Drawing on his experience as a health minister and chair of the Health and Social Care Select Committee, GK’s strategic advisor Steve Brine examines the debate over Palantir’s role in the NHS and argues that evidence of patient benefit, delivery risks and credible alternatives should guide ministers’ decisions.
Senior Vice-President of Palantir UK Tom Watson is an old friend. We were in the Commons together for nine years and our paths crossed many times, not about health it must be said, but music, where we share a vision of creative sovereignty in the face of a powerful challenge from big tech.
Tom’s return to the political headlines was always likely to provoke a reaction but I was pleased to see him back; primarily because it means he once again had the all-clear in his battle with prostate cancer.
Back how? The former deputy leader of the Labour Party has joined American technology company Palantir as a senior vice-president. The appointment comes just as ministers approach a big decision about the company’s future role in the NHS.
Appearing recently on a BBC podcast, Tom warned against allowing what he called “mob rule” to determine government procurement decisions. His intervention was characteristically direct, but it raised a question that goes well beyond one company.
The bottom line is this: should our government allow political discomfort (justified or otherwise) with a supplier to override the practical interests of NHS patients?
At the heart of this debate is the NHS Federated Data Platform (FDP), a programme designed to connect information held across different parts of the health service, helping clinicians and managers make decisions about patient care, waiting lists and capacity.
In November 2023, a consortium led by Palantir secured a seven-year contract to deliver the FDP. Ministers now face a decision over a contractual break point in the early part of 2027.
Two parliamentary committees have recommended that the government take the option to break the contract and seek an alternative supplier.
And so, the search begins for compelling evidence that the existing arrangement is failing, or that a demonstrably better alternative exists, because in my experience the NHS cannot afford another technological false start.
Some NHS users have indeed complained about the technology itself. A British Medical Journal (BMJ) investigation published in July this year reported that NHS leaders had privately described aspects of the FDP as clunky and offering a poor user experience.
Having served as a health minister, and subsequently chaired the Health and Social Care Select Committee, I know how often the ambitions of NHS reform collide with the practical realities of delivery. The entire NHS Ten Year Plan is a case in point!
For decades, the NHS has struggled with fragmented information systems, inconsistent data and organisations unable to communicate with one another.
The consequences are not abstract. They mean operating theatres sitting underused while patients wait for treatment, unnecessary delays in discharging people from hospital, and clinicians spending time navigating disconnected systems.
The FDP was conceived to address precisely these challenges. It is not, as some commentary might suggest, a vast new database into which everyone’s medical records are poured. Rather, it provides a means of connecting and analysing information across existing systems, subject to governance and access controls.
The NHS does not need LESS intelligent use of data. It needs considerably MORE.
Indeed, the government’s wider ambitions for a modernised health service, including that Ten Year Plan, depend on greater use of artificial intelligence and a more joined-up approach to care.
Walking away from an established programme without a credible replacement risks repeating one of the NHS’s most frustrating habits: announcing transformation, beginning implementation and then changing direction.
Supporters of the FDP point to encouraging improvements. NHS England has reported increased theatre utilisation, more patients receiving procedures, improvements in cancer pathways and reductions in delayed hospital discharges.
But there is a necessary qualification.
Recent scrutiny, including by the Financial Times, has questioned whether some of the headline improvements can be directly attributed to the FDP. Some trusts have stopped using certain tools on the platform, and NHS England has acknowledged that improvements observed after the platform’s introduction do not automatically establish cause and effect.
That criticism deserves serious attention. But there is an important distinction between questioning the measurement of benefits and concluding that the technology has no value.
It is also striking that, unlike some previous NHS technology programmes, the FDP has not generated an overwhelming public backlash from the clinicians and operational staff expected to use it. There are complaints about usability and uneven adoption, certainly, but these sit alongside positive accounts from NHS organisations reporting practical improvements.
That is not proof of success, but neither is it the picture of an obviously failing system.
Incomplete evaluation is an argument for stronger scrutiny, not necessarily for abandoning a programme midway through implementation.
The politics of Palantir
Of course, much of the opposition to Palantir has little to do with such cold-eyed scrutiny of NHS operational performance.
The company’s American ownership, its co-founder Peter Thiel, and its work with defence, intelligence and immigration agencies have made it a lightning rod for political controversy.
These are matters on which people can hold sincerely different views, but we should be careful not to confuse opposition to a company’s activities elsewhere with evidence that its NHS contract is inappropriate or unsafe.
Palantir is not the owner of NHS patient data. It acts as a data processor, with NHS organisations retaining responsibility as data controllers.
The NHS relies extensively on international technology companies, pharmaceutical manufacturers, medical device suppliers and other private-sector partners. Their expertise is essential.
Surely the relevant test should be whether a supplier meets the necessary standards of security, effectiveness, accountability and value for money, not whether its founders hold political views with which some Labour MPs happen to disagree?
There are legitimate questions about supplier dependence and public trust, but they deserve to be distinguished from objections rooted primarily in political hostility towards the company.
Tom Watson’s intervention is interesting because it highlights an uncomfortable reality for government. Public procurement must operate within the law, but it must also serve the public interest.
In a recent conversation, someone described the FDP to me as a form of technological infrastructure that another supplier could, in theory, provide.
That is true, but it risks understating the complexity of what has been built. Palantir does not simply provide the cloud infrastructure on which the platform runs; its technology helps integrate data and supports the applications and operational workflows built around it.
There may well be credible alternatives, including other major international technology companies. But replacing the supplier is not necessarily the same as switching electricity providers. The government would need to establish what could be transferred, what would need rebuilding, how long that would take and at what cost.
And if the objective is greater technological sovereignty, replacing one American supplier with another hardly settles the argument.
Ministerial judgement
The government faces a difficult decision but, in my experience, the variables shift considerably when you’re making a political one rather than an operational one.
Palantir should be held rigorously to its contractual obligations, and the evidence of benefit must be strengthened. But – and this is my key point – the burden of proof should also apply to those advocating termination.
What would replace the existing platform? How quickly could an alternative supplier be brought in, and at what cost? What risks would that impose on patients, and would it deliver better results?
These are precisely the questions ministers must ask, and I am confident they are.
The wider lesson is that governments must resist the temptation to confuse a politically attractive announcement with successful public policy. Here I am less confident.
My instinct is that ministers should not exercise the February 2027 break clause unless the evidence presents a compelling case for doing so.
Ultimately, patients waiting for operations, cancer diagnoses or a safe discharge from hospital will not judge ministers by the ideological purity of their decisions.
They will judge them by whether the NHS is ‘fixed’.
