Artificial intelligence news.
Straight from the original announcements.

Data centers may strain power systems more than local water supplies

November 13, 2026, 11:00By the AI News Bites Editorial TeamSource: PR Newswire

A hypothetical 100-megawatt data center could use substantial water, but its electricity demand may pose a greater local infrastructure challenge, according to research from California’s San Gorgonio Pass Water Agency. The white paper also examines how infrastructure costs, tax revenue and public responsibilities could shape the financial impact on communities.

Key points
  • A hypothetical 100-megawatt data center would directly use approximately 419 acre-feet of water annually.
  • That water use equals roughly 5 percent of Banning’s annual water use.
  • The facility would consume nearly six times Banning’s 2024 electricity use.
  • Public costs could change the financial impact of a data center project.
  • No data center projects have been brought forward in the Pass area.

Water use is significant, but limited in the local comparison

The research, conducted by UC Berkeley student Jagger Mattox for SGPWA, examined the potential water, electricity and fiscal effects of a hypothetical large-scale data center. Using industry data, Mattox estimated that a facility with a 100-megawatt IT load would directly consume approximately 419 acre-feet of water each year.

That amount is roughly 5 percent of the City of Banning’s annual water use. SGPWA said Banning was used only as a local comparison because comparable data was available, and not because of any current or potential data center project. The estimate also excludes indirect water use linked to electricity generation and chip manufacturing.

Electricity could require major infrastructure investment

The same hypothetical facility would consume nearly six times as much electricity as the entire City of Banning used in 2024, according to the white paper. Depending on existing capacity, meeting that demand could require significant new electrical infrastructure or a separate power solution.

That comparison is central to SGPWA’s argument that communities should assess data centers through a wider infrastructure lens. “Water suppliers cannot afford to plan only for the demands we already know,” SGPWA General Manager Lance Eckhart said in the agency’s announcement. “Data centers are changing infrastructure conversations across the country.”

The financial outcome depends on who pays

The white paper says data centers generally employ relatively few permanent workers and create limited additional demand for services such as schools, parks and police. A well-structured project could therefore generate significant tax revenue without a comparable increase in many municipal costs, according to the research.

That equation could change if infrastructure and utility expenses are shifted to the public. SGPWA says communities considering data centers should ask what a facility will require, who will pay for those demands and what the community will receive in return.

No data center projects have been brought forward in the San Gorgonio Pass area, and SGPWA is not advocating for or against their development. The agency said it commissioned the research to understand a fast-emerging issue before it becomes an immediate planning question. Mattox’s full white paper, Overview of Data Center Development in the Pass Area: Water Demand, Energy Demand, and Policy Considerations, is available at sgpwa.com.

San Gorgonio Pass Water AgencyUC BerkeleyJagger Mattoxdata centersenergyresearch findings

Related news

AI100 report warns the scaling era is approaching its limits
ResearchNovember 10, 2026, 15:00

AI100 report warns the scaling era is approaching its limits

The third report from the One Hundred Year Study on Artificial Intelligence says the scale-driven approach behind recent AI progress is running into limits involving…

AI energy efficiency varies sharply by model, hardware and task
ResearchNovember 5, 2026, 01:47

AI energy efficiency varies sharply by model, hardware and task

AI models that produce the most valuable work do not necessarily deliver the best return on energy, according to a study from researchers at Stanford, UC Berkeley and UT…

Joules-to-GDP: Smaller Gemma 4 31B beats GLM 5.2 on value per unit of energy
ResearchNovember 5, 2026, 01:47

Joules-to-GDP: Smaller Gemma 4 31B beats GLM 5.2 on value per unit of energy

The AI model that produces the most valuable work is not always the best use of energy, according to Joules-to-GDP, a paper from researchers at Stanford, UC Berkeley and…