As Palantir vies for a £480M contract to build a “federated data platform” for the UK's NHS, experts say the company's Foundry tool is costly and hard to use
Ian Johnston / Financial Times :
Context & Ripple Effects
Palantir's push into UK public-sector healthcare has been building for a year: the company prepared a £360M bid in mid-2022 to underpin NHS operations, and by late 2023 doctors and privacy campaigners were openly organizing against its expected win of the larger £480M federated data platform contract. This FT report adds a second axis of criticism beyond privacy: experts say Foundry itself is expensive and difficult to use.
That matters because the criticism proved durable rather than episodic — Palantir and Accenture ultimately won the £480M overhaul, and by 2026 the platform sat between two poles: NHS data chief Ming Tang praising its results while public and political pressure mounted around the contract's break clause.
First-order effects
- Doctors and privacy campaigners opposing Palantir's candidacy gain a fresh argument that is not about data ethics but about value for money — cost and usability become part of the procurement debate alongside consent.
- NHS procurement officials must weigh expert claims about Foundry's cost and complexity against Palantir's existing 600-person UK footprint and stated plans to add staff.
Second-order effects
- Competing bidders can differentiate on implementation burden and total cost of ownership rather than privacy alone, forcing any winning consortium to justify integration spend.
- A win would channel follow-on work toward systems integrators like Accenture, shifting where the contract's value accrues — services and configuration around the platform, not just licences.
Third-order effects
- If governments keep awarding centralized data platforms to a small set of vendors despite expert reservations, the result is structural lock-in: the 2026 fight over NHS England's break clause shows how politically contested these contracts become only after they are too embedded to unwind cheaply.
- The split between internal endorsement (Tang's 'outstanding results' memo pushing deeper rollout) and external pressure frames the recurring pattern for sovereign AI procurement — official champions inside agencies versus public legitimacy battles outside them.
The trend: Western governments are consolidating national health and civic data onto a handful of commercial platforms, where procurement decisions create dependencies that long outlast the controversies surrounding their award.