Research by Public First, shared with NVIDIA for London Tech Week shows that increasing data centre access has positive economic effects at the local level, with the full return appearing five years after the construction of AI infrastructure such as data centres. 

Doubling data centre access could increase local Gross Value Added (GVA) by up to 2.07%, with a potential aggregate uplift of about £36.5 billion for UK economic output. Even a modest increase in data centre access could add £5 billion to UK economic output as well as producing benefits for local economies. 

Expanding data centre capacity generates targeted economic boosts, with every 1% increase in compute access driving growth in a range of sectors. ICT hubs, pharmaceutical clusters and professional services clusters are among those most likely to benefit. Data centre investments also increase their economic return overtime, peaking around five years after construction.

This peak is likely able to be accelerated, increasing the return on investment. Further research is needed, but faster adoption and uptake of AI products by businesses and academic institutions, as well as improving the local skills base are likely to be key.  

Public First will continue to work with research partners to explore the best policy interventions that can maximise the local economic benefits of data centre access and AI investments. You can read more about our research with NVIDIA here. 

Methodology: 

Our research examines the economic impact of data centre infrastructure using event study and regression analysis at granular geographic level (LSOAs) in the UK.

Key data sources:

• Data centre locations and capacities (MW) from Telegeography.
• Economic indicators (GVA, employment) from UK Office for National Statistics.
• Electricity infrastructure data from UK Distribution Network Operators.

Computer Access Metric:

• Gravity model approach considering data centre capacity and inverse square distance

Empirical Methodology:

• Regression analysis evaluating compute access against local economic output (log GVA)
• Controls for employment density, population, weighted distance to data centre, energy infrastructure, and sectoral composition
• Interaction terms analyze sector-specific effects (ICT, pharmaceuticals)
• Event study identifies economic impact timing around compute access threshold (15-20 MW) 
  

Findings highlight significant delayed economic benefits of enhanced data centre infrastructure.