Large, energy-intensive data centers are expanding quickly as cloud computing, cryptocurrency mining, and artificial intelligence demand more computing power. That growth creates ecological risks that extend beyond electricity use and carbon emissions. Fossil-fuel electricity and on-site diesel generators release pollutants that degrade air quality and contribute to acid deposition; water-intensive cooling reduces freshwater availability and can stress aquatic habitats; continuous cooling noise affects nearby communities and may disturb wildlife; and hyperscale campuses convert land, fragment habitat, and intensify urban heat. Institutional analyses project that the annual public health burden of air pollution associated with United States data centers could approach US$20 billion by 2030, while direct water use is already estimated at more than 100 billion gallons a year. These burdens fall unevenly on ecologically vulnerable and socially disadvantaged communities, raising questions of sustainability justice. This article examines data center impacts through an ecological sustainability lens, drawing on environmental health, freshwater ecology, acoustic ecology, land-use science, and environmental justice research. It proposes a socio-ecological framework that links five interacting impact pathways, maps them onto the planetary boundaries, and grades the evidence behind each claim, separating established findings from modeled projections and untested hypotheses. It weighs these costs against the benefits of digital infrastructure, identifies research gaps, and sets out policy priorities for aligning data center growth with planetary limits and more equitable outcomes.