Proposes a rebound-informed framework with five tests (metric, boundary, reinvestment, burden shifting, governance) showing that AI datacenter sustainability claims often rely on relative efficiency gains without proving absolute reductions in energy, water, and other burdens.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CY 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Plateau That Never Comes: When Efficiency Claims in Datacenters and AI Become Greenwashing
Proposes a rebound-informed framework with five tests (metric, boundary, reinvestment, burden shifting, governance) showing that AI datacenter sustainability claims often rely on relative efficiency gains without proving absolute reductions in energy, water, and other burdens.