Improper redaction reveals Google Data Center water and electricity usage

Original Article ↗ Hacker News Discussion ↗

Improperly redacted state documents have exposed the water and electricity usage of Google data centers in Nebraska, along with tens of millions of dollars in tax refunds to the company. Commenters note that even the largest facility’s hundreds of millions of gallons per year are small compared to agricultural irrigation or municipal consumption, but argue that concentrated demand can still strain local aquifers, grids, prices, and noise limits. Many see the focus on water as a partial red herring that distracts from larger concerns about fossil-fueled power expansion, weak transparency around corporate incentives, and the rapid, AI-driven growth of hyperscale data centers.

Scale of Data Center Water Use

  • 13M gallons/year at the small Lincoln DC is repeatedly framed as modest: ~20 Olympic pools; ~0.27% of Lincoln’s daily use; ~280 US households’ annual water.
  • The larger Papillion site uses ~548M gallons/year; six Nebraska DCs together ~765M. Some see this as still small in state‑level context; others call it significant at city scale.
  • Several argue the media focus on water is misleading given much larger users (agriculture, ethanol, golf courses, refineries).

Local Water Stress, Evaporation, and Aquifers

  • Key distinction: cooling water often evaporates, so it leaves local rivers/aquifers and may not return where it was taken.
  • Others respond that total volumes are “a rounding error” in regional hydrology and that most concerns should focus on aquifer depletion and agricultural irrigation.
  • Some emphasize that even small extra loads matter in already stressed basins or drought periods.

Electricity Demand and Climate

  • Many commenters see electricity use as the real issue: new AI-focused DCs in the hundreds of MW to multi‑GW range, often powered by new gas plants, with implications for CO₂, air quality, and local power prices.
  • Counterpoint: DCs aggregate national demand efficiently; better to expand clean generation than fight DCs per se.

Transparency, Redaction, and Trade Secrets

  • The bad PDF redaction is widely mocked; several say it undermines trust more than the numbers themselves.
  • Some defend secrecy: detailed water/power data are treated as trade secrets revealing fleet size and efficiency.
  • Others argue public utilities use mandates full transparency, not NDAs and shell LLCs.

Economic Deals and Tax Breaks

  • Large tax refunds to DC LLCs (tens of millions each) alarm many; some suspect the state may be net negative on these projects.
  • Locals question why massive incentives go to facilities that provide few permanent jobs.

Local Impacts: Noise, Land, and Neighbors

  • Noise from cooling systems and generators is a major complaint in some locales (reports of ~60 dB+ hum at homes).
  • Concerns about bright night lighting, conversion of farmland or public/wildlife land to DC campuses, and increased utility rates.
  • Others note DCs don’t smell or pollute like dairies or refineries, and their job count per MW is inherently low.

Comparisons to Other Uses

  • Frequent comparisons: irrigated corn and alfalfa, almonds, ethanol production, leaking municipal systems, refineries, golf courses.
  • One recurring theme: if one truly cares about water, agriculture (esp. ethanol, livestock feed, export crops) dwarfs DC usage.

AI, Moral Panic, and Underlying Motivations

  • Several see the water discourse as a proxy war over AI itself: job fears, cultural backlash, and distrust of tech billionaires.
  • Others counter that even if AI is the root objection, resource and siting issues (water, power, noise, emissions) remain legitimate.

Cooling Technologies and Trade-offs

  • Discussion of evaporative vs closed‑loop cooling: water‑heavy but energy‑efficient vs water‑light but power‑hungry.
  • Some argue the “greenest” option depends on local water stress and carbon intensity of marginal electricity.
  • Idea of using DC waste heat for district heating or greenhouses appears, with frustration that this is rarely planned.

Policy and Governance Ideas

  • Proposals include: mandatory public reporting of large industrial water/energy use, uniform disclosure rules across industries, and stricter siting/zoning in water‑stressed or residential areas.
  • Others warn against singling out DCs when many higher‑consumption industries escape similar scrutiny.