As the AI boom rapidly expands data center electricity and water use, the inconsistency of environmental disclosures by Google, Amazon, Microsoft, and Meta is emerging as a problem. A researcher's analysis rates Meta as the most transparent and Amazon as the least.
July 2026 · AI Environmental Disclosure
Big Tech's Data-Center Footprint: Same AI Boom, Wildly Different Transparency
As AI demand explodes electricity and water use, an independent analysis ranks Meta the most transparent and Amazon the least — with no common standard forcing comparable disclosure.
23 GW
AI power demand by end-2025 (201.5 TWh/yr) — up from 5.3–9.4 GW a year earlier
~1T L
Est. water used by North American data centers in 2025 — near NYC's annual demand
25th
AI's global carbon rank if it were a country — on par with Egypt (32.6–79.7 Mt CO₂)
Transparency ranking of the four hyperscalers
Column height reflects overall disclosure transparency (higher = more transparent).
Meta
Most transparent — only firm disclosing indirect water use
Google
Strong energy metrics; withholds global water average
Microsoft
New per-site water & power metrics disclosed
Amazon
Lowest — strong efficiency but no total power use
The hidden water cost of electricity
In Meta's 2025 data — the only such disclosure — indirect water use from power generation dwarfs the water used on-site.
24×
Indirect (power generation)
Progress
Amazon reports 0.03 gal/kWh average water efficiency (a 52% improvement); Microsoft debuts a new AI data-center design using no water cooling . Closed-loop cooling is spreading.
Concern
Google's water use hit 10.9B gallons (+34% YoY) . Investors press for site-level detail as centers cluster in drought-prone Texas and Arizona; third-party facilities stay undisclosed.
The core problem: no common standard
With no shared reporting rules or regulation, firms cite "competitive reasons" and disclosures remain "apples-to-oranges."
UN call
Guterres urges disclosure of data centers' "full footprint"
New initiative
UN AI Environmental Transparency Initiative launched June 2026
Concentration
Hyperscalers hold ~2/3 of the top 15 firms' capacity
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…