On June 16, SpaceX's stock price surged to such an extent that Elon Musk took advantage of the opportunity to purchase a $60 billion artificial intelligence coding venture in an all-share deal, essentially paying for it with value that had emerged that same afternoon. Three weeks later, the stock had fallen by more than a third from its peak, dragged down in part by concerns over a Starlink pricing dispute in Memphis that had nothing to do with artificial intelligence at all. The company that briefly made Musk the world's first trillionaire, on the strength of investors treating it as an AI story rather than an aerospace and telecommunications one, is now waiting on its first public earnings report to establish whether the story holds up.

That correction, uncomfortable as it has been for retail investors who bought in the opening days of trading, is the market doing what markets do. Price finds out, eventually, whether a valuation reflects a business or a narrative about that business. SpaceX's shareholders are being forced through that discovery in real time, in full public view, with a lock-up period due to end around the same time as those earnings, potentially adding fresh supply and fresh scrutiny simultaneously.

The NHS has no analogous process, and that absence is worth examining more carefully than it usually gets.

Over the past two years, the health service has committed itself to a series of technology programmes, the Federated Data Platform chief among them, that were sold on projected efficiency gains, clinical time freed up, and outcomes improved, well ahead of any evidence that those gains would materialise at the scale claimed. Ambient voice technology is following a similar trajectory, moving into wards and consultation rooms faster than the regulatory apparatus, including the MHRA, has been able to establish clear governance around its use. In each case, the pitch resembles the one investors bought into on SpaceX's opening day: a business case built substantially on the promise of what artificial intelligence will eventually do, rather than a verified account of what it currently does.

What SpaceX shows is that even a narrative this compelling, backed by one of the most closely watched executives in the world, cannot indefinitely outrun scrutiny once real numbers are due. Investors who did not wait for the earnings report to test the claims are now underwater. But investors have the option of exit, of repricing, of simply refusing to buy at a given valuation. NHS trusts signing multi-year contracts, integrated care boards adopting platforms under central direction, and clinicians expected to work with tools procured above their heads do not have that option in any meaningful sense. Once a contract is signed and a system embedded, the mechanism for testing whether the technology delivers what was promised is slow, political, and easily deferred, in a way that a share price never can be.

This matters most acutely now because Sir Jim Mackey's accountability agenda and the abolition of NHS England both depend on centralised decision-making moving faster than the evidence base that ought to justify it. A market forces reckoning through price. A centralised health system forces reckoning, if at all, through inquiry, audit, or scandal, mechanisms that arrive years after the money has been committed. NHS leaders evaluating the next wave of AI procurement would do well to ask a version of the question SpaceX's earnings report is about to answer for its shareholders: what proportion of this valuation, or this contract's value, rests on delivered performance, and what proportion rests on a story about the future that has not yet been tested against anything at all.