The place that told Palantir no
Greater Manchester’s managed something rare in NHS procurement: it’s said no, and kept saying it. While nearly every other NHS trust and almost every integrated care board in England has switched on Palantir’s federated data platform in some form, Greater Manchester’s care system has repeatedly stayed out. That makes it the lone holdout in England, which is a slightly awkward place to be when the rest of the room has already signed the attendance sheet.
The region is not some sleepy pilot with a few clinics and a neat spreadsheet. It covers about 3 million people, so this is a full-scale public health decision, not a token act of resistance for the benefit of a committee briefing. In practical terms, the refusal means one of the country’s biggest urban health systems has chosen to keep its own data setup rather than fold itself into Palantir’s national model.
When a place this large says no, the answer stops being local procurement and starts looking like a national argument.
That argument’s heating up because the government’s moving toward a February deadline on the contract. Ministers have to decide whether to pull the plug early or let the deal run until the end of the decade. That leaves a narrow window for anyone trying to argue the case either way, and Greater Manchester’s become the most obvious counterexample for critics who want an exit.
The politics around it have grown louder than the software itself. Petitions have circulated, public anger’s spilled into the usual online corners, and MPs have started treating the issue less like a systems upgrade and more like a test of who gets to shape NHS data policy. In tech news terms, this would usually be filed under procurement and platform architecture. In ai policy terms, it’s drifted into a bigger argument about vendor dependence, public trust and whether the state should hand a single company such a wide field of view.
There’s also a digital culture angle here, though nobody on either side is likely to use those exact words in a boardroom. Patients are being asked to trust a data system they don’t see, built by a company many of them only know through controversy and headlines. That makes every local decision feel larger than it’d on paper. Greater Manchester’s refusal’s become a symbol because it offers a simple, stubborn fact in a debate full of abstractions: one of England’s biggest health regions looked at the offer and decided to keep its own house in order.
What that house contains, and why so many other parts of the NHS have decided to move in with Palantir anyway, is where the story gets more interesting.
What Palantir’s NHS platform is supposed to fix
After the local refusal comes the more awkward question: what, exactly, was the national system supposed to solve in the first place?
In 2023, the NHS commissioned Palantir to build a federated data platform for health records and operational data, with rollout beginning in early 2024. The idea was not to rip out every local database and force everyone onto one giant screen. It was meant to connect national data with the systems already sitting inside hospitals, integrated care boards, and regional planning teams. In practice, that meant giving planners and clinicians a common layer that could pull together information without making every trust rebuild its own stack from scratch. The NHS’s own announcement framed the project as a way to improve care for millions of patients, and that is the pitch the Palantir NHS deal has relied on ever since. NHS England’s 2023 announcement laid out the basic idea.
The promise was plain enough: stop making clinicians hunt across half a dozen systems just to answer a basic question.
That basic question might be, “Has this patient already had the scan?” or “Which ward has a bed?” or “Who’s ready for discharge but still stuck waiting for the paperwork to catch up?” The platform was sold as a way to make those answers easier to find, faster to act on, and less dependent on someone knowing the right person to ring at 7:30 a.m. Before the day gets messy.
The NHS data problem was never glamorous. It was clunky, repetitive and deeply human. Staff were dealing with a patchwork of digital systems that didn’t always talk to each other, plus spreadsheets, paper files, and the occasional whiteboard that still controlled more than it should have. A record might sit in one hospital system, then vanish from view when a patient moved to another part of the service. A discharge note could exist in one place and be invisible in another. Tests got duplicated. Follow-up plans slipped. Someone, somewhere, spent too much time phoning around to reconstruct a patient’s story from scraps.
That’s where the federated part matters. The system was designed to let data stay where it is locally while still being queried and used across the network. Hospitals, regional planners and national teams could look at the same broad picture without forcing every organisation into a single database. In theory, that gives the NHS more control over local information than a centralised system would, while still fixing the age-old problem of fragmentation. It also means a tool built in one area can be reused somewhere else if the underlying setup’s shared. If a dashboard helps one region spot theatre delays or discharge bottlenecks, another region can adapt it without starting from zero. That kind of reuse’s one of the selling points, because no health service wants to pay twice for the same workaround.
The practical aim was blunt. Shorter waiting times, and fewer unnecessary overnight stays. Better use of operating theatres, where minutes lost to poor scheduling can snowball into cancelled procedures and longer backlogs. A planner looking at theatre lists, bed availability and patient flow in one place can make decisions that are less guesswork, more triage. In a system as stretched as the NHS, even a small improvement in scheduling can ripple outward pretty quickly. Not in a magical way. Just in the boring, expensive, staffing-and-timetable way that healthcare runs on.
Still, the pitch has always carried a catch. The same platform that promises smoother operations also asks hospitals and care boards to trust a shared data layer run through a private contractor with a very large reputation. That’s where the conversation starts to slide from software into power and politics, even if the original problem looked like pure admin. Data architecture sounds dry until it decides who sees what, when, and on whose terms.
For now, though, the core claim is simple enough. The NHS wanted a system that could stitch together information already scattered across wards, trusts and regional teams. Palantir’s job was to make that stitching less painful, less manual and less dependent on luck. Whether it does that well enough to satisfy everyone is a different argument, and it’s the one that’s now spilled far beyond procurement meetings.
Why Greater Manchester chose its own route
Greater Manchester didn’t wake up one morning and decide to be difficult for the sake of it. The region’s spent close to a decade building its own Analytics and Data Science Platform, or ADSP, and by the time Palantir’s NHS deal arrived, local officials already had a working system in place. In a May 2025 board meeting, they said the in-house setup was already ahead of the national platform in some areas, which is a fairly awkward thing to hear if you’re trying to sell replacement software to people who’ve already done the homework.
When a public health system asks people to share sensitive data, trust is not a side issue. It’s part of the infrastructure.
That point matters in Greater Manchester because the ADSP wasn’t cobbled together as a quick response to a procurement cycle. It grew through years of local use, with clinicians, analysts, and planners leaning on it for day-to-day work. The region argues that this history has done something the newer national federated data platform still has to earn: public confidence. In a place where health data can feel abstract until it lands on someone’s record, that confidence is doing real work.
There’s also a practical difference in what each setup can reach. Officials in the region say the ADSP can draw on primary care data that Palantir’s platform doesn’t currently have. That’s a big deal in plain English, because a system that can see only part of the patient picture is always going to make decisions with one hand tied behind its back. GP records, hospital data and local operational information don’t always sit neatly in the same place, and Greater Manchester’s case’s that its own platform has spent years pulling more of that into view.
The way the system’s built also gives the region more room to move. ADSP’s modular and stitched together from multiple technologies, which means it can swap out weaker parts without tearing the whole thing apart. That’s a lot less glamorous than the sales pitch for a single grand platform, but it may be more useful in practice. If one component starts to creak, the region can replace it. It can be retired, if a tool gets outpaced. There’s no need to wait for a giant vendor roadmap to catch up with local needs.
That flexibility has a political side too. Once a system starts handling sensitive health records, patients tend to ask a sensible question: who’s holding this, and why should we trust it? Greater Manchester’s answer, at least in part, is that trust has already been built through use, scrutiny and time. Public bodies love to talk about scale. Patients usually care more about whether the thing actually works and whether their data will be treated like private information rather than loose change.
A further wrinkle is that the region’s comparing a homegrown platform with one designed to sit across the NHS at national level. Those are not identical jobs. The local model was shaped around Greater Manchester’s own mix of services, data sources, and planning habits, which may explain why officials believe it can do some things the national setup still can’t. For them, saying no to Palantir’s been less about making a statement and more about refusing to throw out a system they believe already does part of the job.
That local confidence is what gives the argument its bite. If the platform had been flimsy, this would be a much easier story. Instead, Greater Manchester can point to years of use, a broader data picture and a setup it says can be adapted without starting from scratch. In a debate about UK health data, that makes the region less like a holdout and more like an inconveniently prepared adult in the room.
How a software contract became a Westminster fight
Once Greater Manchester drew a line, the argument stopped looking like a local procurement quarrel and started sounding like something much larger. In June, a cross-party group of MPs warned that the NHS’s growing reliance on Palantir had created an unacceptable single-vendor dependency. That phrase did a lot of work in a short space. It suggested not just awkward billing or bad optics, but exposure: if one company sits too close to the plumbing, the public sector can end up with very little room to move.
By July, Parliament had sharpened the criticism. A separate committee backed the idea of using the break clause in the NHS contract and looking for domestic alternatives. That pushed the debate past the usual “is the software any good?” question. The new question was more awkward: should a health system as large as the NHS be locked into one supplier for this long, especially when the government can still walk away if it chooses? For ministers, that’s not a tidy spreadsheet problem. It’s a judgment call with politics hanging off every line item.
In Whitehall, software stops being just software the moment it starts looking like a test of who gets to hold the keys.
Palantir’s critics did not stop at procurement risk. They tied the backlash to the company’s work in conflict zones and its contracts with U.S. immigration enforcement, which gave the NHS deal a moral charge that ordinary public sector technology purchases rarely attract. That matters because the platform was sold as a clinical and operational tool, yet opponents kept dragging the conversation back to the company behind it. If a supplier is already associated, fairly or not, with surveillance, military use, or border enforcement, then a health data contract is never going to read as neutral. It becomes an argument about what kind of state is being built, and who gets invited inside it.
Palantir hasn’t exactly helped itself in the UK. Its founders have a habit of speaking in a way that sounds calibrated to irritate committee rooms, and the company’s own manifesto has only deepened that effect. Lawmakers who might’ve treated the NHS contract as a boring bit of digital housekeeping instead found themselves reading material that felt ideological, combative and a bit too fond of drawing battle lines. For a firm selling itself into public services, that’s a curious way to win hearts and minds. Or, in this case, possibly none.
The result’s that the NHS deal now sits inside a broader argument about AI policy and public sector technology. Some NHS voices say the scrutiny’s drifted away from the actual system and turned into a political theatre piece. They argue that the real question is whether the platform works, whether it can pull data together, and whether it can help staff make better decisions without forcing another round of clunky local integrations. That view has a certain practical appeal. Hospitals don’t need Parliament to cosplay as a help desk.
Opponents answer that public bodies can’t hide behind technical language when the public’s uneasy about who holds the data and what else that company’s done. In their view, “it works” is never the end of the conversation. A public service has to justify its suppliers, not just its uptime.
Even outside Westminster, the mood has spread into the day-to-day machinery of local government. Similar questions about public-sector technology and AI oversight have shown up in committee papers and meeting documents, including a Rotherham council item on AI policy and an East Hertfordshire committee document on digital decision-making. That may sound dry, but it points to something real: this is no longer a niche NHS disagreement. It has become a live test of how comfortable public institutions are with handing core functions to a single private vendor.
If the break clause is used, what happens next?
the first problem won’t be philosophical, if ministers pull the early-exit lever. It’ll be practical. A lot of NHS kit still speaks in mismatched formats, and some of it barely speaks at all. Supporters of the Palantir platform warn that tearing up the contract without a ready substitute could leave hospitals leaning harder on spreadsheets, scanned PDFs, and, in the worst cases, actual paper files. That sounds quaint until you remember that a misplaced form can mean a missed referral, a delayed discharge, or a clinician hunting through three systems for one answer.
In health tech, the ugliest failure is often quiet: data that exists, but not where the people treating you can use it.
The case for keeping the deal’s blunt. Backers say there’s no alternative sitting on the shelf that can pull together information across trusts, integrated care boards and national bodies at the same scale right now. They point to the awkward reality of NHS tech procurement, where patchwork systems have often been built locally, purchased separately and then expected to behave like a single service. That hope hasn’t always aged well. A cleaner replacement sounds lovely in theory. In practice, someone has to build it, test it, pay for it, and get hundreds of organisations to use it without breaking their own workflows.
Yet the political pressure hasn’t gone away, because the argument is no longer about software alone. It’s about data trust. If clinicians think the platform’s clunky, or if patients think their records are being fed into a system they don’t trust, usage drops fast. And once staff start working around a system instead of through it, the whole thing becomes a very expensive filing cabinet.
That’s where the debate gets awkward for ministers. They’re weighing patient care against procurement politics, while a separate anxiety hums in the background: foreign technology use. Palantir isn’t some neutral spreadsheet vendor hiding in the corner. It’s a powerful outside supplier with a lot of control over a setup the NHS wants to use everywhere. For critics, that’s reason enough to keep the brakes on. It’s exactly why the contract should be kept in place until a serious rival exists, for supporters.
Either way, the government’s choice will tell us more than who won a contract fight. It’ll show what sort of digital state the NHS thinks it can build, and how much public consent it needs before it asks patients and staff to hand over even more of their data. The NHS is buying software. It’s also deciding who gets to shape the operating system of public health, and that’s the part of this UK tech news story that won’t fit neatly into a procurement spreadsheet.
