The US States Most at Risk of Falling Behind in the AI Boom

A recent study by TRG Datacenters highlights the states that are least prepared for the upcoming artificial intelligence (AI) infrastructure boom. As the American AI market is projected to grow threefold by 2029, the implications for future growth are significant, particularly for states struggling with infrastructure development.

Key Findings

The study analyzed 24 states based on critical factors influencing AI infrastructure development, including:

• Environmental impact
• Energy consumption
• Established data centers

The findings reveal that states with limited resources and high environmental stress are likely to face more challenges in developing their AI capabilities.

States Most at Risk

1. North Dakota
   – Large data center sites: 1
   – Total AI data center capacity: 68 MW
   – Share of state’s electricity used: 1.7%
   – Estimated CO2 emissions: 219.1K tonnes
   – Share of mapped data centers in high water stress areas: 0%
2. Wyoming
   – Large data center sites: 1
   – Total AI data center capacity: 152 MW
   – Share of state’s electricity used: 6.7%
   – Estimated CO2 emissions: 406.6K tonnes
   – Share of mapped data centers in high water stress areas: 100%
3. Minnesota
   – Large data center sites: 1
   – Total AI data center capacity: 178 MW
   – Share of state’s electricity used: 2.1%
   – Estimated CO2 emissions: 573.4K tonnes
   – Share of mapped data centers in high water stress areas: 44.4%
4. Wisconsin
   – Large data center sites: 2
   – Total AI data center capacity: 369 MW
   – Share of state’s electricity used: 4%
   – Estimated CO2 emissions: 1411.6K tonnes
   – Share of mapped data centers in high water stress areas: 53.8%
5. New Mexico
   – Large data center sites: 2
   – Total AI data center capacity: 87 MW
   – Share of state’s electricity used: 2.1%
   – Estimated CO2 emissions: 226.3K tonnes
   – Share of mapped data centers in high water stress areas: 100%

Implications for Future Growth

The spokesperson from TRG Datacenters noted, “US data center energy demand is projected to nearly double between 2025 and 2028, jumping from 80 to 150 gigawatts. For states like North Dakota, that trajectory creates a hard ceiling on growth that larger states with diversified grids simply don’t face.” This highlights the constraints that smaller states may encounter as they attempt to expand their AI infrastructure.

As the federal push for AI prioritizes speed and scale, the limitations of state-level grid capacity could determine where the next wave of infrastructure is built. States with limited resources are at risk of being left behind in the race for AI development.

For more detailed findings, you can access the full report here.