The Mechanization of Racial Profiling: ICE's $85 Billion Procurement Architecture and the Contractual Infrastructure of Algorithmic Discrimination

Introduction

The Institutionalization of Automated Surveillance and Targeting

The One Big Beautiful Bill Act (OBBBA), enacted by Congress in July 2025, allocated $170.7 billion for immigration enforcement, with $75 billion directed specifically to Immigration and Customs Enforcement through fiscal year 2029.

This appropriation transforms ICE from a mid-tier federal law enforcement agency into the highest-funded law enforcement apparatus in the United States government—exceeding the entire Federal Bureau of Prisons budget by a factor of 8 and constituting the single largest investment in immigration enforcement infrastructure since the Department of Homeland Security's 2003 establishment.

The $85 billion reference in current appropriations encompasses ICE's total allocation across multiple fiscal years within this authorization ceiling.

What distinguishes this appropriation from previous immigration enforcement funding is neither its magnitude nor its stated objectives, but rather the explicit architectural mechanisms through which it operationalizes racial profiling through technological intermediation.

The procurement records, contractor selections, and algorithmic systems authorized by this funding allocation reveal the systematic conversion of immigration enforcement into what might be termed "algorithmic racial profiling as infrastructure"—a system wherein discrimination becomes embedded not in explicit policy directives but in the technical specifications of contracted data analytics, skip-tracing algorithms, and machine-learning systems deployed across detention, targeting, and removal operations.

The Legal and Constitutional Inversion

From Constraint to Authorization

Prior to September 2025, federal courts had begun to impose constitutional constraints on ICE racial profiling practices.

In July 2025, Federal District Judge Margaret Frimpong, examining ICE street arrests in Los Angeles, determined that immigration enforcement operations had violated Fourth Amendment protections by targeting individuals based on "apparent race or identity," "perceived national origin," "Spanish language," "English spoken with an accent," and occupational classification (workers in construction, agriculture, hospitality).

Judge Frimpong explicitly found that these factors were "no more indicative of illegal presence in the country than of legal presence"—meaning that the targeting criteria bore no rational relationship to probable cause or reasonable suspicion.

Yet in September 2025, the United States Supreme Court, in a 6-3 decision, suspended Judge Frimpong's restraining order pending appeal, thereby lifting constitutional constraints on racial profiling.

Justice Sonia Sotomayor's dissent articulated the constitutional horror: "The Government…has all but declared that all Latinos, U.S. citizens or not, who work low wage jobs are fair game to be seized at any time."

The Supreme Court's decision converted what had been constitutional constraint into constitutional authorization. Racial profiling, previously subject to Fourth Amendment scrutiny, became permissible law enforcement methodology.

Critically, the timing of this legal transformation coincided precisely with ICE's procurement cycle for skip-tracing contracts. In October 2025, the Department of Homeland Security issued a Request for Information soliciting vendors for skip-tracing services.

The timing relationship is not coincidental: the legal authorization for racial profiling and the technological contracting for racial profiling infrastructure emerged simultaneously, suggesting deliberate orchestration of legal permissibility and technological operationalization.

The Skip-Tracing Architecture

Outsourcing Racial Targeting to Private Contractors

The skip-tracing procurement, finalized in December 2025, awarded contracts to 10 companies with a combined ceiling of approximately $1.2 billion over 2 years.

The primary contract recipients include Capgemini Government Solutions ($365.8 million ceiling, $4.8 million obligated as of early 2026), Palantir Technologies ($30 million for "ImmigrationOS" platform), and specialized contractors including AI Solutions 87, providing what were candidly marketed as "bounty hunter AI agents."

The architectural design of skip-tracing outsourcing accomplishes several objectives from the perspective of both enforcement effectiveness and legal insulation.

First, it converts detention targeting from a direct government function into a contractor-mediated process, thereby diffusing agency responsibility across multiple private entities.

When ICE receives targeting leads from contractors, the agency claims to be executing objective analytical outputs rather than subjective targeting decisions.

When contractors generate targeting leads, they claim to be applying algorithms programmed by other entities, with ICE oversight consisting of accepting or rejecting recommendations. Responsibility becomes dispersed—no single actor experiences full culpability for targeting outcomes.

Second, skip-tracing outsourcing creates operational distance between the racial targeting decision and the enforcement action. A contractor's algorithm generates a "confidence score" for an individual's address. ICE receives the output.

An immigration officer executes an arrest based on the algorithmic recommendation. The contractor claims to provide technical analysis, not targeting guidance.

The officer claims to be executing leads provided by analytical systems. The originating decision—why this individual was selected for targeting—becomes obscured within layers of technical intermediation.

Third, the profitability incentive structures transform contractors from service providers into active participants in enforcement escalation.

The procurement specifications authorize "performance-based incentives" rewarding successful address verifications, "volume-based incentives" providing bonuses for exceeding caseload percentages, and "quality assurance incentives" tied to positive ICE feedback.

🛑Monthly assignments of 50,000 addresses per contractor create systematic throughput pressure.

The business model incentivizes contractors to locate individuals efficiently and repeatedly, generating continuous enforcement leads regardless of individual threat assessment or Fourth Amendment compliance.

The Palantir-Medicaid Nexus

Health Data Repurposed for Deportation

The most consequential technological development is Palantir's integration of Medicaid data into its "ELITE" (Enhanced Leads Identification and Targeting for Enforcement) platform provided to ICE. Through this system, health insurance information generated by immigrant beneficiaries seeking medical care becomes converted into targeting data for deportation.

The mechanism operates as follows: Individuals without legal immigration status who qualify for certain Medicaid benefits provide address information to state health agencies to obtain coverage.

Medicaid data is shared with the Department of Health and Human Services. Palantir's contracted system ingests Medicaid databases alongside law enforcement records, commercial data, property records, and other sources, synthesizing them into unified dossiers.

The ELITE tool maps individuals' locations, associates, and employment, generating confidence-scored deportation targets.

This process accomplishes what the Electronic Frontier Foundation has termed "turning health care into deportation surveillance."

The chilling effect is immediate and documented: immigrant families reporting decreased health care seeking, prenatal care avoidance, and medication adherence abandonment due to fear that health information will be weaponized for enforcement.

Patients never consented to medical data repurposing; the transformation of health information into targeting data occurs administratively, without judicial process or individual notification.

The constitutional and ethical dimensions require emphasis: Medicaid represents a social insurance program predicated on the assumption that individuals disclosing health information are protected by medical confidentiality.

The redirection of that information for deportation enforcement constitutes a fundamental breach of the implicit social contract.

Moreover, the repurposing disproportionately affects undocumented immigrants, creating a two-tiered system wherein immigrant access to health care exists only at the cost of surveillance exposure.

The Algorithmic Discrimination Problem

Bias in Commercial Data and Skip-Tracing Accuracy

The technical specifications of skip-tracing contracts reveal underlying structural vulnerabilities to algorithmic discrimination.

Contractors are authorized to utilize "government-furnished case data, commercial data verification, automated and manual real-time skip tracing, use of search engines, and collection of photos and documents to verify residence or employment."

The problem inherent in this methodology is that commercial data sources—property records, credit reports, social media data, vehicle registration databases—contain systematic inaccuracies and gaps disproportionately affecting marginalized populations.

Research in algorithmic bias demonstrates that commercial databases consistently misidentify, underidentify, or fail to capture individuals from communities of color.

Address verification algorithms trained on predominately white populations exhibit degraded performance on communities with non-standard address formats, transient housing situations, or multi-family occupancy.

Skip-tracing algorithms that rely on credit bureaus systematically disadvantage undocumented immigrants who deliberately avoid financial system visibility to prevent immigration exposure.

When immigration enforcement targeting depends on algorithmic recommendations based on biased commercial data, the bias becomes operationalized within enforcement architecture.

The skip-tracing procurement documents explicitly authorize contractors to use "physical observation services" in addition to data analytics—meaning that contractors are empowered to conduct surveillance photography of individuals' homes and workplaces.

This physical surveillance component, combined with algorithmic targeting, creates a hybrid system wherein algorithmic recommendations directing contractors to specific addresses are supplemented by physical surveillance verification, creating intensive targeting infrastructure.

The Acceleration of Detention Expansion

From Judicial Constraint to Infrastructure Scaling

The OBBBA authorized $45 billion for detention capacity expansion, with objectives of creating 100,000 new detention beds by fiscal year 2029—increasing total ICE detention capacity from approximately 65,000 beds to 165,000 beds, exceeding the entire federal prison population.

The ACLU's recent FOIA litigation revealed that ICE is considering 7 new detention facilities, several of which are former correctional facilities with documented histories of violence, sexual assault, and corruption.

This detention expansion represents the infrastructural consequence of skip-tracing acceleration. If ICE's targeting capacity expands through algorithmic skip-tracing, enforcement volume necessarily increases.

The detention infrastructure must scale proportionally to accommodate increased arrests. The $45 billion detention authorization provides the physical capacity infrastructure; the skip-tracing contracts provide the targeting infrastructure; together they operationalize mass enforcement.

Moreover, the detention expansion contracts explicitly funnel resources to private prison companies—GEO Group and CoreCivic—with long-term contract extensions. These companies receive 10-year contract renewals creating stable, incentive-aligned profit streams.

The companies' financial interest in detention population expansion creates parallel incentive structures: contractors profit from detention numbers, skip-tracing vendors profit from targeting volume, detention facility operators profit from bed occupancy.

Every participant in the system has financial interest in maximizing enforcement regardless of individual culpability or Fourth Amendment compliance.

International and Domestic Critique

The Emerging Resistance

By late January 2026, the procurement arrangements began generating significant international scrutiny. The French government, represented by Finance and Economy Minister Roland Lescure, demanded transparency from Capgemini regarding its skip-tracing contract with ICE.

French unions CFDT and CGT called for contract termination.

The British Columbia Attorney General urged Canadian companies to reconsider collaborations with ICE.

By February 1, 2026, Capgemini announced it was selling its U.S. subsidiary operating the ICE contract, citing legal review and public controversy.

Palantir faced internal employee resistance, with 13 former employees releasing a joint letter criticizing the company's ICE collaborations and 450+ tech employees signing a public letter demanding CEO cancellation of immigration enforcement contracts.

Palantir internal wiki acknowledged "widespread reports of ICE racially profiling and deporting citizens" and noted "reputational risks" but maintained that ICE was "committed to avoiding unlawful or unnecessary targeting."

The contradiction in Palantir's positioning—acknowledging racial profiling while maintaining contractor relationships—illustrates the moral and legal incoherence of the skip-tracing architecture. If ICE is acknowledged to engage in racial profiling, then data analytics systems providing targeting recommendations to ICE are operationalizing that profiling.

There is no technical mitigation for racial targeting when the targeting agency itself engages in profiling.

The Political Economy of Enforcement Expansion

Diffusion of Responsibility and Profit-Aligned Incentives

What distinguishes the 2025-2026 immigration enforcement apparatus from previous enforcement regimes is the systematic alignment of profit incentives across public and private entities around enforcement volume. Skip-tracing contractors earn performance bonuses for volume.

Detention facility operators profit from bed occupancy. Private security companies contract for facility operations. Technology vendors receive recurring payments for algorithmic systems. Immigration judges operate under threat of termination for adverse decisions.

Every actor in the enforcement apparatus has financial or employment incentive aligned with enforcement escalation.

The system is deliberately engineered to maximize through-put regardless of individual case merit, Fourth Amendment compliance, or humanitarian consequences.

This represents what critical scholars term the "carceral infrastructure"—the institutional architecture wherein incarceration becomes the default mechanism for managing vulnerable populations precisely because that default generates profit and employment.

Cause and Effect

How Procurement Architecture Produces Discriminatory Outcomes

The causal chain operates as follows: Congress appropriates $170.7 billion. The Trump administration designates $75 billion for ICE.

The Supreme Court lifts racial profiling constraints. DHS issues procurement solicitations for skip-tracing services.

Contractors bid on identifying immigrants for targeting. Capgemini, Palantir, and others win contracts worth billions.

These contractors deploy algorithms utilizing Medicaid data, commercial records, and physical surveillance. Skip-tracing outputs identify individuals for targeting. Immigration officers receive algorithmic recommendations and execute arrests. Detention facilities receive deportation candidates and expand operations. Private companies earn profits.

Contractors collect performance bonuses. Immigration judges decline to review cases and avoid termination.

Racial profiling becomes operationalized at scale, diffused across multiple institutions, and transformed from explicit policy into technical infrastructure.

At each step, individual actors claim limited responsibility: the appropriator funded legally authorized functions; the contractor applied neutral algorithms; the officer executed outstanding leads; the judge applied statutory law.

Yet collectively, the system produces discriminatory enforcement targeted at Latino communities at scales unprecedented in contemporary American governance.

Future Trajectory

The Normalization of Algorithmic Racial Profiling

If current trajectories persist without significant constitutional or legislative constraint, several developments are predictable.

First, the successful operationalization of skip-tracing for immigration enforcement will establish precedent and demonstrate viability for application in other law enforcement domains.

Border Patrol, ICE, FBI, and local police may increasingly adopt skip-tracing contracting models. Criminal justice enforcement may incorporate similar vendor-relationship architectures, creating parallel systems of algorithmic targeting across multiple law enforcement domains.

Second, the technological infrastructure assembled through these contracts—ELITE, ImmigrationOS, bounty hunter AI agents—will likely persist regardless of political transitions. Technology procurement contracts typically extend across administrations.

Once data systems are integrated into operational practice, removing them requires explicit action. The machinery of algorithmic targeting, once constructed, develops organizational constituencies—officers trained on systems, contractors integrated into operations, data dependencies embedded in case management platforms.

Third, detention expansion to 165,000+ beds will create permanent infrastructure demanding perpetual population to justify capital and operating costs.

Once detention facilities are constructed, financial and employment constituencies develop around their operation. The incentive structure created by this infrastructure will systematize enforcement escalation.

Conclusion

The Constitutional and Ethical Reckoning

The 2025-2026 transformation of immigration enforcement through procurement-authorized algorithmic racial profiling represents a critical moment in American governance.

The nation faces explicit choice: whether to permit the institutionalization of discrimination through technical intermediation, or to assert that constitutional protections against racial profiling cannot be evaded through privatization of targeting, algorithmic obfuscation, or profit-incentivized enforcement escalation.

The Supreme Court's lifting of racial profiling constraints, Congress's unconditional appropriation, and the administration's procurement of skip-tracing infrastructure collectively operationalize what might be termed the "normalization of systematic discrimination."

What emerges is not a discrete program subject to policy reversal, but rather an institutional architecture deliberately constructed to persist across time, diffuse responsibility across actors, and align profit incentives with enforcement escalation.

The international backlash—French government pressure, Canadian provincial opposition, European media scrutiny—suggests emerging recognition that this represents not merely domestic immigration policy but rather a transformation in how modern democracies govern marginalized populations through technology.

The United States is constructing a publicly visible model of algorithmic racial profiling as infrastructure, potentially establishing precedent that authoritarian regimes and illiberal democracies internationally may follow.

Whether American democracy possesses constitutional and political resources to constrain this trajectory—through courts enforcing Fourth Amendment protections, Congress imposing procurement conditions, or public pressure compelling corporate accountability—remains the critical question of the moment.