Opinion: The Palantir row is really a question about what the NHS is for

Earlier this month the cross-party Science, Innovation and Technology Committee told the government to use a break clause next February and walk away from NHS England's £330m Federated Data Platform, built by the American firm Palantir. The committee's language was unusually pointed: it cited the company's support for 'highly controversial policies and activities', and recommended that the NHS either build a replacement in-house or commission one from 'UK-owned and UK-based providers that are more compatible with UK values'. Amnesty International has backed the campaign, and some clinicians have refused to use the software at all.

There is a real and legitimate unease here, and it deserves to be taken seriously rather than waved away as squeamishness. Palantir is not a neutral plumbing contractor; it is a company with a distinct worldview and a portfolio of government work, including in defence and immigration enforcement, that many find troubling. The question of whose hands hold the nation's most intimate health records is not a technicality. It is one of the defining political questions of the digital state.

And yet the committee's recommendation, however well-intentioned, risks confusing two arguments that need to be kept apart: the argument about sovereignty and the argument about morality. Conflating them produces a satisfying headline and a poor decision.

The sovereignty case is the strong one

The most defensible reason to be wary of the contract has little to do with Palantir's other clients and everything to do with dependency. When a single foreign-owned firm becomes the connective tissue linking patient records across the health service, the state acquires a structural vulnerability. Switching costs balloon, the supplier's leverage at renewal grows, and a future dispute, commercial or geopolitical, could put critical infrastructure at risk. That is a sound, sober reason to want a credible exit option and a domestic alternative on the horizon.

This is the argument ministers should be making, and it points towards a strategy rather than a gesture. Building genuine UK capacity in large-scale data platforms is a worthwhile national objective. But it is the work of years, not of a single break clause, and it requires investment, skills and a tolerance for the failures that have plagued past NHS IT megaprojects. Walking away in February with nothing ready to replace it would not be sovereignty. It would be self-harm dressed up as principle.

The 'values' test cuts both ways

The morality argument is more slippery than its proponents allow. 'UK values' is a comforting phrase that dissolves on contact. The NHS already buys from a long list of multinational suppliers, from pharmaceutical giants to cloud providers, whose conduct in other markets would not survive a purity test. If we are to apply a values screen to procurement, it must be a consistent, transparent framework applied across the board, not a spotlight switched on for the company that happens to be politically toxic this quarter.

There is also a practical danger in elevating supplier identity over patient benefit. The entire point of a federated data platform is to let a clinician see a complete picture, to manage waiting lists, coordinate discharges and spot deteriorating patients faster. If the system is delivering those gains, then ripping it out on the basis of the vendor's reputation imposes a real cost on real patients. That cost has to sit on the scales too.

• Is there a UK-based alternative that can actually deliver at NHS scale today?
• What would migration cost, in money and in clinical disruption?
• Are the data-governance safeguards in the current contract adequate, regardless of the vendor?
• Would an in-house build repeat the cost overruns of past NHS IT programmes?

“The right question is not whether we like Palantir. It is whether the state can ever again allow itself to be this dependent on any single supplier for its most sensitive infrastructure.”

— Priya Nandakumar

A better path than the binary

The committee has done a service by forcing the issue into the open, and ministers should not simply renew on autopilot. But the choice is being framed too crudely as stay or go. A more mature response would treat the February break clause as leverage, not as a cliff edge. Use it to demand stronger data-governance terms, hard limits on how the data may be used and by whom, full transparency over sub-processing, and a binding commitment to portability so the NHS is never again locked in.

At the same time, the government should publish a genuine plan, funded and time-bound, for building or commissioning a domestic capability, and only switch when something credible exists to switch to. That is less emotionally satisfying than a clean break, but it is how serious states manage critical infrastructure. The alternative, ditching a functioning system to make a point and then improvising under pressure, is exactly the kind of decision that produces the next NHS IT scandal.

The deeper lesson runs past Palantir altogether. We have allowed a handful of firms to become indispensable to the running of the state, and we are only now noticing the bill. The answer is not to scapegoat the most unpopular of them, but to rebuild the public sector's own capacity to understand, commission and where necessary build the systems on which our lives increasingly depend. Get that right and the question of which company holds the contract becomes a great deal less frightening.