If AI data centers are so great, why are they being built in secret?

Environmental impact: water, power, pollution

• Strong disagreement on water use:
  • Some claim modern data centers can use “very little to none” via closed-loop cooling, and that their total water use is small compared to agriculture, golf courses, or lawns.
  • Others counter that many facilities still rely on evaporative cooling, especially in hot regions, consuming significant water continuously and often sited in drought-prone areas.
  • One side stresses relative national impact is small; the other stresses local impact can still be severe even if totals are modest.
• Pollution concerns focus on:
  • Fossil-fuel-based power plants needed to feed new load, with air quality and CO₂/NOx/particulate impacts.
  • Linked reports of data center-related power projects allegedly violating or bending environmental rules.
• Some argue water-pollution claims are overstated (closed systems, minimal contamination); others say full lifecycle and power-plant impacts are opaque and should be disclosed.

Local communities, NIMBY, and permitting

• “Secret” siting is framed as:
  • A rational response to adversarial, delay-prone local processes where almost any project (grocery stores, housing, data centers) can be blocked via environmental or procedural challenges.
  • A way to avoid organized opposition, especially in politically weak rural areas that can’t extract strong tax or mitigation deals.
• Critics say:
  • Residents bear costs (higher utility prices, noise, land use, possible health risks) with little local benefit.
  • Public input and legal processes exist for a reason; bypassing hearings or shopping for compliant councils undermines democracy.
• Others argue NIMBY use of environmental law has broadly throttled industrial capacity, not just data centers.

Security and secrecy rationales

• Some justify secrecy as protection against terrorism, sabotage, or theft of high-value hardware.
• Others note large data centers are physically hardened and visible once built, questioning how much secrecy is actually about security vs politics and PR.

Economic value and jobs

• Proponents present AI/data centers as engines of economic growth, tax revenue, and technical jobs (construction, technicians, security).
• Skeptics:
  • Question the scale and quality of local employment (often relatively few, modestly paid jobs).
  • Argue AI and automation may reduce net jobs and funnel gains to shareholders.
  • Compare claims to earlier broken promises (e.g., crypto mining, big-box retail).

AI, politics, and public perception

• Some see anti–data center sentiment as ideologically driven or based on “bunk science” about heat and water.
• Others say distrust is grounded in history: industrial pollution, corporate tax avoidance, socializing costs and privatizing profits.
• There is mention of:
  • Alleged under- or misreporting of GPU and data center capacity to hype valuations.
  • Growing anti-AI populism, with climate concerns and control over society as underlying fears.
• Several commenters draw parallels to earlier industrial revolutions: large aggregate benefits, but highly uneven and sometimes negative local and generational impacts.