The Matt Clifford Anthropic role has forced a swift reversal: within days of announcing he would keep his chairmanship of the UK’s Advanced Research and Invention Agency (ARIA) alongside joining the Claude-maker as managing director of international affairs, Clifford said on 7 September he would stand down from ARIA after all, citing the need to ensure his new position ‘doesn’t become a distraction.’
He will remain as chair until 6 November while the Department of Business, Innovation, Science and Technology recruits his replacement.
The Matt Clifford Anthropic Role That ARIA Knew Nothing About
The detail that sharpened criticism most was revealed at a House of Lords Science and Technology Committee session on 8 September: Research Professional News reports that ARIA’s chief executive, Kathleen Fisher, told peers the agency ‘was not consulted’ about Clifford taking the Anthropic position, adding it was ‘not Aria’s decision’ and that ‘decisions were conversations between Bist and Matt.’ Conservative peer Kulveer Ranger said the appointment had raised ‘a lot of concern in the industry.’
That is a damaging admission. ARIA is a publicly funded body whose chair is appointed by the government. The idea that the chair could take a senior commercial role at an AI company without anyone at the agency being consulted is, to put it plainly, a governance failure, regardless of which direction you assign the blame.
Chi Onwurah, chair of the House of Commons science, innovation and technology committee, had called Clifford’s original plan to stay on at ARIA a ‘clear conflict of interest’ and demanded the AI minister, Kanishka Narayan, set out boundaries. Anthropic had argued that Clifford would step back from any ARIA matter touching the company, and that its commercial relationship with the agency was handled by a separate team. That defence did not survive contact with Kathleen Fisher’s testimony.
A Government-Shaped CV, Now in Private-Sector Hands
The broader unease is about what Clifford carries through the revolving door. He drafted the government’s AI action plan, which backed AI growth zones and the development of UK-owned models. That work led directly to the establishment of a £500 million Sovereign AI Fund, as Chemistry World reports. He was also the UK’s representative to the 2023 AI Safety Summit and helped create what became the AI Security Institute.
Tom Brake, chief executive of campaign group Unlock Democracy and a Liberal Democrat MP from 1997 to 2019, put it bluntly: ‘Taking extensive knowledge obtained within government, about a policy as sensitive as AI, to a private sector company shortly after leaving government gives that company a huge advantage.’
My read is that Brake is correct, and that Anthropic knows it. The company did not hire Clifford for his general charm. His job title, managing director of international affairs, means he will lead Anthropic’s engagement with governments including those of the UK, Europe, Asia-Pacific, and India. The knowledge of which levers move which officials, built up at Downing Street and across Whitehall, is the product. That is not an accusation. It is simply what this kind of hire is for.
Clifford co-founded Entrepreneurs First, which he describes on his personal website as a ‘$1.3 billion holding company.’ ARIA, where he has served since 2022, he describes as a ‘£1.4 billion agency for funding breakthrough R&D, loosely based on DARPA.’ The scale of public resources those figures represent is worth keeping in mind.
Anthropic’s IPO Ambitions Add Another Layer
Clifford joins at a pivotal moment for the company. On 1 June 2026, Anthropic confirmed it had confidentially submitted a draft registration statement on Form S-1 to the US Securities and Exchange Commission (SEC) for a proposed IPO, with no share count or price yet set. Anthropic raised $65 billion in a Series H round in May 2026 at a post-money valuation of $965 billion, according to Direxion, which cited Anthropic’s own announcement. The $2 trillion valuation figure that has circulated widely is an investor estimate; Anthropic has not confirmed any IPO valuation, date, or offering size. Direxion reports the company was targeting a public listing in late September or early October 2026.
A company on the cusp of what could be the largest IPO in history, actively courting favourable regulatory conditions across multiple jurisdictions, hiring the person who wrote the UK government’s AI policy: the incentive structure here is transparent. That does not make it improper in itself. But it does mean the government needs to be considerably more clear-eyed about what ‘safeguards’ actually mean in practice, rather than accepting assurances that separate teams handle the relevant paperwork.
Clifford’s stand-down from ARIA is the minimum the situation required. Whether it is sufficient depends on what happens next: specifically, whether any future ARIA decisions, procurement or otherwise, are ever challenged as having been shaped by knowledge Clifford carried with him. The replacement of his successor, and the terms under which they are appointed, will be the first real test of whether the government has actually learned anything here.
