【Wen/Observer Net Wang Kaiwen】Under the global artificial intelligence (AI) boom, the Starmer government in the UK launched an ambitious "AI Opportunities Action Plan" in January this year, with the goal of becoming a "superpower in artificial intelligence."

After nearly a year, where has this plan reached?

According to a December 27 report by American Consumer News and Business Channel (CNBC), many global tech giants have committed to investing in the UK's AI infrastructure. The UK government has also introduced regulatory provisions favorable to data centers and designated four "AI Growth Zones," which are positive signs.

However, critics point out that the UK's high energy costs and strict restrictions on access to the national grid have become key obstacles to the development of AI in the UK.

"The ambition is great, but the ability to implement it is still lagging behind," said Ben Pritchard, CEO of AVK, a power solution supplier for data centers.

According to the report, one of the key pillars of the "AI Opportunities Action Plan" is to quickly build data centers that can meet the massive computing power demands of AI development, which will be driven by the "AI Growth Zones." These areas offer more relaxed planning permissions and easier access to electricity.

Currently, the UK's AI infrastructure is still in its early stages, and the "AI Growth Zones" are also in the initial development phase.

On January 13, 2025, UK Prime Minister Starmer visited University College London. On that day, Starmer announced the "AI Opportunities Action Plan." IC Photo

This February, the UK's first "AI Growth Zone" was established in Oxfordshire, which is also the location of the UK Atomic Energy Authority, but the growth zone has not yet started construction. In September, another growth zone located in northeastern England began ground preparation work, with formal construction scheduled to begin in early 2026.

The other two growth zones are located in northern and southern Wales, and were announced in November. The northern Wales park is seeking investment partners, and the UK Department for Science, Innovation and Technology revealed that a partner will be confirmed within a few months. The southern Wales park consists of several areas, some of which are already in operation, while others still need to be built.

In July this year, the UK government stated that according to the plan, these core "AI Growth Zones" should meet at least 500 megawatts of power demand by 2030, with at least one growth zone exceeding 1 gigawatt.

Pritchard said the biggest challenge in achieving these goals is the limited capacity of the UK's power grid. New projects face long delays to connect to the grid.

"Developers expect that grid connection will be delayed by eight to ten years, especially in London, where the number of unprocessed connection applications has never been so high," Pritchard said.

He said that AI has significantly increased energy demand, and as more companies and consumers start using AI technology, it adds extra pressure to the UK's already strained energy system.

"These are no longer isolated risks; they are actually slowing down or hindering AI development across the country," Pritchard said.

Spencer Lamb, Chief Commercial Officer of UK data center developer Kao Data, mentioned that the "AI Growth Zones" are open to applications, leading some people who own land with transmission towers or cable equipment to apply for them. As a result, the UK National Grid received a large number of speculative applications, which have little chance of success.

CNBC reported that the public institution responsible for managing the UK's national grid, the National Energy System Operator (Neso), has taken measures to address these issues.

Earlier this month, Neso announced plans to prioritize hundreds of projects to speed up grid connection. When asked by CNBC, Neso refused to disclose whether AI infrastructure projects are on the priority list, but said that a significant portion of them are related to data centers.

On September 16, during US President Trump's visit to the UK, the UK and the US signed the "Technology Prosperity Agreement." This agreement focuses on collaboration in rapidly developing technological fields such as artificial intelligence, quantum computing, and civil nuclear energy.

The UK government later stated in a news bulletin that top US tech and AI companies, including Microsoft, Google, NVIDIA, OpenAI, and CoreWeave, have committed to investing 31 billion pounds (about 293 billion yuan) to improve the UK's AI infrastructure and cutting-edge technologies, covering data centers, computer chips, and processing capabilities behind AI.

At the same time, UK-based AI infrastructure company Nscale will provide AI computing services and build data centers, and plans to deploy tens of thousands of NVIDIA chips at an AI factory near London by early 2027.

On November 6, 2025, Nebius deployed its first AI cloud infrastructure in the UK, which uses NVIDIA's latest network technology. IC Photo

Puneet Gupta, General Manager of NetApp UK and Ireland, told CNBC that investments from major private enterprises have laid an important foundation for the UK's AI development. The momentum around national supercomputers and new computing power is also growing, but the real test is how long it will take for these plans to translate into computing power available for the UK.

Stuart Abbott, Director General of VAST Data UK and Ireland, told CNBC that for the UK to achieve long-term success in AI infrastructure, it must invest in the "full stack," including data pipelines, storage, energy, security, talent, and skills.

"If the UK wants this to continue rather than being a short-term excitement, it must view AI infrastructure as economic infrastructure," Abbott said.

CNBC pointed out that the UK faces huge challenges in developing AI. First, Europe's investment in data centers is negligible compared to the US. At the same time, the UK's current energy costs are the highest in Europe, having risen by 75% compared to before the Russia-Ukraine conflict. Additionally, the aging grid infrastructure may take many years to connect to new sites.

Pritchard said that for projects unable to connect to the national grid, a potential solution is microgrids. Microgrids are self-sufficient power networks composed of engines, renewable energy, batteries, etc.

Pritchard revealed that AVK is designing two microgrids for UK cloud computing partners, although these are not designed for AI. According to reports, these microgrids may take about three years to build, and the current cost is about 10% higher than grid energy.

Abbott believes that placing computing power in locations with existing electricity, rather than starting from scratch to develop new sites, is also a way to accelerate the launch and operation of AI infrastructure.

In Lamb's view, implementation speed will be key. "Unless the fundamental issues of energy availability and price, AI copyright, and AI research funding can be quickly resolved, the UK will miss this era's most important economic opportunity and eventually risk becoming an international AI backwater," he said.

This article is exclusive to Observer Net. Unauthorized reproduction is prohibited.

Original: toutiao.com/article/7588816733541859903/

Statement: The article represents the views of the author.