.”
[5] The Guardian (December 21, 2024). Wilson, Jason. “He’s anti-democracy and pro-Trump: the obscure ‘dark enlightenment’ blogger influencing the next US administration.” / Political Research Associates (June 13, 2025). “Tech Capitalism and the Neoreactionary Movement Behind DOGE.” https://politicalresearch.org/2025/06/13/tech-capitalism-and-neoreactionary-movement-behind-doge / The Nation (December 18, 2024). “A Guide to Some of the Key Movers and Influencers in Trump’s Orbit.”
[6] New Zealand Herald (January 25, 2017). “Controversial billionaire Peter Thiel made a Kiwi after two-week holiday.” https://www.nzherald.co.nz/business/controversial-billionaire-peter-thiel-made-a-kiwi-after-two-week-holiday/H7ZMW57ILS4VX3TTY3LBRGFHRA/ / RNZ (June 30, 2017). “Govt digs in over Thiel citizenship decision.” https://www.rnz.co.nz/news/political/334161/govt-digs-in-over-thiel-citizenship-decision
[7] Wikipedia, “Peter Thiel.” https://en.wikipedia.org/wiki/Peter_Thiel (regarding Pronomos Capital, Praxis, and the Greenland sovereignty dispute over the integration of Palantir in Denmark)
[8] WIRED (June 20, 2026). “Leak Exposes Members of Peter Thiel’s Secretive ‘Dialog’ Society.” https://www.wired.com/story/leak-exposes-members-of-peter-thiels-secretive-dialog-society/
[9] Ibid.
[10] MercoPress (April 23, 2026). “Tech magnate Peter Thiel meets Milei after a week of private contacts with the Argentine government.” https://en.mercopress.com/2026/04/23/tech-magnate-peter-thiel-meets-milei-after-a-week-of-private-contacts-with-the-argentine-government / Buenos Aires Times (April 24, 2026). “Milei discusses taxes, culture war, politics in ‘wonderful’ meeting with Thiel.” / Buenos Aires Herald (April 23, 2026). “Peter Thiel meets with Javier Milei in Buenos Aires.”
[11] Reuters (November 2023). Milei election victory and Musk reaction coverage. / ENACOM Resolución 1291/2020 (Starlink’s operating license in Argentina, effective December 2021).
[12] Amnistía Internacional Argentina (May 2026). “Caso Palantir en Argentina.” / DNU 941/2025 (issued December 31, 2025, amending the National Intelligence Law).
[13] Informáticos.ar (April 23, 2026). “Palantir: una amenaza para la democracia.” https://informaticos.ar/palantir-una-amenaza-para-la-democracia/
[14] Radio Gráfica (May 28, 2026). “Milei lanza en Argentina un sistema de inteligencia artificial bajo la sombra de Palantir.”
[15] DLA Piper (2026). The Isaac Accords: Argentina’s New Strategic Framework with Israel. / The Jerusalem Post (2026). / AP News (2025).
[16] Prime Minister’s Office of Japan. Reporting on Thiel’s visit to Japan and courtesy call (March 5, 2026).
[17] SOMPO Holdings, official announcement (November 18, 2019). https://www.sompo-hd.com/~/media/hd/files/news/2019/20191118_1.pdf / Nikkei Cross Tech (November 18, 2019). “Tackling the World’s Challenges: SOMPO Holdings to Establish New Company with U.S.-Based Palantir for Data Analytics.”
[18] Nishimura & Asahi (2020). “SOMPO Holdings, Inc.: Investment in Palantir Technologies Inc.” / CX Research Institute (2026). “What Is Palantir? Explaining Its Business, Stock Price, and Entry into Japan.”
[19] Nikkei (August 19, 2025). “Fujitsu Signs AI Contract with U.S.-Based Palantir.” / Fujitsu, official announcements (December 5, 2023; August 19, 2025). On the conclusion of the global partnership and license agreement / Acquisition, Technology & Logistics Agency, Defense Innovation Science and Technology Institute, announcement regarding commissioned research with Fujitsu (defense multi-agent AI research).
[20] Sumitomo Corporation, official announcement (March 19, 2026). On the conclusion of a license agreement with Palantir Technologies Japan through Ascend SC.
[21] Japanese-language Wikipedia, “Palantir Technologies.” https://ja.wikipedia.org/wiki/パランティア・テクノロジーズ / JBpress (March 2026). “The Reality of the Data-Analytics Company Palantir.” / Gendai Business (March 2026). “Peter Thiel, to Whom Prime Minister Takaichi Has Suddenly Grown Close” (note: both carry essentially identical claims, and no independent confirmation from an official Ministry of Defense statement or a major national newspaper had been obtained at the time of this paper’s writing).
[22] Prime Minister’s Office of Japan (May 27, 2026). Press conference on the enactment of the National Intelligence Council Establishment Act. https://www.kantei.go.jp/jp/105/statement/2026/0527kaiken.html / Cabinet Secretariat (2026). Outline of the National Intelligence Council Establishment Bill. https://www.cas.go.jp/jp/houan/260313/gaiyou.pdf / House of Councillors, bill information. https://www.sangiin.go.jp/japanese/joho1/kousei/gian/221/meisai/m221080221024.htm
[23] sbbit.jp (2026). “The Soon-to-Be-Established ‘National Intelligence Agency’: Why It Is Neither a Japanese ‘CIA’ Nor an ‘FBI’.”
[24] Japanese-language Wikipedia, “Ankoku Keimo” (Dark Enlightenment). https://ja.wikipedia.org/wiki/暗黒啓蒙
[25] Satoshi Kizawa (2019). Nick Land and Neoreaction: The Dark Ideas Covering the Contemporary World. Seikaisha Shinsho.
[26] Ministry of Internal Affairs and Communications (2021). Press release regarding Fuji Media Holdings’ violation of the foreign-ownership restriction / National Diet Library research material. The foreign-ownership violation by Tohokushinsha and the revocation of its satellite-broadcasting authorization (2021).
[27] The Record from Recorded Future News (April 6, 2022). “Suspected China-backed hackers target 7 Indian electricity grid centers.” / The Diplomat (April 8, 2022). “Chinese Hackers Reportedly Target India’s Power Grid.”
[28] eventussecurity.com (2026). “Top 15 Recent Cyber Attacks in India 2026.”
[29] SGDSN (2026). “National Cybersecurity Strategy 2026–2030.” https://www.sgdsn.gouv.fr/files/files/National%20cybersecurity%20strategy_ENG.pdf
[30] Chambers and Partners (2026). “Cybersecurity 2026 – France.” / Legiscope (2026). “Certification SecNumCloud ANSSI: Guide Complet 2026.”
[31] CNIL (January 29, 2026). “Data breach: FRANCE TRAVAIL fined €5 million.” https://www.cnil.fr/en/data-breach-5million-fine-france-travail / Infosecurity Magazine (January 29, 2026). “France Fines National Employment Agency €5m Over 2024 Data Breach.”
[32] Asterès / Cigref (April 25, 2025). “Technological Dependence on American Software and Cloud Services: An Assessment of the Economic Consequences in Europe.” https://www.cigref.fr/wp/wp-content/uploads/2025/05/TECHNOLOGICAL-DEPENDENCE-ON-AMERICAN-SOFTWARE-AND-CLOUD-SERVICES-AN-ASSESSMENT-OF-THE-ECONOMIC-CONSEQUENCES.pdf (study commissioned by Cigref and Numeum)
[33] Ministry of Defense of Japan (January 16, 2026). “Defense Minister Koizumi’s Visit to the United States.” https://www.mod.go.jp/j/approach/anpo/2026/0112a_usa-j.html / Kyodo News (January 17, 2026). “Defense Minister Koizumi Inspects Cutting-Edge Technology.” / Shinjiro Koizumi, post on X (formerly Twitter) (January 17, 2026).
[34] Anthropic, official statement (June 12, 2026). “Statement on the US government directive to suspend access to Fable 5 and Mythos 5.” https://www.anthropic.com/news/fable-mythos-access / CNBC (June 12, 2026). “Anthropic disables access to Fable 5 and Mythos 5 to comply with government directive.” / Al Jazeera (June 14, 2026). “US asks Anthropic to block global access to top AI models: Why it matters” (citing Reuters reporting).
[35] The Next Web (June 2026). “Anthropic’s model shutdown just handed India’s sovereign AI movement its strongest argument yet.”
[36] Palantir Technologies. “AIP Model Catalog.” https://www.palantir.com/docs/foundry/model-catalog/overview / Palantir Technologies, AWS, and Anthropic, joint announcement (November 2024). “Anthropic and Palantir Partner to Bring Claude AI Models to AWS for U.S. Government Intelligence and Defense Operations.”
[37] CBS News (February 27, 2026). “Hegseth declares Anthropic a supply chain risk, restricting military contractors from doing business with AI giant.” / CNN Business (March 26, 2026). “Judge blocks Pentagon’s effort to ‘punish’ Anthropic by labeling it a supply chain risk.” / Mayer Brown (March 2026). “Anthropic Supply Chain Risk Designation Takes Effect — Latest Developments and Next Steps for Government Contractors.”
Glossary
Libertarianism
A political philosophy that places the highest priority on individual freedom, private property rights, and the market economy, and holds that state intervention and regulation should be minimized. The Dark Enlightenment discussed in this paper is positioned as a more extreme offshoot, in which some proponents of libertarianism moved beyond the claim that “the state should be reduced” to the further conclusion that “democracy itself is an obstacle to freedom.” Peter Thiel’s 2009 essay is a document that symbolizes this branching point.
The Dark Enlightenment (Neoreaction, NRx)
An anti-democratic, anti-egalitarian, anti-progressive intellectual movement that emerged online in the mid-2000s. Systematized by Curtis Yarvin and extended by Nick Land. It rejects democracy as an “inefficient institution” and seeks to replace it with an authoritarian governance structure resembling corporate governance.
The Ideological Infiltration Index
A diagnostic framework proposed by this paper that decomposes the process by which an ideology crosses national borders and infiltrates a nation into four stages: ideological contact, infrastructural substitution, policy codification, and direct political engagement.
Seasteading / Charter Cities
A concept for building an autonomous community, governed by its own rules, in a place beyond the reach of any existing state’s sovereignty — whether on the high seas or in a land area holding special legal status. Thiel was an early backer of the Seasteading Institute, and has since continued to invest in charter cities (such as Próspera in Honduras) through vehicles including Pronomos Capital. The New Zealand case discussed in Chapter Three of this paper is positioned as a practical precedent that occurred within the framework of an existing state, prior to the realization of this concept.
Dialog
A private, invitation-only organization co-founded by Peter Thiel in 2006. In June 2026, WIRED verified and reported on its substance following the leak of its internal records.[8]
The Techno-Sovereign Actor
A group of private actors that is neither a state nor an ordinary multinational corporation, but which holds cross-cutting control over multiple security-related domains — military AI, space, the integration of state data, telecommunications infrastructure, and strategic resources.
Two Types of Autonomy (Isolation Type / Institutional-Separation Type)
Autonomy aimed at avoiding dependence on a techno-sovereign actor takes two distinct forms: an “isolation type” (exemplified by India), which isolates itself from international threat-information sharing and leaves operational vulnerability unaddressed, and an “institutional-separation type” (exemplified by France), which maintains international cooperation on threat-information sharing while institutionally separating authority over core data.
Project 2025
A policy document led by the U.S. Heritage Foundation (released April 2023). It is said to bear an ideological parallel to a neoreactionary vision, packaged in the language of a think tank. Its central drafter was Russell Vought.
Russell Vought
A figure who served as Director of the Office of Management and Budget in the first Trump administration. The central drafter of Project 2025, who, through the think tank he leads, the Center for Renewing America (CRA), produced draft executive orders ready for signature immediately upon the administration taking office. His posture of confrontation with a “corrupt bureaucracy” is said to be substantively consistent with what Yarvin calls the dismantling of “the Cathedral.”
© 2026 Institute for Japan Institute of Crisis Management. All rights reserved.
Author: Miho Funayama
Miho Funayama is a strategic analyst specializing in international standardization, technology intelligence, intellectual property analysis, and geopolitical and crisis risk management. She holds a degree in International Politics from Sophia University and completed graduate studies in International Political Economy, Philosophy, and Psychology at Aoyama Gakuin University. She previously led patent research and strategic analysis on international standards and emerging technologies at Canon Inc., where she served as Deputy International Secretary of ISO/IEC JTC 1/SC 28 and received the ITSCJ Award for Contribution to International Standardization three times. She is currently a Director and Chief Researcher at the Japan Institute for Crisis Management Research, a Fellow at the Institute of Middle East–Asia Information Strategy, and a Professional Associate Member of the Foreign Correspondents’ Club of Japan.
Her research focuses on the intersection of cybersecurity, geopolitical risk, and information warfare, examining decision-making structures and psychological operations through an interdisciplinary lens spanning international politics, philosophy, and psychology.