US Big Tech tightened its stranglehold over UK politics last week, as Matt Clifford, the face of the UK’s ‘Sovereign’ AI strategy, joined Anthropic as ‘Managing Director of International Affairs’.
With Clifford, the US AI major now boasts an arsenal that includes not just former UK Prime Minister Rishi Sunak, but the man Sunak hired to shape how the UK develops and invests in British companies and startups building the models, compute, and infrastructure that will shape and power the UK’s approach to artificial intelligence.
UK startups – including ‘sovereign’ AI companies – will now be competing for procurement contracts against a lobbyist that has the mobile phone number of every single senior politician and official who matters, access he got because he was meant to be their champion in the face of the onslaught of the handful of US companies —particularly Anthropic and its competitors OpenAI and Google
In a statement, Clifford said he was joining Anthropic because they “took seriously” the belief that “people everywhere want real agency over how [AI] is used”. These words ring hollow. In reality, those seeking to give the UK agency by growing their companies here – and thereby sharing in the value of this technological transformation – will now find it harder to do so.
Clifford is only the latest to be hoovered up in the great brain drain – of leading figures from across academia, politics, and business going to work at the frontier AI labs. The impact of this ‘soft’ yet insidious corruption of public life is only set to grow.
Big tech uses the same playbook on both sides of the Atlantic; Dean Ball, who previously advised the US administration on AI policy, recently went in-house to do a similar job for OpenAI. Ball, a prominent public commentator on AI matters, was then shocked to discover people attributed his online musings to his employer. But that’s the bargain you have made, guys: you become rich beyond most people’s wildest dreams, but we don’t have to take you seriously as independent actors anymore.
That’s why it was a shock to read that Clifford initially intended to stay on as Chair of the Advanced Research and Innovation Agency. ARIA, which was set up by Dominic Cummings to fund high-risk, high-reward British science including AI, to hand out £800m of taxpayers’ money to promising innovations. There were already concerns about the institution’s closeness to Big Tech: a Democracy for Sale investigation earlier this year found an eighth of its research funding over two years had gone to US tech firms and venture capital groups.
It was always completely untenable for the agency to be chaired by an Anthropic executive – with access to confidential information on the value and activities of the very UK companies his employer might look to buy. Acquisitions of smaller companies are increasingly a key business strategy for the firm – Anthropic bought at least four startups in 2026 alone: Bun, Stainless, Vercept, and, for a reported $400m, the biotech firm Coefficient Bio.
With Clifford having initially offered a narrow recusal from Anthropic-related matters, he then said in a statement on X on Monday that he had “decided to step down to ensure my new role at Anthropic doesn’t become a distraction”. He has said he will stay on as Chair for six months on an interim basis, during which time he will have access to the business-sensitive information described here. It is noteworthy that this announcement came after a significant backlash at his apparent plan to continue in the role. In a reply to me posted on X late on Monday, he insisted no one had asked him to resign.
To understand how a sovereign gamekeeper could so readily turn international poacher, it’s important to situate this as part of a wider failure of the British tech sector and policymaking worlds.
For many UK tech players, now typified by Clifford, their understanding of British sovereignty sees the UK cast as little more than an outpost in the US empire. The reaction from the tech sector was telling in this regard. Dom Hallas, Executive Director of the Startup Coalition, the UK’s trade body for tech startups, commented on X: “Matt [Clifford] ensuring that Anthropic continues their run as the most British-coded major lab 😂 V[ery] exciting stuff!”
Generously, some of the reaction is likely a recognition of (what they at the time believed was) Clifford’s continued status and influence as a funder through ARIA. But it’s hard to avoid the sense that all the congratulatory social media posts betray an emptiness at the heart of the appeals to sovereignty. If the call came from Silicon Valley and the number was right, many of these people would be on the first plane out of Heathrow just like Clifford.
The yank-ification of British tech also shows up in policy. Earlier this year we at The New Contract found, through Freedom of Information requests, that more than a third of ‘Sovereign AI’ funding has gone to companies ultimately owned in the US. As it stands, there is nothing to prevent any of them upping sticks – or being bought by the likes of Anthropic. This is not a serious industrial strategy.
So what can be done?
Despite a new system in place following the abolition of the Advisory Committee on Business Appointments (ACOBA), the government should introduce a statutory cooling-off period for anyone moving from policymaking into the very companies they were meant to regulate and design policy around.
Public money must come with conditions, including continued UK ownership, and a guarantee that the IP and data it funds stay in Britain. If a strategic company is targeted for acquisition, the government should be willing to use Competition and Markets Authority intervention and/or ministerial powers to block this.
Andy Burnham assumed the premiership amid reports he was planning to ‘revamp’ AI policy; the abolition of the Science and Technology department was rumoured to be because he rightly saw Starmer’s administration as too close to Big Tech. Clifford’s resignation must be the catalyst for this more significant reset.