Advisory Analysis of the DGSI-Palantir Renewal: Three Years of Strategic Dependence
DGSI Renews Its Contract with Palantir Until 2028
The General Directorate for Internal Security has renewed its contract with Palantir for three years. The official announcement of December 15, 2025 confirms this unambiguously. This renewal covers the Gotham platform and integrated operational support services.[1][2][3]
Nine Years of Entrenchment Concealed Under the Discourse of Temporariness
DGSI declares it is awaiting "the deployment of a new sovereign tool." This formula conceals reality: the collaboration initiated in 2016 has produced structural entrenchment. Gotham now constitutes "the core" of the Directorate's analysis tools. Any migration requires complete reconfiguration of processing chains, revalidation of control protocols, retraining of teams. The political and operational cost of change grows inversely to the promise of transition.[4][5]
French Alternatives: Promises Without Candidates
Since 2018, French leaders have repeated an internal promise: develop a "French or European offering." Laurent Nuñez (DGSI director, now Minister of the Interior) announced it in 2018. The CEO of Thales confirmed it in 2020: two years would suffice. In 2022, an official call for tenders (OTDH — Heterogeneous Data Processing Tool) selected three candidates: Athea (Thales-Eviden), ChapsVision and Blueway.[6]
In December 2025, no solution has reached operational stage. Selection remains ongoing. DGSI renews Palantir for lack of approved substitute, not by strategic choice. The temporary has crystallized into structural dependence.[7]
Data Control: A Technical Guarantee That Obscures Deep Asymmetry
Palantir affirms that processed data is hosted in France, exclusive property of DGSI, accessible only by authorized agents. This technical description, accurate on its points, conceals an asymmetric reality.[8]
The software code remains proprietary — inaccessible to French expertise. Software updates are decided by Silicon Valley. The global architecture escapes national control. Palantir, founded with CIA financing (In-Q-Tel), has progressively diversified toward American healthcare (Medicare-Medicaid integration), toward the IRS for tax audit prioritization, toward private operators. This trajectory reveals a company designed to industrialize centralized interconnection of sensitive data — an objective that far exceeds the stated counter-terrorism needs.[9]
Three Structural Factors Underpin the Renewal Despite Everything
The security threat is real. Terrorist threats, cyber threats and foreign interference are not declining. Gotham has proven its operational utility, notably during the 2024 Olympics.[10]
Migration cost is prohibitive. Nine years of integration make any change technically risky. Reconfiguring analysis chains, revalidating protocols and retraining teams constitute considerable investment during a period of threat.
No competitor offers required guarantees. The three French candidates remain in selection phase. None will offer before 2027-2028 the immediate operational robustness required by an intelligence service.
The Paradox: The Trajectory of the Temporary That Takes Root
DGSI renews Palantir "pending" a sovereign solution. This three-year renewal pushes back to 2028 the concrete possibility of transition. At this pace — calls for tender launched in 2022, no finalists in 2025, new contract until 2028 — technological sovereignty will be pushed back to 2030-2035 minimum. Meanwhile, teams, operational procedures, analytical habits will become further entrenched. The argument of temporariness empties of content in the reality of time.[11]
Denmark and Germany have also chosen Palantir after security audit. Their official pragmatism contrasts with French rhetoric of permanent transition.[12]
What This Renewal Reveals: Strategic Failure
DGSI did not choose Palantir in 2025. DGSI had no other viable option. This distinction is decisive. It transforms a commercial contract into state technological dependence.
The French state now depends on an American company for intelligence exploitation methodologies, technical evolution capabilities, and implicitly, for the analytical priorities that the tool privileges. Declarations about "data ownership" or "French control" do not erase this asymmetry. It is inscribed in the very structure of the tool and decision-making power.
Three Imperatives for the 2025-2028 Period
Accelerate the OTDH program without delay. The three candidates must achieve security certification before 2027. This requires immediate political clarification: budget? Timeline? Public order guarantees? Institutional improvisation has not worked so far.
Plan transition from 2026. Begin coexistence tests between Gotham and the selected French platform, on limited flows. Each lost year increases the cost of exiting dependence.
Re-internalize critical competencies. Do not entrust technological sovereignty solely to Thales, Atos or Eviden. Create a sovereign competency core within the state itself, capable of designing, maintaining and evolving the tool independently.
Irrevocable conclusion: The three-year renewal with Palantir is established. It responds to a real security urgency. But it proclaims the absence of sovereignty strategy implemented since 2016. The three coming years constitute the ultimate window to reverse this trajectory. Beyond 2028, technological dependence will become irreversible. The choice belongs to the French government: build or accept dependence.[2][1][7][11]
