UK minister warns Britain must invest or be at AI’s ‘mercy’

Liz Kendall’s remarks are a reminder that AI industrial policy is no longer just about ethics white papers—it’s about whether mid‑sized economies can shape the trajectory of the technology or are forced to buy it on someone else’s terms. By framing AI as both “engine of economic power and hard power” and warning that the UK could be at its “mercy and whim,” she is explicitly linking model access, compute, and chip supply to national security.

The proposed Sovereign AI venture fund and forthcoming hardware plan suggest the UK wants to move beyond being a pure customer of US hyperscalers. Even modest on‑shore chip design, packaging, or specialized data‑center build‑out could give British regulators more leverage over safety, auditability, and export controls. At the same time, Kendall’s emphasis on partnering with other “middle power” nations hints at a bloc strategy: pool demand and regulatory heft to negotiate with the handful of companies that will control frontier‑scale compute.

For the AGI race, this doesn’t directly change model capabilities, but it does affect who can run what, where. If more countries insist on sovereign infrastructure and stake in local AI champions, we may see a less centralized, more federated development landscape—still dominated by a few US labs, but with more room for regional players to influence norms and deployment choices.