But there are deeper issues with the FDP. It runs the risk of stealing oxygen – and funding – from other critical work already underway to help the NHS join up its patient data for good. For example, openSafely, a flagship national data platform for health research, was developed by Ben Goldacre and a team at Oxford and was used for vital Covid research. It’s completely open source, safe and lights a way forward for trustworthy health research. It also costs a fraction of what Palantir does.

What’s more, pushing so much access and control to the centre may not make sense. For some issues – vaccination, workforce planning – there is a clear case for a national solution. But ultimately, most care is delivered locally and planned regionally. There are already places, such as London, that have pioneered solutions to pool patient data to plan care better – at a fraction of the FDP’s cost. It is far from clear how this will interact with the FDP, or whether it can survive the new system.

Other competitors – like a UK consortium of universities and open-source firms that are apparently bidding for the deal – would have loved a fair crack at the FDP contract. But let’s be honest: they probably haven’t got a snowball’s chance at beating Palantir’s incumbent advantage, won through a mixture of insider influence and watermelon cocktail lobbying.

Once Palantir’s in, it will be hard to get it out. The technical architecture is proprietary – and other government agencies have struggled to get off Palantir when they’ve tried. Having a single supplier to help you join up data and analyse it also risks creating a dangerous private monopoly over vital NHS infrastructure.

Indeed, if you take Palantir chief executive Alex Karp at his word, that’s the plan. “We are working towards a future where all large institutions in the United States and its allies abroad are running significant segments of their operations, if not their operations as a whole, on Palantir,” he wrote. “Most other companies are targeting small segments of the market. We see and intend to capture the whole.” That reads like an express statement of an intention to seek monopoly power.

It’s also clear they’re in it to profit. Their chief technology officer, Shyam Shankar, recently wrote: “The problem with defen[c]e contracting is not the popular narrative that contractors make too much money. It is actually that they make too little money… Innovators will need outsized profits to motivate progress.” Monopoly and profiteering may be good for Palantir’s share price, but they sit uncomfortably with the ethos of a public health service.

Joining up the NHS’s disparate health data systems better will present stiff challenges, and the NHS will face trade-offs – buying in consultants may be easier in the short term, for example, but may prove more expensive in the long run. But at the moment the government is stonewalling legal letters asking even basic questions about the FDP. And they are also creating facts on the ground that could be seen to favour Palantir. The legal basis for all of this, now that the pandemic’s suspension of protections for patient data has lapsed, is unclear.

People care deeply about how their health data is used. We go to the doctor to share our worries, our fears, and our pain – and if we don’t trust that conversation to be private, we may not go at all. People want to feel safe to contribute their health data for the good of the NHS – but when the government runs out ahead of patient trust, overhauling patient data systems without explaining what it wants to do, who will see the data, and what safeguards there are, people baulk. In 2021 more than a million people in a month opted out of sharing their health data because they didn’t trust the government’s last plans to pool their GP records. The history of the NHS is a boneyard of such schemes: massive, expensive white elephants that all failed because the government didn’t take the time to get the governance or consent right.

It is past time for the government to learn from these mistakes. We can build a better future for our patient data – if we take the time to design carefully, honouring patient choice and thinking about what system will serve the NHS for the long haul. Anything less is likely to fail and set the cause of progress back another five years.