The NHS should end its dependence on Palantir and move to an open-standards data platform after a short transition
Source: https://www.theguardian.com/profile/ben-coleman. "British people don’t trust Palantir – it’s not too late to reverse the NHS’s dependence on it | Ben Coleman | The Guardian." November 9, 2026. www.theguardian.com
The Gist
The author, a Labour MP on the health committee, argues the government shouldn't renew Palantir's NHS data contract beyond a short transition. The company's benefits are unproven, the public doesn't trust it, and the NHS may be locked in. He wants a system built on open, shared standards so the NHS can switch suppliers and keep control of its own data.
Conclusion
Ministers should not extend Palantir's contract beyond a short, conditional transition (max two years), and should instead build a federated data platform based on national open standards that gives the NHS full control of its technology and patient data.
Premises
- The NHS is at risk of vendor lock-in with Palantir: it owns the applications but rents the platform, ministers have not clarified which arrangements end with the contract, and they declined to audit which Palantir tools trusts depend on.
- Palantir's benefits are contested or unproven: the Health Foundation found no noticeable change in discharge delays, one trust accounts for 84% of the claimed outpatient waiting list reduction, NHS England admits cause and effect are unproven, and a full evaluation is not due until 2029.
- Take-up and trust are weak: a third of signed-up trusts have never logged in, 16 trusts and health boards say they do not use the platform, and mistrust is driving growing patient opt-outs, which undermines a data platform that depends on patient trust.
- Open standards would let regions choose and swap suppliers (NHS in-house teams or small British firms) without starting again, while data stays regional and is shared only under strict permissions.
- Britain has the talent and existing examples (Greater Manchester, NHS OneLondon) to build an equivalent, as experts like Nigel Shadbolt and the national data guardian's call for open standards support.
- Sovereignty over health data and AI requires a system the NHS controls, not merely data held on British soil, and replacing Palantir with a single British firm would not provide that control.
Assumptions
- Vendor lock-in and dependence on a foreign firm are serious enough harms to outweigh the disruption and cost of switching.
- Public distrust of Palantir is well-founded and will persist, and opt-outs will materially damage the platform's value.
- Low use in some trusts shows that reversal is feasible and that the NHS is not irreversibly dependent.
- A decentralised open-standards model can be built and can deliver comparable functionality to Palantir within about two years.
- Conditions placed on Palantir during a transition (freeze on new applications, publication of costs, a test migration) are enforceable and will be honoured.
- Absence of proven benefits means the benefits are small or non-existent, rather than simply not yet measured.