Britain shouldn’t try to compete with America or China in the race to create cutting-edge AI models, and instead focus on widespread AI adoption. Even this will require an increase in local compute capacity. The Tony Blair Institute for Global Change warns
that the UK must “take infrastructure seriously” in order to remain competitive in an AI-era and fix systemic issues that cannot be solved overnight. The Tony Blair Institute for Global Change, a non-profit, says that this will require time and investment, as well as the government making building datacenters a priority.
Blair’s think-tank, in its reportsays that Britain lacks the resources it needs to compete with the US, China, and Gulf States who are all investing hundreds of billions into energy-hungry, vast datacenters. It adds that the UK does not have the resources, land or money to follow suit.
The focus should be on the deployment and widespread adoption of AI. “demonstrating to the world how to effectively apply it across sectors including health, education, government, defence and science.” Here, the report states, is where the economic benefits will be found. The jury is still out on whether or not this can be done.
This strategy will still require significant investment in AI infrastructure. The UK government has already acknowledged this in its AI Opportunities Action Plan ( ) unveiled at the beginning of the year and the Compute Roadmap recently detailed.
According to the Tony Blair Institute, these steps are insufficient, and claim “the situation is now dire” that the UK has placed AI as the centre of its goals for growth and security, but lacks the infrastructure necessary to achieve them. The country will not be able to reach its 2030 goal of 6 GW AI-ready capacity in the UK at the current pace of construction, says the Tony Blair Institute. It blames delays with planning and permitting, constraints on the national grid, and rising industrial energy costs as the main reasons for the slowdown.
This report is backed by a separate one from fDi Intelligence (part of the Financial Times), which claims that the UK could be facing a 5 GW shortage in datacenter compute capacity. This is based upon an analysis commissioned from the Department for Science, Innovation & Technology. It assumes that supply will almost double, from 1.8GW to 3.3GW, and that demand is expected to range between 5.1 GW to 8.5 GW.
Britain is unlikely to reach the 6 GW goal, as the majority of the 1.8G W bit barns in London are not optimized for AI.
Tony Blair Institute recommends that the government follow “accelerated diversification,” a strategy which means rapidly building out resilient infrastructure. It will be necessary to make changes in order to attract new investments, reduce risk and spread capability regionally in order to improve resilience. This is easier said than accomplished.
The government should also ensure that the National Energy System Operator (NESO), integrates bit-barn demand (as estimated from DSIT), into national plans, and builds to support it in “dynamic updates.”
But the AI Energy Council (19459034), formed last year, is tasked with ensuring that. However, it operates under a traditional British Government cloak, declining to disclose what has been discussed, or any decisions made during the two meetings The report recommends that a team of AI experts and datacenter experts should be formed within NESO in order to support demand planning and speed up AI integration into the energy system.
It is not surprising that the report also recommends modifying the planning process so that decisions are made within eight months, and utilizing ministerial call in powers for high investment datacenter projects and grid improvements.
The government has already taken these actions, namely designating these sprawling servers farms as critical infrastructure (CNI), allowing developers to override local oppositionagainst datacenters being constructed in a particular area. Developers can bypass local authorities by categorizing them as Nationally Significant Infrastructure Projects.
The government should also adopt a strategy for developing a series new gigawatt nuclear energy station projects and reform the way nuclear is regulated both to speed up building while reducing costs.
According to the Tony Blair Institute report, the Trump administration has also taken similar actions, including changing the rules for co-location of AI Datacenters and energy generation sources. It also suggests identifying government land which could be used to build bit barns to offer to private developers.
These suggestions are too much for the UK government. They are known for their sclerotic decision-making and slow pace of action. The current administration is struggling with chronic budget restrictions and the urgent need for investment in a variety of other areas.
Tony Blair Institute concludes by warning that the UK will fall behind if it fails to build. If it does it right, the UK has a real chance to lead in AI by gaining expertise in how to use it and building the right infrastructure to support.
Despite all the calls to action there are growing fears that all this AI investment might turn out to be a hype-bubble. McKinsey & Company’s report identified widespread unease, as nobody knows what level of AI demand will be in the future. Other research found that generative AI had not had a significant impactyet on earnings or hours recorded in any occupation, despite billions being poured into the building and training of the models.
Gartner recently forecast that 40 percent of projects would be abandoned by the end 2027 due to rising costs or unclear business value. Zuck tries justifying AI splurge by talking of’superintelligence for all’
In the past, Baidu’ The UK government and Tony Blair Institute are not perturbed by the generative AI critics. (r)
