Home » Insight Collections » How AI Workloads Are Transforming Data Centre Location Planning in 2025
The data centre landscape is experiencing fundamental change in 2025, propelled by extraordinary AI workload demands and their consequent energy requirements.
LandGate’s Q1 2025 Data Centre Activity Report reveals American data centres will likely consume over 600 terawatt-hours (TWh) of electricity by 2030 – a remarkable 308% increase from 147 TWh in 2023.
This represents 11.7% of total US electricity demand, creating substantial pressure on site selection strategies and energy infrastructure planning.
This transformation presents both challenges and possibilities for market intelligence professionals. Established data centre hubs, including Northern Virginia, Silicon Valley and Dallas, now face grid congestion and exorbitant land costs, forcing a tactical shift towards secondary markets like Indiana, Texas, Georgia and Ohio.
This geographical redistribution is altering competitive dynamics and generating new investment focal points that require sophisticated market analysis.
The implications for competitive intelligence are profound. Major technology firms are competing intensely to secure prime locations with dependable power access.
Microsoft’s $2.1 billion AI-focused data centre in Phoenix demonstrates how organisations are implementing energy diversification approaches, utilising combinations of solar, nuclear and natural gas power sources.

Oracle currently dominates Q1 2025 investment at roughly $100 billion, with AWS following at $90 billion and Microsoft at $80 billion.
For strategic planning analysts, these shifts mark a decisive turning point requiring forward-thinking intelligence.
Data centres are evolving from conventional 30-50 MW facilities to 100+ MW hyperscale operations, with the average American facility projected to expand from 40 MW to 60 MW by 2028.
The anticipated $137.5 billion in US data centre market revenue for 2025 highlights the economic importance of this transformation.
This analysis explores how market intelligence platforms help organisations navigate this rapidly changing landscape, identifying emerging opportunities in secondary markets whilst delivering the data-driven insights necessary for strategic site selection decisions in an increasingly competitive environment.
Research Context
This analysis draws from LandGate’s Q1 2025 Data Centre Activity Report, a comprehensive market intelligence document that tracks data centre growth patterns, energy consumption trends and geographical distribution of investments across the United States.
LandGate, a leading provider of data solutions for site selection and market analysis of US renewable energy and infrastructure projects, brings credibility through its specialised focus on land availability and energy infrastructure analytics.
The report methodology combines quantitative data on power consumption, investment volumes and geographical distribution with qualitative insights on market shifts and technological innovations.
The findings have been compiled through analysis of public and private sector investments, regulatory developments and infrastructure planning across multiple states, with particular emphasis on emerging secondary markets.
For market intelligence professionals, this research is particularly valuable as it identifies nascent trends in site selection strategy, illustrates how AI workloads are reshaping infrastructure requirements and provides forward-looking insights into the competitive landscape of data centre development.
The data-driven approach offers competitive intelligence teams actionable intelligence on where major players are investing and how energy infrastructure considerations are influencing strategic decision-making.
Main Themes
The AI-Driven Power Paradigm Shift in Data Centre Planning
The explosion of AI workloads has fundamentally altered data centre power requirements, creating a new competitive landscape for site selection.
Traditional facilities requiring 30-50 MW of power are being supplanted by next-generation data centres demanding 100+ MW capacity. This 200% increase in power density necessitates completely different approaches to location strategy and energy infrastructure planning.
Strategic intelligence teams must now prioritise locations with access to diverse energy sources capable of providing continuous, high-output power. Sites near natural gas pipelines, nuclear plants and renewable energy facilities have become premium assets, as evidenced by the strategic placement of Microsoft’s $2.1 billion Phoenix facility.
This represents a significant shift from traditional site selection criteria that primarily emphasised network connectivity and cooling capacity.
The market intelligence implications are profound. Competitive intelligence analysts must now monitor energy infrastructure developments as closely as they track traditional IT metrics. With US data centres projected to consume 200 TWh in Q1 2025 alone, energy access has become the primary constraint on growth and a key differentiator in site selection strategy.
Organisations with superior intelligence on energy infrastructure planning gain significant competitive advantage in securing optimal locations.
For market research managers, this trend necessitates incorporating energy resource mapping into standard competitive intelligence frameworks.
Companies that can accurately forecast power availability in emerging markets will identify investment opportunities before competitors, creating significant first-mover advantages in the race to secure prime data centre locations.
Geographical Diversification: The Strategic Shift to Secondary Markets
The data centre industry is experiencing a geographical redistribution that presents new strategic opportunities for forward-thinking organisations.

LandGate’s analysis reveals that while Virginia maintains the largest state acreage dedicated to data centres (approximately 43,000 acres), emerging markets like Wyoming are witnessing explosive growth, with the highest percentage increase in data centre development during Q1 2025.
This geographical shift is driven by three key factors that competitive intelligence teams should monitor closely.
First, traditional hubs face prohibitive land costs and grid congestion, with Northern Virginia real estate values tripling in recent years.
Second, secondary markets offer compelling incentives, as exemplified by Meta’s $800 million investment in Jeffersonville, Indiana and AWS’s $3.5 billion expansion in Columbus, Ohio.
Third, these emerging regions often provide more accessible approval processes for large-scale power infrastructure development.
The most significant strategic insight for market intelligence professionals is that this geographical diversification is creating entirely new competitive landscapes.
In Georgia alone, over $35 billion in data centre investments were announced in Q1 2025, with AWS committing $11 billion across Butts and Douglas counties on 500+ acres. Google’s operations in Georgia now exceed 1,000 MW capacity, generating $18 billion in economic activity.
For strategic planning analysts, these developments require a recalibration of competitive intelligence frameworks to account for regional variations in energy costs, regulatory environments and infrastructure readiness.
The companies gaining competitive advantage are those with superior market intelligence on emerging hubs like Ohio, Nevada, Minnesota, Wyoming, Kentucky and Wisconsin, regions previously overlooked in traditional data centre strategy.
Economic Impact and Infrastructure Transformation
The data centre boom is creating profound economic ripple effects that extend well beyond the technology sector, generating strategic implications for multiple industries.
With the sector contributing over $700 billion to US GDP and creating more than 4 million jobs, its economic significance now rivals traditional industrial sectors.
Competitive intelligence analysts should note the disproportionate fiscal impact on local economies. In Loudoun County, Virginia, data centres occupy just 4% of commercial parcels yet contribute nearly 50% of property tax revenues.
The average assessed value per data centre parcel has reached $172 million – 16 times higher than hotels, the next highest use at $11 million per parcel. This tax base transformation is creating new political dynamics that influence infrastructure planning and zoning decisions.
The market intelligence significance extends to infrastructure development patterns. The $1 billion GigaPop project in North Texas exemplifies how data centre investments are driving grid modernisation, requiring up to 540 MW of power and necessitating substantial upgrades to transmission infrastructure.
This creates strategic opportunities for companies in adjacent sectors, from renewable energy development to advanced cooling technology.
For market research managers, tracking these economic impacts provides crucial context for understanding how regional development priorities are evolving. Communities increasingly view data centres as economic development anchors, creating favourable regulatory environments to attract investment.
This shift in political economy creates strategic openings for organisations with sophisticated market intelligence capabilities to identify locations where incentives align with infrastructure readiness.
Key Statistics and Insights
- US data centres are projected to consume over 600 TWh of electricity by 2030, representing 11.7% of total US electricity demand – a critical constraint on industry growth.
- Traditional data centres requiring 30-50 MW are being replaced by AI-optimised facilities demanding 100+ MW, fundamentally altering site selection strategy.
- The data centre industry contributed over $700 billion to US GDP in 2025 Q1 while creating more than 4 million jobs across construction, operations and support services.
- Oracle led Q1 2025 investment at approximately $100 billion, followed by AWS ($90B), Microsoft ($80B), Google ($50B) and NVIDIA ($45B).
- In Loudoun County, Virginia, data centres occupy just 4% of commercial parcels yet generate nearly 50% of property tax revenue, with average assessed value of $172 million per parcel.
- Georgia attracted over $35 billion in data centre investments in Q1 2025 alone, including AWS’s $11 billion commitment spanning 500+ acres across Butts and Douglas counties.
- The average US data centre facility is predicted to grow from 40 MW to 60 MW by 2028, driving an urgent need for more sophisticated site selection strategies.
Technical Glossary
Hyperscale Data Centre: Massive facilities exceeding 10,000 square feet and 5,000 servers, designed for maximum efficiency and typically operated by major cloud providers. Distinguished by standardised infrastructure that enables rapid scaling.
Power Usage Effectiveness (PUE): Ratio of total data centre energy consumption to energy consumed by computing equipment. Industry average has improved from 2.0 in 2015 to 1.3 in 2025, with state-of-the-art facilities achieving 1.1.
Immersion Cooling: Thermal management technique that submerges servers in dielectric fluid to manage heat from high-density AI workloads. Enables 40% higher power density than traditional air cooling.
Small Modular Reactors (SMRs): Advanced nuclear power plants under 300 MW that can be factory-assembled and transported to sites. Increasingly viewed as viable power sources for hyperscale data centres requiring reliable, carbon-free energy.
ISO (Independent System Operator): Organisations that coordinate, control, and monitor electrical power system operation. Critical stakeholders in data centre development due to their role in grid capacity planning.
Direct-to-Chip Liquid Cooling: Targeted cooling system that brings liquid coolant directly to processor heat sources, enabling significantly higher computational density. Essential technology for advanced AI workloads.
Colocation Facility: Data centres where equipment, space, and bandwidth are available for rental to retail customers. Distinguished from hyperscale facilities by multi-tenant design and standardised power delivery.
MW (Megawatt): Unit of power equal to one million watts. Primary metric for measuring data centre capacity, with modern AI-optimised facilities requiring 100+ MW connections.
Key Questions & Answers
How is AI changing data centre power requirements?
AI workloads have driven power demands from traditional 30-50 MW facilities to 100+ MW hyperscale operations, necessitating entirely new approaches to energy infrastructure and site selection.
Which secondary markets are emerging as data centre hubs?
Georgia, Indiana, Ohio, Nevada, Minnesota, Wyoming, Kentucky, and Wisconsin have seen significant investment growth, with Georgia attracting $35+ billion in Q1 2025 alone.
What energy sources are proving most strategic for data centres?
Diversified approaches combining renewable energy with nuclear and natural gas provide optimal reliability. Sites near natural gas pipelines, nuclear plants, and renewable facilities are increasingly valuable.
How do data centres impact local economies?
Beyond direct investment, data centres create disproportionate tax revenue (50% of commercial property taxes in Loudoun County despite occupying just 4% of parcels) and drive infrastructure modernisation.
What cooling technologies are enabling higher densities?
Immersion cooling and direct-to-chip liquid cooling technologies are replacing traditional air cooling, enabling the thermal management of high-performance GPUs used for AI workloads.
How are workforce challenges affecting data centre development?
Labour shortages in electrical, mechanical, and plumbing systems are constraining growth, prompting industry investment in specialised training programs through community colleges and technical schools.
What role do energy analytics platforms play in site selection?
Tools like LandGate’s provide real-time intelligence on land availability and energy infrastructure, enabling data-driven decisions that match site characteristics with specific data centre requirements.






