Skip to content
SINGULARITY FARMER
LearnInsightsUpdatesFollow on X ↗

The learning desk · Computing, power & water

What is a data center, anyway?

Data centers are infrastructure for a computerized future. Rural communities should learn how they work, welcome the opportunity and negotiate to own more of the upside.

By Singularity Farmer · Published · Updated and sources rechecked September 15, 2026

Share on X ↗

The short answer

A data center houses computing equipment: servers that process requests, storage that holds information, and networks that connect it all. Electrical and cooling systems keep that equipment operating. Some facilities support AI; others handle services such as websites, business software or data storage. The label alone does not describe a facility’s workload or size. Berkeley Lab studies these systems together ↗

A simplified flow: electricity powers computing, computing produces heat, and cooling moves heat out.
A simplified explanation, not a site design or energy balance. Cooling also uses energy. See Berkeley Lab’s systems research. Open the shareable field-guide graphic.

More compute is an opportunity.

Singularity Farmer is pro-compute and pro-data-center. Computing capacity is becoming strategic infrastructure—the next well, railhead or grain elevator is increasingly likely to include banks of machines. Communities should not be taught to fear that future. They should be equipped to shape it.

That means welcoming serious proposals while negotiating serious terms: who funds the power and connectivity, who owns the land and equipment, what revenue stays local, how water and noise are measured, and who pays when the site changes or closes. Private, off-grid and community-owned compute deserve experimentation too. Those are strategic directions, not promises that every design will work everywhere.

Start by asking what a number measures.

Berkeley Lab’s 2024 report estimates past U.S. data-center electricity consumption and models a range of possible demand through 2028. A scenario is not a measured result. A national total is also not a prediction for a particular building. Compare a proposal with its own engineering documents, assumptions and time period. Read the report and its scope ↗

Chart of Berkeley Lab's U.S. data-center electricity-use estimates: 58 terawatt-hours in 2014, 176 in 2023, and a modeled range of 325 to 580 in 2028.
National historical estimates and a future scenario range from Berkeley Lab’s 2024 report—not measurements or a forecast for one local facility. Berkeley Lab’s 2025 update extends its national scenarios through 2030.

Water use needs a definition.

Withdrawal is water taken from a source. Consumption is the portion no longer immediately available for another use—for example, because it evaporated. The two numbers answer different questions. USGS explains the terms ↗

Cooling equipment and operating conditions matter. A useful description names the cooling system and water source rather than assuming every data center works the same way. Berkeley Lab’s research covers air and liquid cooling, monitoring and water use. Explore its research areas ↗

A seven-part deal sheet to keep.

Use this beside an actual project plan. Record the document, date and page for each answer. A pro-building community can still insist on a good deal.

  1. Ownership: Who owns the land, facility, generation and data? Can local people or institutions hold a stake?
  2. Power: Who funds generation, transmission, storage and backup? Who bears curtailment and rate risk?
  3. Water and cooling: Which system and source are proposed? Are the figures withdrawals, consumption, or both?
  4. Community value: Which taxes, payments, infrastructure, training or contracts are binding rather than promotional?
  5. Resilience: Can the project strengthen local power or communications during an emergency, and what evidence supports that design?
  6. Land and neighbors: What enforceable terms cover noise, setbacks, lighting, traffic and restoration?
  7. Exit: Who pays to modify, decommission or remediate the site?

Go to the sources.

  1. Berkeley Lab: Data Center Energy Analysis. Overview of computing, power and cooling research.
  2. 2024 United States Data Center Energy Usage Report. Published December 19, 2024. National historical estimates and scenarios; not measurements for a local site.
  3. USGS: Water-Use Terminology. Definitions of withdrawal and consumptive use.
  4. United States Data Center Energy Usage Report: 2025 Update. Published June 2026. Updated national scenarios through 2030.

This guide expands the homepage’s short introduction. Prepared with AI assistance from the linked primary sources; published by Singularity Farmer. Its pro-compute position is an editorial thesis; its factual claims are linked to sources. It makes no findings about a particular proposed facility. First edition September 14, 2026; updated September 15, 2026.

Keep the notebook open.

Read more updates, follow by RSS, or join the conversation on X ↗.