Project Hummingbird proposes a hyperscale AI data center on 1,400 acres of a former coal mine in Greene County, Pennsylvania, pairing servers with on-site 910 MW gas turbines and a water treatment plant.
The facility would be a massive, self-powered load, drawing up to 17 million gallons of water per day for turbine and server cooling, with steam discharge, which raises obvious questions on river impact and thermal footprint.
Local officials frame it as an economic reset: thousands of construction jobs, 150-200 permanent data center roles, 40-50 power plant jobs, and significant tax revenue, all while forcing long-delayed brownfield remediation.
Community pushback centers on noise, air concerns, industrialization of a now-quiet area, and skepticism that a highly automated AI campus will truly replace the breadth of lost coal and steel jobs.
The project currently lacks a data center operator/investor, underscoring that the power plant piece is more defined than the actual AI workload or vendor stack.
Strategically, this is a textbook “coal-to-compute” play: turning stranded industrial land and regional gas into AI infrastructure, with energy, water, and social license as the real gating factors, not GPUs.
Worth reading for a ground-level view of how hyperscale AI buildouts collide with legacy energy communities and environmental constraints.
Source: Greene County residents debate proposal to put data center on former coal mine – Farm and Dairy
