Here’s a scenario that sounds like a procurement footnote but is really a much bigger argument in disguise.
Earlier this month, Sadiq Khan blocked a £50m deal that would have seen the Metropolitan Police use Palantir’s AI to automate intelligence analysis in criminal investigations. The mayor’s office didn’t frame it as a values judgement. Officially, it was about process: a “clear and serious breach” of procurement rules, and the awkward fact that the Met had apparently looked seriously at exactly one supplier before deciding that supplier was the answer.
Scotland Yard, for its part, called the block “disappointing” and went further, warning that without new technology it’ll be forced into choices that “cannot avoid reducing officer numbers.” That’s a heavy thing to say out loud. The Met is already staring down a £125m shortfall and around 1,150 job cuts. So the subtext is hard to miss: block our tech and you’re effectively cutting coppers.
So we’ve got two stories sitting on top of each other. One is a dry tale about procurement governance. The other is a much spicier one about whether the politics of who you’re buying from should ever override the question of what the kit can do. Let’s pull them apart.
The bit nobody’s saying out loud (but everyone’s thinking)
You can’t talk about this deal without talking about Palantir itself. The company was co-founded by Peter Thiel, it works with the Israeli military and with US immigration enforcement, and its CEO recently published a manifesto that one MP cheerfully described as the ramblings of a supervillain. Khan has previously been explicit that Londoners want public money going to firms that “share the values of our city.”
So even if the official reason for the block is procurement, the political weather around it is impossible to ignore. And that brings us to the real question:
Should a company’s politics, ownership, or ethics influence whether the state buys its technology?
I genuinely don’t think this has a clean answer, and I’d be suspicious of anyone who tells you it does.
The case for keeping politics in the room is pretty strong. This is public money. The police are an arm of the state, and the public has a legitimate stake in not being made complicit in things they find objectionable. Procurement has always carried ethical conditions; we already exclude suppliers over labour practices, sanctions, environmental records. “We don’t like your founder’s tweets” is a sillier version of that,
sure, but “we’re uneasy about a vendor whose business model is built around surveillance and who serves regimes we have concerns about” is not a frivolous position. Democratic accountability is supposed to reach into spending decisions, not stop politely at the door.
But the case against is just as real, and it’s the one I’d want a tech-minded reader to sit with. Once “values” becomes a valid reason to block a contract, “values” becomes a wonderfully flexible tool. It can be a genuine ethical stand, or it can be political theatre, or it can be soft protectionism dressed up as principle, and from the outside those three look almost identical. Tech bought on vibes rather than capability tends to be bad tech. And there’s something uncomfortable about a system where the same software is fine for the NHS and the Ministry of Defence (both already Palantir customers, to the tune of £330m and £240m) but not for London’s police, on grounds that have nothing to do with whether it works.
Here’s the part I find genuinely clever about how this played out: the block didn’t have to invoke values at all. It rode in on procurement. And maybe that’s the honest answer to the whole question. Politics shouldn’t be the stated reason to reject technology, but good governance often gives you a perfectly legitimate reason anyway. If you only ever seriously considered one vendor for a £25m-a-year contract, you’ve got a real problem regardless of whose logo is on the invoice.
But what does this actually do to policing on the street?
This is where I want to slow down, because both sides have an incentive to overstate things.
The Met’s framing is essentially: this tech lets us do more with less, so blocking it means cutting officers. I’d treat that with a raised eyebrow. “Give us this or we lose frontline staff” is one of the oldest moves in the public-sector budget playbook, and the causal chain from “intelligence-analysis automation” to “more uniforms on the high street” is a lot longer and wobblier than that soundbite suggests. Automating analyst workflows frees up analyst time. Whether that converts into visible, street-level presence, the thing the public actually feels, is an assumption rather than a guarantee.
And yet. The under-resourcing is real, not rhetorical. The shortfall is real, the cuts are real, and the appetite to modernise is a rational response to being asked to police a huge city on a shrinking budget. The promise of tools like this is precisely that they let a stretched force triage faster, connect intelligence it would otherwise miss, and spend fewer human hours on grunt work. If that promise even half-lands, blocking it has a genuine cost, and that cost is paid not by the mayor or the commissioner but by the cases that move slower and the things that get missed.
Then there’s the long game, which Chi Onwurah put her finger on: vendor lock-in. This is the unglamorous concern that I think actually matters most. Khan’s point about “loss leaders” (a supplier giving you a sweet deal up front so you become dependent, then holding the pricing whip later) is the oldest trick in enterprise software, and the public sector falls for it constantly. A dependency on a single large US provider for something as sensitive as police intelligence isn’t just a procurement risk; it’s an operational risk that could squeeze policing budgets harder, five years from now, than any single contract being delayed today.
So the honest read on street-level impact is probably this: the block doesn’t make the resourcing problem better, and the Met’s “we’ll cut officers” warning is at least partly leverage. But the thing being protected against (lock-in, dependence, a contract waved through without competition) is a slower-burning threat to policing that’s easy to dismiss precisely because it doesn’t bleed in the next budget cycle.
Where that leaves us
The need doesn’t go away because the deal did. The Met still has to modernise, still has to do it with less, and now has to run the procurement properly, which costs time, and the resourcing gap stays open while that runs. Nobody really “wins” here in the short term.
But maybe the useful takeaway for anyone thinking about technology adoption, in policing or anywhere else, is that how you buy is not separate from what you’re buying. The most defensible reason to block a deal is rarely “we don’t like them.” It’s “we didn’t run a real process, we didn’t test the market, and we were about to make ourselves dependent on a single supplier for something we can’t afford to get wrong.”
Should politics play a part in tech adoption? Honestly, it already does, and pretending otherwise is its own kind of fiction. The better question is whether we’re being upfront about it. Procurement discipline and democratic values aren’t really rivals here. When they point in the same direction, you should probably listen. When they don’t, that’s exactly the moment to ask which one is doing the actual arguing.
What do you think? Is “shares our values” a fair test for public technology, or a back door for politics to override capability? I’d genuinely like to know where you land.
