The UK and the EU are lobbying to gain exemptions from the U.S. government’s recent export control directive against Anthropic’s Fable 5 and Mythos 5, which prompted the AI lab to disable access to both models for all users worldwide days after release.
Reports indicate that the Trump administration will not heed British pleas to restore UK access to the two frontier models, while EU and American officials have discussed the possibility of establishing a ‘trusted partner’ scheme, without any commitments yet from Washington.
The ongoing saga has brought the issue of ‘sovereign AI’ sharply into focus for the UK and Europe, underscoring the need for Britain and the continent to develop their own AI infrastructure and models.
However, current activity in the European AI sector has shown that truly sovereign AI may be difficult, with recent deals, announcements and public spending highlighting how the UK and Europe continue to rely heavily on American companies, despite the prevailing rhetoric.
The ‘debilitating sovereignty tax’
Participants in the UK tech industry aren’t especially surprised that the U.S. government has placed restrictions on Anthropic’s latest models, with Mark Boost — the CEO at London-based cloud provider Civo — telling TechRepublic that advocates for sovereign AI have been warning of such a scenario for a while.
“Relying on foreign-controlled tech means access can be revoked at any time, especially given the rapid pace of AI’s development,” he says.
Given the rapidity of development, Boost suspects that denial of access could occur more frequently in the future, with administrations increasingly classifying AI models and tools as “matters of national security” that warrant restricted access.
Such restrictions will cause major disruption to British and European organisations, Boost warns, particularly if they’re locked into ecosystems founded on American cloud platforms.
He adds, “This is especially true for UK companies already locked into these ecosystems, who face a debilitating sovereignty tax: a compounding penalty of aggressive technical debt, interoperability barriers, and punitive egress fees that make exiting these platforms extremely challenging for certain workloads.”
On a practical level, UK and European companies that had begun integrating Fable 5 into their workflows are likely to have experienced delays and efficiencies, having to roll back systems to earlier models.
“The sudden access loss to Fable 5 could have exposed vulnerabilities in the UK’s current technology strategy,” says Boost. “With the US government switching the tool off 72 hours after launch, organisations with single-model dependencies may have experienced operational disruption, had they planned to integrate it into production workflows.”
Some business owners in the UK have already posted the disruption they’ve suffered as a result of the removal, with one suggesting that the episode has taught him not to rely too heavily on AI as a founder, and to have contingency plans.
Access could be reinstated
More broadly, Boost affirms that the whole incident should be regarded as “the biggest wake-up call yet for UK organisations” that they need to prioritise tech sovereignty.
“When British firms are disrupted in such a way or forced to scramble for alternatives, UK tech growth is stifled,” he says.
The removal of Fable 5 and Mythos 5 is not only disruptive for UK and European firms in the shorter term, but it also raises the troubling possibility that they could be permanently deprived access to the most advanced AI models.
“This is one of the clearest reasons why the British government needs to invest in domestic alternatives and why UK companies should prioritise local tech providers, which often boast similar capabilities,” Boost adds.
The Software & Information Industry Association (SIIA) — which represents Amazon, Apple and Google — has written an open letter opposing the “unprecedented action” against Antropic, while sources close to discussions between the UK and the U.S. have indicated that the Trump administration may reach a deal with tech firms to restore access globally.
However, the fact remains that Fable 5 and Mythos 5 are still inaccessible, while the U.S. government continues to flirt with the idea of blacklisting DeepSeek and other Chinese AI companies, underlining the possibility that national security concerns may continue to intervene in the industry.
More must-read AI coverage
- SS&C Intralinks DealCentre AI vs. Datasite: Which platform is built for the future of dealmaking?
- SS&C Intralinks FundCentre AI vs. Juniper Square: Which platform better supports modern private markets fund managers?
- Why Data, Not Models, Determines AI Success
- The Rise of the AI-Native Factory: How Physical AI Is Transforming Manufacturing
Inconsistencies amid desire for tech sovereignty
As such, Boost is one of numerous CEOs calling for more action to be taken in terms of sovereign AI, in case the UK and Europe is shut out of growth.
“Organisations can no longer ignore that data sovereignty is a live infrastructure risk rather than the compliance box-ticking exercise it may have been considered previously,” he says. “This must catalyse UK tech leaders and policymakers to aggressively invest in and champion home-grown alternatives that will guarantee the country’s future resilience.”
Such affirmations parallel recent comments in Europe, where Publicis CEO Maurice Lévy has called on France and Germany to take the lead in establishing a €100bn fund that would support the development of European AI and infrastructure alternatives.
There’s little doubt that authorities in the UK and Europe were already aware of the value of sovereign tech, with the UK announcing a £500 million Sovereign AI Unit in April and a £1.1 billion AI Hardware Plan earlier in June, and with the EU announcing its own comprehensive Tech Sovereignty Package several weeks ago.
Many industry participants strongly believe that more should be done, however, something reflected in Civo’s recent report on tech sovereignty.
“72% of UK IT leaders agree that UK innovation requires domestic infrastructure, and 43% explicitly require AI workloads to be hosted under UK law,” affirms Boost, explaining the report’s findings. “Furthermore, 90% of these leaders are calling for a more proactive stance to back British tech.”
Boost notes that 20 MPs in the UK Parliament are supporting an amendment to the Cyber Security and Resilience (Network and Information Systems) Bill that would oblige the UK Government to publish a digital sovereignty strategy within 12 months, a sign that support for sovereignty is growing.
However, he notes that glaring contradictions and inconsistencies remain, as public and private entities continue using U.S. tech companies for key infrastructure and services.
He says, “While the government talks about sovereignty, it continues to funnel 90% of its cloud spend to US hyperscalers, including recent multi-year £400m+ deals with AWS for HMRC and Google Cloud for the MOD.”
Some of the UK’s most promising tech companies are doing something similar, including AI lab Ineffable Intelligence, which has only just signed an agreement to use Google Cloud’s infrastructure to develop its self-learning AI models.
In Europe, Thales has recently signed a comprable partnership with Google Cloud, with which it will work on developing a “sovereign cloud” for Germany.
Given such preference for American incumbents over local providers, Boost urges the UK Government to support and help develop national champions, so that organisations have credible British (or European) alternatives.
“We need the government to put its money where its mouth is when it comes to its own tech-buying strategy,” he concludes. “Crucially, we also need even more tangible state-level collaboration to actively back UK-rooted cloud alternatives.”