Stop underestimating British tech

Tech sovereignty is all the rage. Parliament’s DSIT committee this week warned policy makers of of dependencies which would leave the UK unable to function in either in peacetime or conflict. With a capricious Trump administration eyeing Greenland and blocking access to the latest AI models, perhaps it isn’t entirely fatuous. The prospect of our forces being shut down en route to defend the Falkland Islands does not require much imagination. Those remote kill switches exist.

But what does it mean? The House of Commons Library provides us with a handy definition. Digital sovereignty is “the agency and capacity of any organisation to make intelligent, informed choices to shape its digital future by design.” 

So it must be intelligent and informed, and this is where the trouble really starts. Whitehall and the right’s policy intelligentsia have become so incapable of thinking of Britain as a technology power, and so deferential to the idea that only Californians can do digital properly, I propose coining a word for it. Forget Anglo-Futurism. This is Anglo-Wurzelism.

For example, take the ongoing story of Palantir and the NHS.

Palantir is a defence logistics and data analytics company that has leaned heavily into the culture wars. Andy Burnham’s administration will reportedly cancel a £330m NHS contract granted to a consortium of which Palantir is a member, exercising its break clause. 

(I should confess that Palantir’s CEO Alex Karp is by some distance my favourite tech CEO, as he is so thoughtful and so very legible, and the firm provided access to me for a recent Sunday Telegraph essay on the firm).

For the online right, cancelling Palantir would be akin to state sanctioned murder. Palantir’s FDP (Federated Data Platform) is revolutionising a sclerotic and dysfunctional state health monopoly which would rather use faxes (three machines are actually still in use) than improve productivity or the welfare of its patients, they argue. From there, the logic takes us to removing Palantir kills people.

“It’s hard to avoid the reality that the opponents of Palantir would prefer that patients die rather than use the best available piece of software,” wrote Matthews Lesh of Freshwater Strategy, a public affairs company. Stephen Pollard writing at CapX, writes of “cancer victims and patients waiting for diagnosis or operations” as if it was Palantir’s software doing the surgery. We must choose Palantir, or more people will die of sepsis in hospitals, suggested GBNews’ Tom Harwood. More on that in a moment, but look how simple and neat it appears.

Consider this, then.

“Why is the government and the NHS so lacking in confidence in UK tech?,” asks Suranga Chandratillake of Balderton Capital, questioning why Palantir is so favoured. Saul Klein, another prominent UK VC, wondered why Faculty AI and Accenture could not do the job. Procurement is unpatriotic. What do they know that our policy intelligentsia do not?

The key to understanding this story is some domain knowledge and context. The “intelligent, informed choices” required by digital sovereignty require knowing whether Palantir has unique capabilities, some magic proprietary sauce, that others don’t possess, and knowing what is required to build operational clinical data systems. If Palantir tops the bill, then reason suggests it should be chosen as a contractor; this then should be a straightforward procurement decision. But what if it’s more complicated than that?

That’s the case made by Chandratillake and Klein, and many in the UK tech sector too, including our own data teams and scientists. They point out that far from importing some magic into the NHS, FDP is composed of open industry standard parts, such as Apache Spark and Databricks, and the logic to bind it together can be written by any competent development team. In fact some NHS organisations, such as the Greater Manchester Integrated Care Board, have done just this, and consider their own version to be far superior to the centrally-mandated Palantir one. The Board’s tech gurus chose its own data pipeline and data visualisation specialists, and the University of Manchester’s eLab. 

For them, the enthusiasm for Palantir in Westminster is a great betrayal. We have many examples of how tech companies bamboozle civil servants who make poor procurement decisions, and write poor contracts as a result. Motorola could extract over £1.1 billion in “supernormal profits” from the Airwave radio contract for police and emergency services, ESN. 

The rosy data has been questioned, too: a third of Palantir trusts are performing fewer operations. The fear is the NHS becomes dependent on Palantir, just as our emergency services became dependent on Motorola’s radio network.

For my Palantir long read, I was given access to dozens of applications and was surprised how basic many were. The improvements of applying technology are real, but incredibly, it’s new digital tools and methods that are being introduced for the first time. And guess what, it can work.

Palantir’s success at detecting sepsis early, cited by Harwood, is laudable. But this is not novel. The data is based on results from one hospital, Tampa General Hospital in Florida, a Palantir flagship, and similar or better results have been recorded in other hospital trials: Duke’s down 27 per cent, John Hopkins Trews, with risk reduction down 18.7 per cent, and absolute mortality down 4.5pc. 

One spectacular gamble against Palantir as a vanguard technology company has been made by Michael Burry, the investor immortalised in the movie The Big Short. Burry’s short position is based on the bet that it is really a labour-intensive consultancy, a Deloitte with some culture wars sizzle, rather than a Google or an Anthropic. That there’s no secret sauce. If he’s correct, then it seems an even stranger hill for the online right to die on. 

For my Telegraph essay, I spent much time exploring this topic, of how much of a lock-in Palantir really holds, and whether it is unique. I gave the last word to Tom Bartlett, the former NHS executive who worked on the Federated Data Platform, and who now as an independent is one of Palantir’s most public supporters. Bartlett sees the case, but cautions that it may take seven to ten years, even given a fair wind. There’s a price to being locked into Palantir, then, and a price to evading the lock-in. Starmer has been too anxious to accept the Silicon Valley investment, and not see the cost, he thinks. 

So where does this leave us in assessing issues of digital sovereignty?

One lesson from the Palantir saga is that Westminster’s cadres of policy experts are oblivious to what’s happening outside the SW1 bubble, and specifically the subtleties of technology development and procurement. As Westminster right-leaning think-tanks gather on summer lawns for the summer, support for Palantir’s FDP will be unanimous. 

British technological prowess is being killed by the prejudice and ignorance of SW1

This is a gulf that once would have been filled by high quality research — the Conservative’s Research Department was once formidable. But the right has lost that analytical capacity. And after so many years repeating the mantras of globalisation, it has also lost its capacity to think dispassionately in terms of our own tech sector. Where papers from the think tanks discuss technology, they invariably side with the US tech giants, not the British companies fighting for market share against them. 

Writing about David Potter’s trailblazing computer company Psion recently — Potter died earlier this month — I was struck by the fearlessness of the engineers at firms like Acorn and Psion in the 1980s and early 1990s. They didn’t wait for Japan or the USA to develop something, or determine the parameters of what was possible. Today, American capital determines the winners and losers, and we tug our forelocks in deference. We can’t even begin to have a conversation about “digital sovereignty” if we can’t assess what we do here well. British technological prowess is being killed by the prejudice and ignorance of SW1, and the online right has become the undertaker.