Nearly two-thirds of upcoming AI data centers in the US are planned in regions experiencing drought conditions, spotlighting growing tensions over water use and power supply reliability amid increasing AI infrastructure demands.
- About 66% of new AI data centers are being built in drought-hit US regions.
- Water use linked to AI includes significant indirect demand from power generation and semiconductor production.
- Some US cities, like Seattle, have paused new data center approvals due to resource concerns.
What happened
A recent study reveals that nearly two-thirds of planned AI data centers in the US are located in areas that have experienced drought conditions over the past year. This expansion is occurring despite widespread regional water shortages and concerns from local communities. Operators such as AWS employ advanced cooling systems designed for water efficiency, but the sheer volume of data center projects means cumulative resource demands are rising sharply.
Power and water consumption linked to AI is not solely from data centers themselves; indirect water use from electricity generation and semiconductor manufacturing comprises the vast majority of the demand. Estimates show only 4% of AI-driven water use is directly related to data centers, while approximately 54% is tied to power generation and 42% to chip fabrication. This complex interplay underscores the extensive infrastructure challenges behind growing AI workloads.
Why it matters
These developments are taking place as the US electric grid faces ongoing stress, raising concerns about the reliability and cost of power supply in drought-impacted regions. Consumers are expected to bear increasing electricity costs linked to expanding AI demands. The water supply is also under pressure, impacting agricultural and residential users in these drought-affected areas.
Local residents and farmers have voiced frustration about rising resource costs and limited benefits from data center growth. In some communities, these tensions have led to regulatory actions; for example, Seattle imposed a one-year moratorium on new data center construction to manage electricity prices and resource allocation. These conflicts highlight a broader debate about balancing technological advancement with environmental sustainability and equitable resource access.
What to watch next
Stakeholders will closely monitor how AI companies and regional authorities respond to resource constraints. Advances in closed-loop cooling and other efficiency measures may mitigate water use, but sustained growth in AI data centers will require significant investments in power infrastructure and sustainable water management strategies.
Policy developments in cities and states could shape the pace and location of future AI facility expansions. Monitoring regulatory trends, utility capacity, and community responses will be important for anticipating potential shifts in the AI infrastructure landscape. The evolving balance between AI innovation and resource stewardship remains a key story to watch.