AI data centers draw infrasound complaints from neighbors

Tom's Hardware reports that residents near some AI data centers are filing complaints about persistent low-frequency noise, including infrasound, which is below the threshold of conscious hearing but can be felt, according to the article. Tom's Hardware reports claims that measured noise near certain sites has reached 96 dB, and the article highlights conventional audible noise from on-site power systems such as natural-gas-powered turbines. The story notes that local governments have raised infrasound and noise concerns when considering moratoria on new data center projects, per Tom's Hardware. The article includes social-media postings and embedded recordings from residents documenting the sounds.

Editorial analysis - technical context: Low-frequency acoustic energy and infrasound are distinct from standard A-weighted decibel measurements used in many noise ordinances. Industry literature and acoustic engineering practice show that A-weighting under-represents very low frequencies; specialized instrumentation and measurement approaches are required to characterise infrasound and low-frequency vibration accurately. For practitioners, this means conventional noise surveys may not capture community experience when complaints center on sub-audible or felt vibrations.

Industry context: Community opposition to data centers has previously focused on water use, energy demand, and visual impact. The Tom's Hardware report places noise, especially from on-site power generation and cooling systems, as a rising local flashpoint. If municipalities consider moratoria or stricter permitting tied to noise impacts, project timelines and permitting contingencies for large hyperscale or off-grid sites could be affected. This is a regulatory and community-relations risk that sits alongside environmental permitting and grid interconnection challenges.

Indicators to follow include municipal permitting language on low-frequency noise, deployment of low-frequency measurement protocols in environmental reviews, utility or developer adoption of quieter genset and cooling technology, and any peer-reviewed studies linking infrasound exposures to health outcomes. Observers should also track whether more systematic acoustic monitoring replaces anecdotal recordings in contested siting decisions.