The UK is dropping £100M to build its own ChatGPT rival

Watching ChatGPT take over the world from across the Atlantic, the UK government has decided it cannot afford to be a bystander.

Prime Minister Rishi Sunak and Technology Secretary Michelle Donelan have committed an initial £100 million (around $124.5 million) to build a Foundation Model Taskforce. The mission is simple: develop the UK's own "sovereign" AI to rival models like ChatGPT and make the nation globally competitive.

Borrowing the vaccine playbook

The structure of this new initiative draws direct inspiration from the UK’s COVID-19 vaccine taskforce, which bypassed bureaucratic red tape during the pandemic. The AI team will report straight to the Prime Minister and the Technology Secretary, with a chairperson set to be named this summer.

This £100 million pledge is not acting alone. It lands alongside roughly £900 million ($1.1 billion) already allocated in the UK budget for an exascale supercomputer and dedicated AI compute resources.

Downing Street is also pitching safety as its main differentiator. Amid calls from researchers for a six-month pause on giant AI training runs, UK officials want their sovereign systems to avoid the hallucinations, refusals, and weird behavioral quirks seen in early commercial chatbots. Technology Secretary Donelan believes trustworthy AI can give Britain an edge, especially in healthcare, clean energy, and public sector automation.

DeepMind is in London, but the glory went to the US

The UK has never lacked AI talent. Google DeepMind operates primarily out of London and drives some of the most groundbreaking breakthroughs in computer science.

Yet the cultural and commercial lightning strikes of this boom belong entirely to the US, led by San Francisco-based OpenAI. Having world-class researchers down the street doesn't help much if the intellectual property and consumer mindshare stay tethered to Silicon Valley.

By launching a state-backed taskforce, Britain is trying to ensure its domestic industry stays relevant rather than becoming entirely dependent on imported American APIs.

Why it matters

Let’s be honest: £100 million is pocket change compared to what Microsoft, Google, and OpenAI are pouring into training frontier models. You cannot build a GPT-4 killer on that budget alone.

However, we think the strategic bet here is smarter than it looks on paper. Instead of racing Silicon Valley on raw brute-force scale, the UK is anchoring its sovereign model to reliability, domestic public services, and strict safety benchmarks. For government agencies, hospitals, and local enterprises terrified of sending sensitive data overseas, an official, British-hosted foundation model is a compelling alternative.

This move also marks a turning point for the industry. AI development is officially shifting from a private corporate arms race into an arena of national sovereignty. Once a G7 government treats foundation models like vaccines or critical energy infrastructure, expect other nations to follow suit with their own state-funded compute and domestic models.