First published on the Forbes Technology Council, 27 May 2026
Enterprises should not wait. The faster route to value is to work with the global AI providers that can already deliver at scale, under a governance framework that controls data access, auditability, accountability and exit routes. Where a provider is headquartered matters far less than how the enterprise governs the relationship. This is the argument our co-founder Tarek Nseir made in Forbes, and it is the practical position we take with clients weighing the same choice.
Why is technology sovereignty the wrong thing to optimise for?
Sovereignty debates ask which companies should be trusted. The more useful question is how systems should be governed. As Tarek put it in Forbes, "Technology debates are too often framed around which companies should be trusted, rather than how systems should be governed."
For most UK enterprises the sovereignty decision was settled long ago. The cloud platforms that now run critical workloads were chosen years back, and the country built its digital economy on them. Reopening that question today slows the next wave of adoption without changing the underlying dependency. "The UK's most consequential decisions on sovereignty were made decades ago," Tarek wrote. Treating sovereignty as a precondition for AI progress holds critical infrastructure back over where the builders happen to be based.
What does good governance mean when working with a global AI provider?
A workable governance framework answers four questions before any global provider touches enterprise data. Who can access the data, and under what conditions. Whether every action the system takes can be traced and reviewed afterwards. Who is answerable when something goes wrong. And how the enterprise moves off the provider later without stranding its data or its models. Answer those four and the provider's location becomes a secondary concern. The last of them, a clear exit route, is also the enterprise's protection against vendor lock-in, which is why it belongs in the contract from the start rather than the renewal.
Does working with a US provider mean giving up control?
No. Control comes from how the relationship is governed, and an enterprise can hold it without owning the underlying capability. An enterprise can work with a global provider at scale and still hold the decisions that matter, who can access the data, how the system is audited, who is accountable, and how it would exit. What it lets go of is the assumption that it has to own the hardest capability to build in order to stay in control of it.
What is the cost of waiting for perfect conditions?
The cost is already visible. Government-commissioned research from the Department for Science, Innovation and Technology puts UK business AI adoption at roughly one in six, and in April 2026 OpenAI paused its Stargate data centre project in the UK, citing energy costs and regulatory friction. Meanwhile the prize keeps growing, with research from Microsoft and Public First estimating that AI could add around £550 billion to UK GDP by 2035 if it is scaled well. Every quarter spent waiting for an ideal domestic option is a quarter competitors spend compounding an advantage. As Tarek wrote, enterprises should "act decisively, accepting the constraints that exist and continuing to innovate with AI anyway.”
"A more pragmatic approach accepts current realities while focusing on control through governance." — Tarek Nseir, Forbes Technology Council
First published on the Forbes Technology Council, May 2026.
