The Second Operating System of Power: How Palantir Connects and Divides Germany’s Security Agencies – Part 1

by Michael Hollister
Published at apolut media on September 08, 2026

5.134 words * 27 minutes readingtime

The First Operating System of Power read here:
BlackRock, Aladdin, and the Invisible Infrastructure

The same software that 4 German state police forces consider indispensable was rejected by the German Armed Forces in April of this year as an unacceptable security risk. While Bavarian police use the U.S. company Palantir’s analytics platform to search around 39 million personal-data records, the Bundeswehr’s Chief of Cyber and Information Domain Service said it was “inconceivable” to allow company employees access to the national data holdings. The Bundeswehr rejected Palantir’s operating model. 2 agencies of the same state, the same company, 2 opposite judgments – and no bridge between them, but an open rupture.

The contradiction comes in a year when Palantir is bludgeoning every doubt about its business model with numbers. In the second quarter of 2026, the company reported a 93 percent increase in revenue to 1.935 billion dollars; revenue from U.S. government customers rose by 90 percent, while business with U.S. corporate customers grew by almost 150 percent. In August alone, the stock climbed by more than half, and market capitalization temporarily exceeded 440 billion dollars – only a few percentage points below the all-time high reached in November 2025. The market has long since stopped valuing Palantir as a software company; it values Palantir as infrastructure that will be difficult to bypass for years.

That is precisely the core of this analysis. When the Federal Constitutional Court ruled on the use of Palantir’s software in Hesse on February 16, 2023, it used a remarkable term in the English version of its judgment: Gotham, the company’s core product, was an “operating system.” The term is not journalistic hyperbole, but the description used by Germany’s highest court. It captures what Palantir has become for the security state – just as Aladdin, the analytics system operated by the asset manager BlackRock, became the operating system of the financial markets examined in an earlier analysis. Palantir is the second of these private operating systems of power, and it sits deeper inside the state than most debates care to acknowledge.

The Question Beneath the Dispute

Public controversy over Palantir is almost always framed in moral terms: Is the software dangerous? Is its co-founder Peter Thiel’s support for Trump a problem? Should a U.S. corporation be allowed to process German police data? These are legitimate questions, but they obscure the structurally more interesting one. They reduce a conflict over power to a question of personality.

The deeper question is this: How can a private operating system simultaneously serve as the backbone and the security risk of the same state? 4 German states build their investigative work around it, while the federal military and domestic intelligence service reject it on principle. Behind that lies the real question of power: Who controls the data infrastructure on which the state performs its most sensitive functions when that infrastructure belongs to a publicly traded company based in the United States? This first part maps the anatomy and the rupture – what the software does, who uses it, who refuses it, and why Karlsruhe has imposed limits on it. The second part will examine who profits from this operating system and who owns it.

Anatomy of an Operating System

Palantir’s software does something technically unremarkable with far-reaching consequences: It brings together data holdings that were previously separate. Police case files, communications data, registry information, contact details, movement and travel information – and, in Hesse, vast quantities of cell-site data – are stored in different systems and formats, then linked in a single environment and rendered visible as a network of people, places, and events. Scattered individual data points become a searchable web of relationships that can be displayed on maps and in network diagrams. The manufacturer promises that this produces entirely new investigative leads; critics call it a super-database that should never have existed in this form.

Frankfurt police have repeatedly demonstrated to journalists what this looks like in practice. During a traffic stop, officers discover that a so-called Gefährder is sitting in the car – an Islamist believed capable of committing the most serious acts of violence. That threshold permits the Hessian software to be activated. It links the available data, identifies the stopped vehicle, produces the names of all offenders living in the surrounding area, and shows in real time where an attached tracking device is moving. Research that once required days of manual work across multiple applications and databases happens within seconds through a single interface. This gain in speed is precisely the selling point advanced by the interior ministries, and it is real.

The downside is equally real. Anyone who reports a crime, becomes the victim of one, or was simply in the wrong place at the wrong time can enter this web. According to the data-protection authority, the Bavarian platform searches around 39 million personal-data records, including data on witnesses, victims, expert witnesses, and people who merely called the police emergency number. The system has no presumption of innocence; it recognizes only connections. The software is also used in some cases for predictive policing, calculating the probability of crime so that authorities can respond with a greater police presence. Another criticism concerns traceability: The source code and algorithms are not public, so the analytical results cannot be independently verified – a black box that generates suspicion without revealing its calculations.

The term operating system is more precise than it initially sounds. An operating system does not manage content itself, but the environment in which content becomes usable; it determines which programs may communicate with one another and in what language. That is precisely what Palantir’s platform does for government data. The company divides its products into 2 lines: Gotham targets security and defense agencies, while Foundry serves companies and public administrations. Both share what Palantir calls an ontology – a digital model that represents people, places, vehicles, events, and the relationships among them within a unified structure. Only this ontology turns a data silo into a searchable representation of reality. In Germany, Palantir operates not merely as a U.S. company; since 2015 it has maintained its own office and a German subsidiary that handles sales and adaptation for domestic authorities.

This operating system is rarely run by civil servants alone. Palantir does not deliver its products as finished, off-the-shelf software; it adapts them to the structures of each agency, which means company employees work inside the security authorities. They import data from police source systems, keep the platform running, and continue developing it. This interlocking of a private provider and a sovereign government function is not a peripheral detail. It is the point at which the state will later split against itself, because the same support arrangement that the German states take for granted becomes a reason for exclusion by the military and intelligence services. Whoever maintains the operating system potentially sees what runs on it – and once that thought has been voiced, it cannot be taken back.

The Map: Who Uses It and Who Refuses

Germany’s use of Palantir began in Hesse. In 2017, under a Christian Democratic-Green coalition government, the state purchased Gotham and operated it under the name Hessendata – becoming the first German state to do so and remaining a pioneer for years. A year later, the state parliament created a specific legal basis intended to account for the right to informational self-determination; according to the Constitutional Court’s findings, that authority was used thousands of times every year. Bavaria followed with a strategically more significant arrangement: The state concluded a framework agreement with Palantir that makes it easier for other states to adopt the same product. The cross-case Research and Analysis Platform, known as VeRA, has been in use since 2024, including in several dozen cases involving property and financial offenses. North Rhine-Westphalia has used Palantir technology for several years, and Berlin recently authorized its use by law. The German states have nevertheless taken sharply different positions on Palantir. Costs in the participating states add up to tens of millions of euros. Behind this sober inventory lies an unresolved question of principle that is openly voiced even within the states using the system: Is Gotham a technological bridge until a European solution is ready, or the beginning of a permanent software-based surveillance architecture from which there is no way back? Baden-Württemberg’s 5-year contract without an exit clause suggests that this is no longer an abstract question, but one already being answered by facts.

Baden-Württemberg is the latest and, in some respects, the most revealing addition. In November 2025, the state parliament cleared the legal path through an amendment to the Police Act – over the protests of the Green coalition partner, which criticized the purchase but did not prevent it. The contract had long since been signed: around 25 million euros over 5 years, with no exit clause. Interior Minister Thomas Strobl said the data would be stored in secure data centers in Germany, exclusively under police control, and that access from abroad was impossible; the Parliamentary Oversight Committee would monitor work with the software and receive regular IT-security reports. The police are scheduled to begin using the software in practice in the second quarter of 2026 – a commitment lasting years, with no contractual escape route.

The persistence with which the participating states cling to the software is striking. After experts issued a public appeal warning of the platform’s risks, the interior ministers of the affected states indicated that they would not change course. Those ministers are fully aware that dependence on a U.S. product is delicate; in principle, they reportedly agree that German software should be used wherever possible in the future. Yet this declaration changes nothing in the present – no German alternative is currently ready for operational use, and Bavaria’s framework agreement has effectively made Gotham a quasi-standard that additional states can adopt without conducting their own procurement process. The commitment to sovereignty and the practice of procurement are moving in opposite directions.

Opposite them stand the holdouts, and they are more numerous than the expansion suggests. Lower Saxony, Bremen, and Hamburg reject its use or rely on alternative solutions; Lower Saxony Interior Minister Daniela Behrens explicitly ruled out Palantir software in her state. Schleswig-Holstein planned a separate tender for analytics software and emphasized that it was not working with Palantir. Skepticism is also growing beyond Germany: A Swiss review concluded that the software posed serious risks. The result is a fragmented map: no nationwide standard, but a patchwork of states reaching opposite conclusions about the same private operating system – with Bavaria’s framework agreement waiting as a gateway should the hesitant states reverse course. The company’s share price has more than doubled within a year, partly because an increasing number of European authorities are using the software or considering it.

The Divided State

The deepest fault line runs not between the German states, but within the federal government – and leads directly to the central question of sovereignty. In April 2026, Vice Admiral Thomas Daum, the Bundeswehr’s Chief of Cyber and Information Domain Service, made clear that Palantir would not participate in building the planned Bundeswehr cloud. His reasoning targeted not the technology, whose capabilities he expressly valued, but the operating model. At NATO, company employees operated the software themselves; that model was out of the question for the Bundeswehr. However interested the armed forces might be in the system’s capabilities, there was no justification for giving employees of a private provider access to national data holdings unless the military retained complete control.

The alternatives under consideration have European names: a Stuttgart provider certified by the Federal Office for Information Security, a Berlin start-up, and the French company ChapsVision. For German defense and security technology, this is more than a procurement question; it is an attempt to build a sovereign layer of critical software that is not subject to a foreign legal order. The path is considered longer and more expensive, but safer.

A few weeks later, the Federal Office for the Protection of the Constitution made a decision pointing in the same direction. The domestic intelligence service awarded the contract for new software to analyze large quantities of data not to Palantir, but to the same French company, ChapsVision, and its ArgonOS system. The federal government’s response to parliament documents the procurement decision. Here too, the stated motive was digital sovereignty: Control over sensitive data holdings was to remain within the national or European sphere, and technological dependence on U.S. service providers was to be reduced. The domestic intelligence service, whose core business involves handling the most sensitive personal data, therefore deliberately rejected the market leader. The same software that 4 state police forces use as indispensable is considered an unacceptable risk by the federal military and domestic intelligence service. The rupture does not run between advocates and opponents of surveillance, but straight through the security apparatus itself – separating the level with the most to lose from the level quickest to adopt the software.

The contrast with the international level sharpens the finding. On March 25, 2025, NATO tasked its own communications agency with acquiring a customized Palantir product, the Maven Smart System, for Allied Command; the platform originated in a U.S. image-analysis program. According to the Bundeswehr’s service chief, that system too is operated by company employees. At the beginning of 2026, the United Kingdom extended its contract for 36 months at a cost of 240 million pounds, while the United States already uses the technology extensively. Allied armed forces rely on the operating system, while the German military expressly rejects it – not because it doubts the system’s capabilities, but because it is concerned about control over its own data, a concern others apparently do not share with the same intensity.

The contradiction calls for an explanation, and it lies in the nature of the data involved. A state police force works with case data already in its possession; the risk of outside access carries a different weight there than it does for an intelligence service, whose methods of gathering information are among the state’s most closely guarded assets, or for a military, whose operational picture can determine life and death in a conflict. What the states accept as a manageable support relationship represents a loss of control that the military and intelligence service cannot afford. The same mechanism – company personnel accessing national data holdings – produces a different judgment depending on the sensitivity of the data.

At the political level, the rupture is reflected in the coalition dispute and follows the party line within the CDU/CSU-SPD government. Interior Minister Alexander Dobrindt of the CSU considers Palantir useful for the Federal Police and Federal Criminal Police Office, citing the threat from terrorist and criminal structures that requires AI-assisted tools; he said he had “no qualms” about the software merely because of its provider. Justice Minister Stefanie Hubig of the SPD publicly disagreed: She did not see Palantir coming to the federal agencies; Germany must not enter into dangerous dependencies, and tools for automated data analysis required particular caution because they involved sensitive data. CDU Digital Minister Karsten Wildberger, by contrast, expressed openness to the software, arguing in part that it could help protect democracy, while the SPD parliamentary group’s digital-policy spokesperson firmly rejected nationwide deployment. The coalition remains divided over Palantir’s use at the federal level.

The government cannot agree on its own operating system – ministry against ministry, party against party, over the same question that divides the states. The dispute has been intensified by the broader foreign-policy climate: As conflicts between European governments and the U.S. administration have escalated over the course of the year, dependence on a U.S. provider carries more weight than it did only months ago. What began in 2024 as a technical procurement question has become a matter of strategic vulnerability in 2026.

Karlsruhe and the Open Flank

Above all of this stands a judgment the government knows and that obstructs its plans. On February 16, 2023, the First Senate of the Federal Constitutional Court ruled that Hesse’s legal basis for automated data analysis was unconstitutional and declared the largely identical Hamburg provision void. The provisions violated the general right of personality in its manifestation as the right to informational self-determination because they lacked an adequate threshold for interference. The mere combination of previously stored data into a comprehensive picture constituted a separate and serious interference with fundamental rights. Data analysis was permissible only to protect particularly important legal interests and in the presence of a sufficiently specific danger supported by factual evidence. Karlsruhe did not prohibit the technology, but it erected guardrails – Hesse was given until the end of September 2023 to revise its legal basis.

Whether the states comply with those guardrails is disputed. In a resolution dated September 17, 2025, the Conference of Independent Data Protection Supervisory Authorities of the Federal and State Governments concluded that the police laws in Bavaria, Hesse, and North Rhine-Westphalia still failed to meet the requirements. It also called for the state’s digital sovereignty and pointed to the federal-state P20 program as an option for an open-source alternative. The Society for Civil Rights has filed constitutional complaints against all 3 state laws. In Bavaria, 8 complainants are suing, including a criminal defense lawyer, activists, and members of a football fan legal-aid group; the Chaos Computer Club supports them. A litigation coordinator for the organization summarized the criticism with the phrase “Bavaria is not Gotham City.” In Hesse, the organization had already succeeded with an earlier complaint and forced the legislature to tighten the law.

And the federal government? It is responding to the Karlsruhe problem with legislation. 6 weeks after Hubig publicly rejected Palantir, the cabinet adopted the Draft Act to Strengthen Digital Investigative Powers in Police Work on April 29, 2026, as part of a larger security package. It amends the Federal Criminal Police Office Act, the Federal Police Act, and the Asylum Act, authorizing automated data analysis, biometric comparisons with publicly accessible online images, and the testing and training of IT products, expressly including artificial intelligence and personal data. Its explanatory memorandum cites the same 2023 Constitutional Court judgment paragraph by paragraph. The Bundesrat raised no objections on June 12, 2026; the Bundestag referred the bill to the lead Interior Committee after its first reading on July 08, 2026.

The reference to P20 is more than a footnote. P20 is the joint federal-state project to place police IT on a shared information architecture – a state-run data environment on which analytical services can operate. The data-protection authorities see in it an opportunity to create an open-source, controllable alternative to the private operating system. The federal government, by contrast, has adopted a modular, hybrid approach to P20 that can expressly include commercially available products alongside in-house development. Whether P20 will produce emancipation from the private provider or merely wrap that provider in a state-owned shell therefore remains open – and will be addressed in Part 2.

The legal situation also remains in motion. In addition to the proceedings in Bavaria and Hesse, a constitutional complaint against the North Rhine-Westphalian provision remains pending. In Baden-Württemberg, meanwhile, the Police Act that first permits use of the software was temporarily stalled even as the state was already paying the company without using the product. The legal dispute over the operating system is not over; it has only just begun.

What the draft does not specify is decisive: a provider. The infrastructure is being created in a product- and technology-neutral form – Palantir-capable, but not Palantir-bound. Hubig’s rejection thus remains intact as a statement while the substantive door stays open. The Society for Civil Rights considers the drafts largely unconstitutional and calls, among other things, for stricter thresholds and the exclusion of private providers based outside the European Union; the German Bar Association has also expressed doubts about their constitutionality, while the Green parliamentary group opposed a blanket expansion of digital powers. Who ultimately fills that infrastructure may be decided by a future government, a different minister, or a different political climate. The legislature is building the house; who moves in remains open.

The Case for Palantir

A finding that hears only 1 side is no finding at all. There are substantial arguments in favor of the software, and they must be stated. The most important concerns the most common suspicion – the covert transfer of data to the United States. Commissioned by the Bavarian State Criminal Police Office, the Fraunhofer Institute for Secure Information Technology examined the software’s source code and found no evidence of hidden backdoors. That rebuts the sharpened claim that Palantir secretly siphons off police data, although it does not resolve the underlying legal question of access through the front door.

The benefits are also asserted by an interested party, but not for that reason an unreliable one. Palantir CEO Alex Karp said in a newspaper interview that his company’s software had helped prevent several major terrorist attacks in Germany. That claim can neither be verified nor disproved from the outside; it remains a statement by the manufacturer and must be identified as such. The interior ministries point to the speed gained during dangerous situations, and that point is undisputed: Anyone who can connect scattered data in seconds rather than days gains time in an acute crisis, and that time can save lives. In the manufacturer’s advertising and the language used by some authorities, the platform is therefore described as a quantum leap for police work – an exaggeration that overstates the benefit just as some critics overstate the danger.

Intellectual honesty requires that the benefit not be minimized merely because the provider is controversial. A tool that locates a potential threat in Frankfurt does not become useless because its manufacturer is owned by a billionaire whose political views are polarizing. Conversely, it does not become harmless merely because it works. Both statements stand side by side, and their simultaneity is precisely the difficulty that cannot be resolved into a simple position for or against.

Computer scientist Felix Freiling of Friedrich-Alexander University Erlangen-Nuremberg has captured the ambivalence precisely: The more authorities know, the more they can prevent and solve, especially after the fact – but society is damaged when everyone becomes transparent, and in the long run a fear of surveillance takes hold. A free society, Freiling argues, needs data protection and privacy in the digital sphere as well. The real balancing exercise lies between these 2 truths. Opposite them stands the other side: deliberately opaque software, dependence lasting years, and a contract with no exit clause. This analysis does not render the judgment; it belongs to the reader, who is shown both columns side by side.

Sovereignty and the Long Arm of the CLOUD Act

Interior Minister Strobl’s assurance that the data is securely stored in German data centers answers the technical question while missing the structural one. The location of the servers alone does not determine who may access the data. The U.S. CLOUD Act requires U.S. companies to provide authorities in their home country access to stored data regardless of where in the world that data is located. A company subject to U.S. law can face a dilemma in which European and U.S. law contradict each other: If it complies with a U.S. order to disclose European data, it violates European law; if it refuses, it violates U.S. law. This is not a hypothetical scenario. The U.S. Supreme Court confronted precisely this question in litigation involving a U.S. company whose data was stored on European servers; the industry association Bitkom and dozens of European trade associations warned that direct access by U.S. authorities to European data was incompatible with European data-protection law. The CLOUD Act did not resolve this tension; it codified it.

For the Bundeswehr and the Federal Office for the Protection of the Constitution, this was precisely a reason for exclusion. The concern is not the hidden backdoor Fraunhofer did not find, but the law’s open front door: the possibility that a U.S. provider could be compelled by an order from its government. The concern has intensified not in the abstract but politically, as European conflicts with the U.S. administration have become more pronounced over the course of the year. In the United States itself, the government has recently expanded Palantir’s role in combining state-held data and spread the technology across agencies – a process raised in Congress as an example of how easily separate data sources can be merged into a comprehensive profile of an individual citizen. What Germany is debating as a data-protection question is already operating elsewhere as government architecture. A state that obtains the operating system for its most sensitive tasks from a provider subject to a foreign legal order surrenders part of its control – no matter how secure the data center may be.

The sovereignty question is therefore no longer merely about data protection, but about power. It asks not only whether data is protected, but who controls the system that processes it and who has access to that system. This question leads directly to the second level explored in the next part of this series: the business behind the operating system, the parliamentary oversight gap, and the question of who actually profits from the so-called European path – choosing ChapsVision instead of Palantir.

Finding

The state has delegated control over part of its most sensitive data infrastructure to a private company – and cannot agree with itself about that company. 4 state police forces build on it; the federal military and domestic intelligence service refuse it; the coalition argues; Karlsruhe draws limits; the data-protection authorities consider those limits breached. This is not a scandal with a clearly identifiable perpetrator, but a structural condition: A private operating system has grown so deeply into the performance of state functions that the state can neither use it consistently nor reject it as a whole.

The capability is documented – the networking of millions of data records, the linking of what was previously unconnected, the speed in an emergency. The intention behind it remains an open question that this analysis neither answers nor seeks to answer. What is remarkable is not that a company sells its software; that is its business. What is remarkable is that a state that has elevated digital sovereignty into a guiding principle partially surrenders that sovereignty precisely where it matters most – in its citizens’ data and in control over the tools that make those data readable. What this analysis shows is the structure, and that structure speaks its own language. While politicians invoke digital sovereignty in Sunday speeches, the Bundeswehr rejects in its weekday business what the state police forces purchase. It is worth following an old grandmother’s rule here: Do not listen to what they say – watch what they do. Anyone who does so sees an operating system of power whose switches are no longer solely in state hands. Who guides that hand and who profits from it is the subject of Part 2.

No advertisers, no sponsors, no political patrons – this publication is funded by its readers. Support independent geopolitical analysis from $5 per month and unlock all member-only content.

You are also welcome to support my work with a one-time donation.

About the Author

Michael Hollister
is a geopolitical analyst and investigative journalist. He served six years in the German military, including peacekeeping deployments in the Balkans (SFOR, KFOR), followed by 14 years in IT security management. His analysis draws on primary sources to examine European militarization, Western intervention policy, and shifting power dynamics across Asia. A particular focus of his work lies in Southeast Asia, where he investigates strategic dependencies, spheres of influence, and security architectures. Hollister combines operational insider perspective with uncompromising systemic critique – beyond opinion journalism. His work appears on his bilingual website (German/English) www.michael-hollister.com and in investigative outlets across the German-speaking world and the Anglosphere.

Sources

Court, Legislative, and Parliamentary Documents

1. Federal Constitutional Court, February 16, 2023 – Judgment of the First Senate, 1 BvR 1547/19 and 1 BvR 2634/20, German version: https://www.bundesverfassungsgericht.de/SharedDocs/Entscheidungen/DE/2023/02/rs20230216_1bvr154719.html
2. Federal Constitutional Court, February 16, 2023 – Same judgment, English version, using the term “operating system”: https://www.bundesverfassungsgericht.de/SharedDocs/Entscheidungen/EN/2023/02/rs20230216_1bvr154719en.html
3. Federal Constitutional Court, February 16, 2023 – Press Release No. 18/2023: https://www.bundesverfassungsgericht.de/SharedDocs/Pressemitteilungen/DE/2023/bvg23-018.html
4. Federal Ministry of the Interior – Legislative Procedure to Strengthen Digital Investigative Powers in Police Work, procedure page: https://www.bmi.bund.de/SharedDocs/gesetzgebungsverfahren/DE/OESI3/ermittlungsbefugnisse-polizeiarbeit.html
5. Federal Ministry of the Interior – Cabinet Version of the Draft Act, PDF: https://www.bmi.bund.de/SharedDocs/gesetzgebungsverfahren/DE/Downloads/kabinettsfassung/OESI3/regE-ermittlungsbefugnisse-polizeiarbeit.pdf
6. German Bundestag, July 08, 2026 – Text Archive on the First Reading, Week 28/2026: https://www.bundestag.de/dokumente/textarchiv/2026/kw28-de-strafprozessordnung-1192534
7. German Bundestag – Today in Parliament, Brief Notice on the Draft Act: https://www.bundestag.de/presse/hib/kurzmeldungen-1187686
8. German Bundestag – Today in Parliament, Brief Notice on the Deliberations: https://www.bundestag.de/presse/hib/kurzmeldungen-1182466
9. German Bundestag, Printed Paper 21/7079 – Federal Government Response Concerning Analytics Software at the Federal Office for the Protection of the Constitution: https://dserver.bundestag.de/btd/21/070/2107079.pdf
10. German Bundestag, Printed Paper 21/4599 – Minor Interpellation Concerning Palantir at the Federal Police and Federal Criminal Police Office: https://dserver.bundestag.de/btd/21/045/2104599.pdf
11. German Bundestag, Printed Paper 21/4923 – Federal Government Response: https://dserver.bundestag.de/btd/21/049/2104923.pdf
12. German Bundestag, Printed Paper 21/5119 – Minor Interpellation on Intelligence-Service Use of Palantir Gotham: https://dserver.bundestag.de/btd/21/051/2105119.pdf
13. German Bundestag, Printed Paper 21/5625 – Federal Government Response: https://dserver.bundestag.de/btd/21/056/2105625.pdf
14. U.S. House Committee on the Judiciary – Hearing on the CLOUD Act and Foreign Access to Data: https://democrats-judiciary.house.gov/committee-activity/hearings/foreign-influence-on-american-s-data-through-the-cloud-act
15. U.S. Senate Committee on Homeland Security and Governmental Affairs, March 19, 2025 – Letter Concerning Government Data Integration, PDF: https://www.hsgac.senate.gov/wp-content/uploads/2025.03.19SRP-Letter-to-DOJ.pdf

Data-Protection Authorities and Civil Society

16. Society for Civil Rights – Press Release on the Constitutional Complaint Against Police Data Analysis in Bavaria: https://freiheitsrechte.org/ueber-die-gff/presse/pressemitteilungen-der-gesellschaft-fur-freiheitsrechte/blackbox-palantir-gff-erhebt-verfassungsbeschwerde-gegen-massenhafte-datenauswertung-durch-polizei-in-bayern
17. State Commissioner for Data Protection and Freedom of Information of North Rhine-Westphalia – Decision Against Automated Data Analysis in Hesse and Hamburg: https://www.ldi.nrw.de/urteilt-gegen-automatisierte-datenanalyse-hessen-und-hamburg
18. Conference of Independent Data Protection Supervisory Authorities of the Federal and State Governments, September 17, 2025 – Resolution on Automated Data Analysis; institutional reference: datenschutzkonferenz-online.de, search term: Entschließung 17. September 2025 automatisierte Datenanalyse.

Press and Trade Media

19. Behörden Spiegel – “Bundeswehr Abandons Palantir”: https://www.behoerden-spiegel.de/2026/05/12/bundeswehr-verzichtet-auf-palantir/
20. Augen geradeaus – “Bundeswehr Cyber Commander Rejects Palantir Software”: https://augengeradeaus.net/2026/04/cyber-kommandeur-der-bundeswehr-lehnt-software-von-palantir-ab/
21. Yahoo Finance – “Bundeswehr Rejects Palantir over Security Concerns”: https://de.finance.yahoo.com/nachrichten/wegen-sicherheitsbedenken-bundeswehr-verzichtet-palantir-151401965.html
22. it-daily – “Federal Office for the Protection of the Constitution Chooses ChapsVision over Palantir”: https://www.it-daily.net/shortnews/palantir-verfassungsschutz
23. taz – “Palantir’s Use in Germany and the Positions of the Federal States”: https://taz.de/Einsatz-von-Palantir-in-Deutschland-Fast-alle-Bundeslaender-sind-dagegen-ein-Grund-zur-Entwarnung/!6186360/
24. taz – “Constitutional Complaint Against Palantir”: https://taz.de/Verfassungsbeschwerde-gegen-Palantir-/!6102522/
25. Staatsanzeiger Baden-Württemberg – “Palantir in Southwestern Germany and What the Police Software Means”: https://www.staatsanzeiger.de/nachrichten/politik-und-verwaltung/palantir-im-suedwesten-was-die-polizei-software-bedeutet/
26. Tagesspiegel Background – “How the German States Use Palantir”: https://background.tagesspiegel.de/digitalisierung-und-ki/briefing/wie-die-bundeslaender-palantir-nutzen
27. Tagesspiegel – “Access to Millions of Data Records and the Constitutional Complaint Against Police Software”: https://www.tagesspiegel.de/politik/digitalisierung-ki/palantir-hat-zugriff-auf-millionen-daten-verfassungsbeschwerde-gegen-regeln-fur-polizei-software-eingereicht-14071667.html
28. ZDFheute – “The Palantir Manifesto and the Interior Ministers’ Response”: https://www.zdfheute.de/politik/deutschland/palantir-manifest-polizei-innenminister-bundeslaender-reaktion-kritik-100.html
29. ZDFheute – “Palantir’s Use and the Search for an Alternative”: https://www.zdfheute.de/politik/deutschland/palantir-einsatz-polizei-deutschland-alternative-100.html
30. Legal Tribune Online – “Federal Constitutional Court Declares the Use of Hessendata Unconstitutional”: https://www.lto.de/recht/nachrichten/n/bverfg-1BvR154719-1BvR263420-einsatz-software-hessendata-durch-polizei-verfassungswidrig
31. Legal Tribune Online – “Society for Civil Rights Files Constitutional Complaint Against Palantir in Bavaria”: https://www.lto.de/recht/nachrichten/n/gff-verfassungsbeschwerde-gegen-palantir-bayern-polizei-ueberwachung
32. netzpolitik.org – “Federal Constitutional Court Judgment on Automated Data Analysis”: https://netzpolitik.org/2023/urteil-des-bundesverfassungsgerichts-automatisierte-datenanalyse-fuer-die-vorbeugende-bekaempfung-von-straftaten-ist-verfassungswidrig/
33. netzpolitik.org – “Assessment of Palantir Software in Switzerland”: https://netzpolitik.org/2025/schweiz-palantir-software-hat-verheerende-risiken/
34. heise online – “Missing Link: Palantir as a Center of Power and How the Software Works”: https://www.heise.de/hintergrund/Missing-Link-Machtzentrale-Palantir-eine-Software-lenkt-Organisationen-10463034.html
35. Police-IT – “Consequences of the Constitutional Court Judgment for VeRA and Other Palantir Systems”: https://police-it.net/folgen-des-bverfg-urteils-fuer-vera-und-andere-palantir-systeme
36. t-online – “Coalition Divided over the Use of Palantir in Germany”: https://www.t-online.de/nachrichten/deutschland/innenpolitik/id_101092782/palantir-software-koalition-uneinig-ueber-einsatz-in-deutschland.html
37. Tageblatt – “SPD Seeks to Prevent Federal Use of Palantir Technology”: https://www.tageblatt.lu/International/SPD-will-Einsatz-von-Palantir-Technik-im-Bund-verhindern-24524.html
38. it-fachportal – “Justice Minister Does Not Expect Palantir to Be Used at the Federal Level”: https://www.it-fachportal.de/68270-justizministerin-rechnet-nicht-mit-palantir-einsatz-auf-bundesebene/
39. Notebookcheck – “Digital Industry Warns U.S. Supreme Court Against Access to Data from Germany”: https://www.notebookcheck.com/Digitalwirtschaft-warnt-Obersten-Gerichtshof-der-USA-Kein-Zugriff-auf-Daten-aus-Deutschland.279238.0.html
40. Handelsblatt – Interview with Palantir CEO Alex Karp on Preventing Attacks; institutional reference: handelsblatt.com, search term: Alex Karp Interview Palantir Anschläge Deutschland.
41. Fraunhofer Institute for Secure Information Technology – Source-Code Review of the Analytics Platform Commissioned by the Bavarian State Criminal Police Office; institutional reference: sit.fraunhofer.de, search term: VeRA Quellcodeprüfung Palantir.

International Use and Contracts

42. NATO Communications and Information Agency, March 25, 2025 – Commissioning of the Maven Smart System; institutional reference: ncia.nato.int, search term: Maven Smart System NATO 2025.
43. United Kingdom Ministry of Defence – Extension of the Contract with Palantir; institutional reference: gov.uk, search term: Palantir Ministry of Defence contract 2026.

Corporate and Financial Data

44. Palantir Technologies – Second-Quarter 2026 Earnings Report, Investor Relations: https://investors.palantir.com/
45. TradingKey – “Analysis of Palantir Stock with a Breakdown of 2026 Quarterly Results”: https://www.tradingkey.com/de/analysis/stocks/us-stocks/262125217-is-palantir-stock-buy-now-deep-dive-aip-valuation-risks-price-targets-tradingkey

© Michael Hollister – All rights reserved. Redistribution, publication or reuse of this text requires express written permission from the author. For licensing inquiries, please contact the author via www.michael-hollister.com.