For a programme sold to Parliament as a near five-to-one return on investment, the NHS Federated Data Platform has spent much of the past year quietly shedding the numbers that justified it. The benefits once projected for the platform were overestimated by £1.7 billion, a correction that lands against separate figures from the National Infrastructure and Service Transformation Authority showing the programme's whole-life cost has climbed to £1.1 billion while its forecast financial benefit has fallen to £808 million.

The FDP was conceived to solve a basic and long-standing NHS problem, that operational data on beds, waiting lists, discharge readiness and theatre capacity sits in disconnected systems across 42 integrated care boards and more than 200 trusts. Built on Palantir's Foundry software and awarded under a contract worth up to £330 million, the platform was meant to let staff see that information in one place and act on it faster. Ministers defending the deal have leaned heavily on the scale of the promised return. As recently as April, a health minister told the Commons the FDP could deliver up to £2.4 billion of benefit over time, citing independent estimates that went well beyond NISTA's own business case.

That gap between political framing and official appraisal is now the story. NISTA's own figures have moved only modestly over the past year, from a benefit forecast of £777 million and a cost of £1.042 billion in May to £808 million and £1.1 billion now, changes the authority attributes to increased delivery at scale and updated assumptions about adoption and usage. It is the wider, more ambitious claims made in Parliament and in NHS England's public messaging that the £1.7 billion correction cuts down. For a programme repeatedly held up as proof that centralised NHS technology can pay for itself, the retreat from £2.4 billion of projected benefit to under £1 billion is not a rounding error.

It also lands amid a broader credibility problem NHS England did not need. In July the Office for Statistics Regulation ordered the organisation to strengthen its caveats after concerns that its published FDP statistics conflated correlation with causation. Freedom of information requests from the campaign group Foxglove found that nearly a third of trusts using the platform's inpatient care coordination tool were performing fewer procedures than before adoption, with a single trust accounting for the bulk of the headline improvement in outpatient waiting lists. Separate analysis reported found no measurable improvement in delayed discharges among trusts using the OPTICA tool, a finding Palantir disputes. Both the Science, Innovation and Technology Committee and the Health and Social Care Committee have urged ministers to use the break clause available in the Palantir contract from 2027.

None of this means the FDP has failed outright. NISTA continues to rate the programme green for delivery, and NHS England maintains that benefits will grow as adoption widens beyond the minority of trusts currently using its core products. But the revised numbers strip away the scale of return that made the FDP politically defensible in the first place. Ministers weighing the break clause now face a starker calculation than the one presented when the contract was signed, one that has to account not only for what replacing the platform would cost, but for how much of the case originally made for keeping it has already been quietly withdrawn.