How the UK Saved Millions by Ditching Palantir — and What It Reveals About Government Tech Procurement

The Ministry of Housing, Communities and Local Government (MHCLG) has confirmed that replacing Palantir's Foundry platform with an in-house system for the Homes for Ukraine refugee scheme is saving "millions of pounds a year in running costs" [1]. The transition, completed in September 2025, followed a procurement process that the government's own chief commercial officer described as contrary to public procurement principles [2].

The story is not simply about one department saving money on one contract. It is a case study in how a major technology vendor gained a commercial foothold in UK government through a zero-cost offering, escalated costs rapidly, and how the civil service ultimately proved it could build a cheaper, functional alternative — raising uncomfortable questions about why it wasn't done that way from the start.

The Contract: From Free to £10 Million in Two Years

When Russia invaded Ukraine in February 2022, the UK government scrambled to launch the Homes for Ukraine scheme, which matched displaced Ukrainians with UK hosts offering accommodation. The task required combining data from tens of thousands of visa applications with hundreds of thousands of accommodation offers — quickly [1].

Palantir offered to build and run the system for free for six months using its Foundry platform, waiving fees worth £3.5 million. The system was operational within nine days [2]. Speed was the priority; the government later acknowledged it "had prioritised delivery over finding an optimal value provider" [3].

Source: National Audit Office / TechMonitor

Data as of Sep 1, 2025CSV

After the free period ended in September 2022, the Department for Levelling Up, Housing and Communities (DLUHC, later renamed MHCLG) directly awarded Palantir a 12-month contract worth £4.5 million. In August 2023, a second 12-month extension followed at £5.5 million — a 22% increase [2]. Neither contract was competitively tendered.

The total paid spend reached £10 million. Including the initial waived fees, the system's notional cost exceeded £13.5 million over roughly 30 months [2].

What the System Actually Did

Palantir's Foundry platform served as a case management system for local authorities administering the Homes for Ukraine scheme. It integrated data flows from the Home Office (visa applications), local councils (safeguarding checks, accommodation inspections), and sponsors — providing a unified view of each refugee's journey from application to placement [4].

Palantir's own account described the system as handling "a data-driven humanitarian operation" that processed applications, tracked safeguarding for unaccompanied minors, and enabled local authorities to coordinate housing placements [4]. By July 2023, the Home Office had processed more than 182,000 visa applications through the broader scheme, and approximately 227,000 people arrived in the UK under Ukraine visa routes [5].

The Replacement: Built In-House

Coco Chan, a senior digital leader on the project, described the replacement as a system built on existing government technology standards rather than a proprietary commercial platform. The goals were explicit: "save significant support costs, control the system data and code" [1].

The in-house team delivered a working replacement by September 2025. MHCLG described it as "better, easier to use, and cheaper" [1]. Chan noted that the department "set a precedent by moving a complex live system to an in-house set up, reducing reliance on external suppliers" [1].

Former government technology advisor Stefan Czerniawski commented that "when given suitable resources the Civil Service can often outperform private companies like Palantir" [1]. Contributors to technical discussions noted that the scale of the problem — hundreds of thousands of records — was well within the capability of standard government digital services [6].

The Procurement Failure

The National Audit Office (NAO) investigation, published in October 2023, documented how the initial free offering created conditions that made competitive procurement impractical [3].

The government's Chief Commercial Officer wrote to Palantir expressing concern about "the practice of offering services to public sector customers for a zero or nominal cost to gain a commercial foothold, contrary to the principles of public procurement which usually require open competition" [2].

When DLUHC investigated alternatives in early 2023, it concluded that switching would be "costly and risky." Palantir itself noted that when alternatives were assessed, they were "viewed as likely to 'cost more', as well as carrying 'risks to implementing on time' and 'potential quality issues'" [2]. This is the classic vendor lock-in dynamic: the free offering creates switching costs that make continued use appear economically rational even at inflated prices.

The House of Commons Public Accounts Committee examined the scheme and found that the speed of deployment meant "the usual testing was not carried out and some local authorities found the system confusing to use" [3].

The Steelman Case for Palantir

Defenders of the original contract make several points. First, the system was delivered in nine days during a genuine humanitarian emergency — no government digital team could have matched that speed [2]. Second, Palantir's Foundry platform offers capabilities that go beyond simple case management: cross-database linking through its "ontology" layer, real-time analytics, fraud detection, and the ability to integrate disparate data sources into a unified operational picture [7].

In other contexts, these capabilities have proved substantial. US Immigration and Customs Enforcement has spent over $200 million on Palantir contracts for investigative case management that integrates travel histories, visa records, biometric data, and social media [7]. The UK's Financial Conduct Authority has contracted Palantir specifically for fraud detection capabilities [8]. The platform's ability to find connections across millions of records — what Palantir calls creating a "unified data landscape" — represents genuine technical depth [7].

The question is whether the Homes for Ukraine scheme actually required that depth. Multiple technical observers argued it did not — that matching accommodation offers to visa applications is "a completely standard kind of challenge" for government digital services [6].

What We Don't Know

MHCLG has not published a precise annual savings figure, stating only "millions." Without knowing the in-house system's development cost, ongoing maintenance budget, and staffing requirements, the net saving is impossible to independently verify [1].

Nor has the government published comparative data on system performance — whether case throughput, error rates, or processing times have changed since the switch. The weekly Ukraine visa statistics publication was discontinued in December 2024 [5], making independent assessment of outcomes difficult.

There is no public evidence that the Information Commissioner's Office (ICO) conducted a formal review of either the Palantir system or its replacement. Civil liberties groups have raised concerns about Palantir's data handling in other UK government contexts — including police surveillance networks and NHS data access — but specific ICO findings on the Homes for Ukraine system have not been published [9][10].

The Broader Pattern: Palantir and UK Government

The Homes for Ukraine case sits within a larger story of the UK government's relationship with Palantir — and growing pressure to reduce dependence.

The most significant parallel is the £330 million NHS Federated Data Platform contract, which faces a break clause in spring 2027. Junior minister Zubir Ahmed stated: "If, at the point of the break clause, we evaluate and find that there are other providers that can do the job better, then of course that needs to be looked at" [11]. Science minister Patrick Vallance has pledged "a different approach" aligned with sovereign technology policy [11].

MP Martin Wrigley highlighted a structural concern: under the NHS contract, "all the specially written software and intellectual property rights belong to the supplier," leaving the NHS without ownership of connection software after the agreement ends [11]. This mirrors the Homes for Ukraine situation — the decision to move in-house was partly motivated by the desire to "control the system data and code" [1].

Greater Manchester has refused to participate in the NHS platform with Palantir [12]. Green Party MPs have called for "an immediate inquiry" into Palantir contracts, warning that "the UK's reliance on Palantir risks making it even more difficult for the UK to confront or even publicly disagree with the US administration" [9].

Source: UNHCR Population Data

Data as of Dec 31, 2025CSV

International Context: The UK as Cost Outlier

The UK's refugee-related spending is already an international outlier. The Center for Global Development reports UK per-refugee costs of nearly $26,000 annually — two-and-a-half times Canada's $10,100, more than three times Italy's $7,800, and over four times France's $6,400 or Germany's $6,100 [13].

Source: Center for Global Development

Data as of Jun 1, 2024CSV

These figures reflect total hosting costs (dominated by accommodation), not technology costs specifically. But they establish that the UK's refugee system operates at premium prices across the board. The Palantir contract — £10 million over two years for what proved to be a replaceable system — fits that pattern.

Germany, which hosts 2.7 million refugees (more than any other European country), uses a mix of federal and state-level IT systems built largely in-house or through European contractors [14]. France's OFPRA (asylum determination body) runs its own case management infrastructure. Neither country has comparable reliance on a single US technology vendor for refugee processing.

Source: UNHCR Population Data

Data as of Dec 31, 2025CSV

Accountability and Lessons

The NAO report and Public Accounts Committee scrutiny represent the formal accountability mechanisms triggered by this case. But no individual has faced consequences for the procurement decision, and no formal finding of impropriety has been made [3].

The lessons are clear but recurring: free offerings from technology vendors create dependencies. Sole-source contracts escalate costs. Government digital teams can build functional alternatives when resourced to do so — but are often bypassed in favour of commercial solutions during emergencies, creating long-term cost problems.

The MHCLG replacement demonstrates that exit from vendor lock-in is possible. Whether the UK government applies this lesson to its larger Palantir entanglements — particularly the £330 million NHS contract — remains an open question.

What Remains Unclear

Several questions lack satisfactory public answers:

• Exact savings: MHCLG claims "millions" annually but has not disclosed the in-house system's total cost of ownership, including development, hosting, and staffing.
• Capability comparison: Whether the replacement handles the same data integration complexity or has reduced scope to match reduced cost.
• Processing outcomes: No published data comparing system performance before and after the switch.
• Data protection: No public ICO assessment of either system's compliance with UK GDPR requirements, particularly regarding data sharing with other government departments.
• Lobbying and influence: No disclosed commercial relationships or revolving-door appointments connected to either the original Palantir award or the replacement decision.

The "millions saved" headline is likely accurate in narrow terms — the annual Palantir licence exceeded £5 million, and an in-house system built on existing government infrastructure will cost less to run. But the full accounting — including transition costs, reduced functionality (if any), and the long-term maintenance burden — has not been made public.