Analysis: Data Centers in Chesapeake
Version 1.1, 7/16/2025
Data Center Benefits
- Meets growing need for computational capacity for internet, smartphones, streaming video, social media, and now AI.
- Potential of new tax revenues, especially through business personal property on computer equipment and real estate taxes. https://netchoice.org/wp-content/uploads/2020/07/2020-DC-report-Virginia.pdf
- Return on City investment in the 175-mile, city-owned fiber optic network built to serve as the digital backbone for our community that will connect to a larger regional ring network. https://www.cityofchesapeake.net/380/Chesapeake-Broadband
- Employment during construction and a small, but well-paid, workforce, on the order of 50 full-time employees per data center during operation. https://jlarc.virginia.gov/landing-2024-data-centers-in-virginia.asp
- Diversifies local economy and attract high-paying high-tech jobs and industries that may co-locate with data centers. https://rva757connects.com/global-internet-hub
Data Center Issues
- Power Consumption substantially higher than other types of commercial or industrial operations.
- Energy consumption for a 300,000 sq. ft. facility can run anywhere from 40 MW to 120 MW, or more depending on the “energy density” of the computer equipment. Historically, data centers consumed about 150 W/sq. ft. of floor space, but now 200 to 300 W/sq. ft. are becoming standard, and AI implementations can use over 400 W/sq. ft. 40 MW to 120 MW is, equivalent to usage from 24,000 to 72,000 homes. https://vantage-dc.com/wp-content/uploads/2020/04/Vantage-what-a-high-density-data-center-means-today.pdf
- Assuming that a 300,000 sq. ft. facility uses 75 MW of power (250 W/sq. ft.), a single such facility would use 7% of the projected 9.5 million MWh/year output of the Coastal Virginia Offshore Wind Project. That is enough power for 45,000 homes. Alternatively, 75 MW would require a more than 500-acre solar farm, just for daytime operation. https://coastalvawind.com/resource/faqs, https://insideclimatenews.org/news/30012025/virginia-solar-development-debate/
- Persistent Low-Frequency Noise from large array of required HVAC equipment.
- This large power input in a tightly confined space results in a tremendous heat load that has to be dissipated through cooling systems, requiring a large array of air-cooled chillers and air-handling units. This results in severe noise and vibration impacts. Assuming 40 MW of heat dissipation requirement, over 11,000 tons of cooling capacity would be required – about 22 x 500-ton chiller units. https://www.engineeringtoolbox.com/cooling-loads-d_665.html
- Air-cooled chillers, cooling towers, air handling units, and internal server fans are key noise sources. Air-cooled chillers pose a significant noise issue. Chiller fans and compressors produce noise levels of up to 100dB(A) at the source. Backup generators used during power outages, or even periods of mandatory power conservation, can generate noise up to 110 dB(A) at the source. https://ketchumandwalton.com/5-sources-of-noise-in-data-centers-how-to-control-it/
- Persistent noise, resulting in severe degradation of perceived quality of life, is a major complaint from neighbors of existing data centers.
- Over time, long-term exposure to low-frequency noise leads to symptoms of cognitive impairment, sleep disturbances, stress, anxiety, and increased risk of cardiovascular disease. https://neurolaunch.com/psychological-effects-of-low-frequency-sound/, https://www.datacenterknowledge.com/data-storage/what-are-the-5-main-causes-of-noise-in-data-centers-
- The acceptable levels for low-frequency noise are not well established. Certain individuals are more sensitive than others, including the middle-aged and elderly, with a majority of them being women, and children with autistic characteristics. Pregnant women are advised to avoid exposure to low-frequency noise. https://www.sciencedirect.com/science/article/abs/pii/S0149763420306722, https://www.cdc.gov/niosh/reproductive-health/prevention/noise.html, https://journals.lww.com/nohe/fulltext/2004/06230/low_frequency_noise_and_annoyance.6.aspx
- Low-frequency fan hum or drone from the equipment can travel long distances, doesn't register in dB(A) noise measurements, and silencers or barriers are not effective in reducing this type of noise. https://invc.com/noise-control/data-center-noise-attenuation/
- For the Etheridge Data Center proposal, City staff examined proposed regulations for low-frequency noise from Prince William County to develop proposed proffers for the Etheridge Data Center. However, Prince William County has not yet adopted the low-frequency noise ordinance and is, in particular, still looking at challenges for measurement and enforcement of the proposed ordinance. https://www.pwcva.gov/department/planning-office/data-center-ordinance-advisory-group
- Noise analysis pre-design is only a simulation and the results of the analysis depend on the assumptions that go into the simulation. This limits its value because the final design of the building, the quantity of cooling, the quality of the equipment used, and the implementation of sound attenuation could result in very different real-life outcomes. Adequate safety margins must be incorporated and effective testing of actual operation post-construction is needed. Enforcement mechanisms are also needed when the actual performance does not match the simulation. https://invc.com/noise-control/data-center-noise-attenuation/
- Note: The Etheridge Data Center noise analysis only assumed 18 MW of cooling capacity, likely far less than required for a facility of that size, which likely means the analysis underestimated the noise impact of that proposal.
- The City Noise Ordinance currently specifically exempts “Noises generated by the operation of heating, ventilation and air conditioning units (HVAC units) attached to a building or structure”, “Noises created by the operation of generators during power outages or under other emergency or necessary circumstances”, and “Noises resulting from lawful activities in the M-1 Light Industrial, M-2 General Industrial and M-3 Waterfront Industrial Districts”. If a data center were proposed in an area currently zoned for industrial use, the City would likely have no enforcement ability on noise issues under the current ordinance. Even when a voluntary proffer on noise limits is offered as part of a rezoning, it is unclear how enforcement may be handled if the developer or subsequent operators of the data center find it unreasonable to meet the proffer requirement. https://library.municode.com/va/chesapeake/codes/code_of_ordinances?nodeId=PTIICOOR_CH26EN_ARTVNO_DIV4EXSPPE_S26-141EX
- Huge Potable Water Consumption through evaporative cooling in standard HVAC systems
- Most data centers use water to cool the computer equipment with water circulated in tubing through the equipment to absorb heat. The heated water is then moved to the external chillers to be cooled. Substantial water is lost to evaporation in the cooling process, on the order of 0.5 gallons per kWh of energy usage. An often-quoted average usage for a mid-sized data center is 300,000 gallons of water per day. Google estimated that their average data center used 450,000 gallons of water per day. 300,000 gallons of water per day would be equivalent to the average daily water usage of 1000 homes and about 3% of the output capacity of the Northwest River Treatment Plant. https://www.datacenterdynamics.com/en/opinions/an-industry-in-transition-1-data-center-water-use,
https://blog.google/outreach-initiatives/sustainability/our-commitment-to-climate-conscious-data-center-cooling/,
https://www.cityofchesapeake.net/863/Plants-Sources - Microsoft has announced that it is piloting new technology closed-loop, zero water evaporation design, with the first facilities coming online in 2027. But the upfront costs are higher than traditional systems and these are generally only being considered for areas with water restrictions. https://www.datacenterdynamics.com/en/news/microsofts-upcoming-data-centers-to-use-closed-loop-zero-water-evaporation-design/
- Rapidly Expanding Electricity Infrastructure for generation and distribution across the region.
- On a state-wide level, unconstrained new data center construction is projected to cause an unprecedented increase in energy generation requirements, nearly tripling the power requirements in the state from around 11,000 GWh/month today to over 30,000 GWh/month by 2040. Even if only 50% of the proposed data centers are built, providing enough power would require adding 2 GW of electricity generation in the form of solar, wind, biomass, gas, and nuclear every year for 15 years. Data centers also present a risk of stranded costs for Dominion Energy and its ratepayers, if electricity infrastructure is scaled up and then data centers are not built or are shut down; while electricity infrastructure represents decades-long investments, specific computing technologies may become obsolete in a few months or years. https://jlarc.virginia.gov/landing-2024-data-centers-in-virginia.asp
- Meeting the growing power need will have local impacts potentially including
- Adding thousands of acres of solar farms, constructing local gas-fired power plants, adding new transmission lines, and expanding electric power substations, with all the disruption associated with such projects.
- The cost of the rapid expansion of the electrical supply system has historically been passed on to all the rate-payers, likely resulting in higher electricity bills.
- As demand outstrips the supply of electricity, economic theory predicts higher costs for electricity generation in addition to the cost of expanding the infrastructure.
- Meeting the goals of the Virginia Clean Economy Act will be extremely challenging adding to the growing impact of climate change on the region.
- Excessive electricity demand can also result in mandatory power conservation or “rolling blackouts” where electricity is restricted to certain customers. During any power outage or such periods of mandatory power conservation, data centers will switch to their onsite backup generators. These typically are diesel or gas-fired turbine generators that emit significant harmful air pollutants that include nitrogen oxide compounds (NOx), sulfur dioxide (SO₂), and particulate matter (PM) during such emergency operations and periodic testing. https://www.staxengineering.com/stax-hub/the-environmental-impact-of-data-centers/, https://www.wpr.org/news/microsoft-to-use-diesel-fired-generators-as-backup-power-for-data-centers