Why UK companies are shifting their AI workloads overseas

High electricity costs are prompting UK companies to relocate AI workloads overseas

Energy is becoming the deciding factor in where AI lives and grows. Right now, high electricity costs in the UK are forcing many companies to take their compute workloads abroad. These are not marginal experiments, AI workloads demand enormous power and cooling capacity. When domestic operating conditions drive those costs up, businesses move to where scaling AI is cheaper and more practical.

This shift is about survival in a market where efficiency wins. One in five UK firms have already moved some of their AI operations overseas. A third of leaders say energy costs limit how far they can scale, and nearly half say cost and performance outweigh sovereignty concerns when they choose where to deploy AI.

For executives, this shouldn’t come as a surprise. When power bills start dictating growth, AI deployment becomes a cost-management decision, not a national one. If the UK wants to keep AI innovation at home, businesses and government need to align on lowering energy barriers. Infrastructure investment and energy pricing reform aren’t optional, they’re essential if the UK expects to compete at scale.

Limited infrastructure is undermining the UK’s AI sovereignty ambitions

AI is not purely code or algorithms. It’s a physical industry built on power, equipment, land, and reliable grid access. Right now, the UK’s physical infrastructure is lagging behind its ambition. Businesses want to anchor AI operations locally, but they’re running into reality, expensive power, limited land availability, and grid constraints make that difficult.

Matt Hawkins, Chief Executive of CUDO Compute, summarized the problem clearly: “AI sovereignty is being hotly discussed as a priority for UK organisations, but it only works if the infrastructure exists to support it.” He’s right. Policy discussions are promising, but infrastructure must deliver. When it doesn’t, cost and performance will force companies to make decisions based on economics, not geography.

Executives leading AI adoption must see this as more than a national strategy issue, it’s an operational capacity challenge. The UK can talk about sovereignty, but unless it can power it, the discussion stays theoretical. The next few years will test whether the country can close the gap between ambition and execution. If it acts quickly on energy, land use, and grid expansion, it still has a chance to lead in AI deployment across Europe. But that window won’t stay open for long.

Geopolitical risks and regulatory considerations

Geopolitics and regulation are shaping how executives think about where to run AI infrastructure. For many UK companies, keeping workloads within domestic borders feels safer. It simplifies compliance, protects sensitive data, and reduces exposure to cross-border policy changes. About 46% of UK decision-makers say geopolitical instability pushes them to keep AI on home ground, and 45% cite data sovereignty and compliance as major influences in their deployment strategy.

Still, cost remains the overriding factor. Power prices continue to rise, and energy availability remains uneven. One in three UK organisations are considering relocation despite these security and compliance concerns. At board level, this represents a hard trade-off, choose between control and cost, knowing both are critical to long-term competitiveness.

For executives, the opportunity lies in balance. AI strategies must factor in not just power and compute costs, but global political shifts and data regulations. Companies that can manage these variables with precision, deploying in stable, cost-effective regions while maintaining data integrity, will gain resilience and maintain performance without getting trapped by short-term pressures.

The United States, India, and Eastern Europe are emerging as the most attractive markets for new AI cluster capacity

Globally, the momentum is shifting to regions ready to host large-scale AI infrastructure. The United States remains at the top of that list. It’s viewed positively by 72% of respondents in the CUDO Compute survey, thanks to its scalable infrastructure, energy supply, and robust digital ecosystems. India follows at 62%, supported by growing talent pools and improving energy accessibility. Eastern Europe comes next at 58%, attracting attention with competitive electricity prices and available land for data centre development.

Western Europe and the Nordics, while still considered stable, trail behind due to higher costs and capacity limits. The trend is straightforward: companies are pursuing regions that can host dense compute clusters more efficiently and affordably.

For business leaders, this data offers a clear takeaway. Global AI deployment is not bound by traditional tech hubs anymore, it follows cost-effective infrastructure and operational predictability. Executives should evaluate these environments not only as alternatives but as strategic extensions of their compute networks. Investing early in scalable, efficient regions positions companies to deploy AI faster, cheaper, and with fewer interruptions, key advantages in a global AI race driven by both performance and practicality.

AI-first, compute-intensive firms are more prone to relocating workloads than traditional enterprises

AI-first companies are under greater strain from power and infrastructure limits than traditional enterprises. Their operations depend on continuous access to high-performance compute, massive energy input, and efficient cooling. When electricity prices rise or local grid capacity becomes unreliable, the economics turn against them quickly. In contrast, traditional enterprises, those using AI as one of several functions, can absorb higher costs more easily or delay scaling.

According to the CUDO Compute survey, 32% of AI-first firms are considering moving workloads abroad because of energy costs, compared with only 18% of enterprise organisations. That difference underscores how sensitive compute-driven companies are to environmental factors and cost volatility.

Executives leading AI-first organisations must treat energy and infrastructure strategy as core business priorities, not secondary logistics. Where the compute runs determines product performance, innovation speed, and revenue scalability. Securing lower power costs and stable compute supply, whether through partnerships, energy sourcing, or relocation, directly impacts competitiveness. As AI workloads grow exponentially, these decisions will define which firms stay ahead and which fall behind on capability and cost.

Broader infrastructure constraints, encompassing land, energy, and grid capacity

Scalability is the real barrier between AI ambition and performance. Right now, the UK is facing constraints across all the physical inputs that matter: land availability, reliable grid access, and affordable electricity. Even with advanced chips and software stacks, without those resources the industry stalls. The result is a widening gap between policy ambitions and what the infrastructure can actually support.

Matt Hawkins, Chief Executive of CUDO Compute, has been clear about the issue. He points out that “AI is not abstract software. It is physical infrastructure that depends on power, land, cooling and grid access.” His warning aligns with what many businesses are experiencing, the tension between wanting to lead in AI and struggling to maintain operational feasibility inside the UK’s current energy and land framework.

For executive leaders, the direction forward is quantifiable and urgent. Strengthening AI infrastructure through grid modernization, renewable energy expansion, and better land-use planning will directly impact the UK’s ability to retain AI innovation. Without progress on those fronts, the country will keep losing high-value workloads to markets offering faster deployments and lower operating costs. The short-term lesson is operational, optimize where the compute runs best. The long-term solution is national, build an infrastructure base capable of sustaining the next generation of AI growth.

Key takeaways for leaders

• Rising power costs are forcing AI operations abroad: One in five UK firms have shifted AI workloads overseas to cut energy expenses. Leaders should evaluate energy sourcing and operational locations to protect scalability and margin stability.
• Infrastructure constraints are stalling AI sovereignty: High energy prices, limited land, and grid constraints are weakening local AI ambitions. Executives should align with policymakers to accelerate infrastructure investment before competitiveness erodes.
• Geopolitical and compliance concerns still matter, but cost rules: Nearly half of UK firms prioritize domestic stability and compliance, yet many still relocate due to rising energy costs. Senior teams should balance data protection with operational efficiency in deployment planning.
• The US, india, and eastern europe are emerging AI powerhouses: These regions lead in affordability and infrastructure capacity, attracting global AI investment. Decision-makers should explore partnerships and capacity expansion in these high-potential markets.
• AI-first firms feel the pressure faster than traditional enterprises: Compute-intensive companies are two times more likely to move workloads abroad. Executives in AI-focused firms should treat energy strategy as a competitive differentiator and secure predictable access to compute power.
• Energy and land access define the next phase of AI growth: The UK’s core AI challenge lies in physical limitations. Business and government leaders should prioritize grid modernization, renewable energy expansion, and land optimization to enable sustainable AI growth.