Google South Carolina Data Centers Are Expanding as Lowcountry Resistance Grows

Google South Carolina data centers now represent a planned $9 billion investment, despite growing resistance to large computing facilities across the state. Google is expanding its Berkeley County campus while building two more sites in neighboring Dorchester County. Those projects place the Lowcountry near the center of America’s race to secure computing capacity for artificial intelligence and cloud services.

The construction also arrives as residents question who pays for the electricity, water, roads, and power plants that support these facilities. Colleton County has already paused data center approvals after a proposed 859-acre campus triggered litigation and sustained public opposition. Other counties have considered similar restrictions.

The conflict is not simply Google versus nearby residents. The deeper contest pits rapid infrastructure development against public demands for measurable local benefits and enforceable protections. Google chose established business parks, unlike the disputed Colleton County proposal near the ACE Basin. Still, every large project draws from the same regional supply of electricity, water, workers, and political tolerance.

What Google Is Building in the South Carolina Lowcountry

Google is turning three Lowcountry sites into a connected base for cloud and AI infrastructure.

The company has operated a data center at Mount Holly Commerce Park in Berkeley County since 2007. That facility sits near Moncks Corner, northwest of Charleston. It was Google’s first data center campus in South Carolina.

In September 2024, Google announced two additional campuses in Dorchester County. One is at Pine Hill Business Campus in Ridgeville. The other is at Winding Woods Commerce Park in St. George.

The original announcement described a combined $3.3 billion commitment. It allocated $2 billion to the two Dorchester County campuses and $1.3 billion to the Berkeley County expansion. The projects were expected to create 200 operational jobs in Dorchester County.

Google increased the scale one year later. In October 2025, the company announced a further $9 billion investment in South Carolina through 2027. The money supports continued construction in Dorchester County and expansion of the existing Berkeley County campus.

The $9 billion commitment also included workforce funding. A Google.org grant supports the Electrical Training Alliance, which planned to prepare more than 160 apprentices for work in technology and energy infrastructure.

These facilities will support Google Cloud, Search, Maps, and artificial intelligence workloads. A hyperscale data center is a large server campus designed to process enormous, continuously changing computing workloads. Its supporting systems include cooling equipment, substations, backup generators, and high-capacity network connections.

The three-site footprint matters because AI demand does not arrive as a conventional software expansion. Adding users to an AI service often requires additional accelerators, servers, cooling capacity, and electricity. The physical infrastructure grows alongside model training and daily inference, which is the computing used to generate an AI response.

Dorchester County expects the projects to produce construction activity beyond their permanent workforce. Its data center facts describe 200 full-time positions and 1,500 construction jobs connected with the two campuses.

Those numbers illustrate an important distinction. Data centers create considerable work during construction, but their operational workforces remain relatively small compared with factories carrying similar investment totals.

The state’s 2024 announcement said the three Google projects would create hundreds of jobs across Berkeley and Dorchester counties. However, it identified 200 new operational positions specifically for Dorchester County.

Google South Carolina data centers therefore offer two different economic stories. The first concerns billions of dollars in construction, equipment, and taxable property. The second concerns a comparatively limited number of permanent jobs after construction ends.

That gap has become central to the Lowcountry debate. Communities want to know whether tax revenue and infrastructure improvements will outweigh long-term demands on public resources.

Why South Carolina Became an AI Infrastructure Target

South Carolina offers available land and a business-friendly development system, but electricity has become the decisive constraint.

Google’s existing Berkeley County operation gave the company a proven regional base. Its expansion into adjacent Dorchester County allows Google to build near established transportation, utility, and telecommunications networks.

The selected Dorchester locations are designated business or commerce parks. That distinction separates them from proposals that would introduce industrial-scale facilities into rural residential areas.

State and county leaders have welcomed the investment. Their argument centers on construction work, operational jobs, tax revenue, and South Carolina’s role in the national AI economy. The state’s original announcement presented the campuses as infrastructure supporting both cloud services and growing AI demand.

Google’s timing also reflects a broader increase in AI capital spending. Training and serving large models requires clusters of specialized chips. Operators place those chips inside data centers with extensive power delivery, cooling, and networking equipment.

However, identifying suitable land is only the beginning. Developers must secure enough electricity to operate equipment around the clock. They also need reliable transmission capacity and backup systems.

South Carolina utilities have acknowledged that electricity demand is rising. Population growth, manufacturing projects, electrification, and large data centers all contribute to their forecasts.

In March 2026, Dominion Energy South Carolina President Keller Kissam said the utility lacked available energy for all the data centers seeking service. He said additional generation would be necessary beyond a proposed power plant already under review.

That statement exposes the central pressure created by South Carolina data center growth. A campus can move through local development channels before the wider public sees its full electricity requirements. The consequences then appear in debates about generating plants, transmission lines, fuel supply, and utility rates.

The Google projects do not exist in isolation. Meta is developing an AI-oriented campus in Aiken County. Other developers have pursued projects in Spartanburg and Colleton counties. Several proposed sites would demand hundreds of megawatts.

A megawatt measures electrical power at a point in time. Large campuses can require hundreds of megawatts continuously, making them fundamentally different from offices or distribution warehouses.

Google has emphasized efficiency across its global fleet. The company says its data centers deliver more computing work per unit of electricity than earlier generations. Efficiency reduces the resources required for a given workload, but it does not guarantee lower total consumption.

This effect matters because demand for AI computing continues to expand. More efficient servers can support more queries and larger models. Total power use can still rise when deployment grows faster than efficiency improves.

The infrastructure question therefore extends beyond whether Google operates efficient equipment. Regulators must determine how much new generation and transmission the combined development pipeline requires. They must also decide who bears the financial risk when forecasts change.

For technology customers, these investments promise additional computing capacity. More regional infrastructure can improve the availability of cloud services and reduce pressure on crowded computing markets.

For residents, the calculation is different. They experience the physical system through utility bills, water planning, industrial construction, and land-use decisions. Those costs remain local even when the digital services have global users.

Google South Carolina Data Centers Face a Water Accounting Test

Water is the clearest example of why communities want operational data instead of broad sustainability promises.

Data centers generate heat whenever processors perform calculations. Cooling systems remove that heat so equipment can operate safely. Some designs use evaporative cooling, while others recirculate water through closed systems.

These approaches can produce very different water requirements. Climate, workload, building design, and the source of the water also influence consumption. A single industry-wide estimate cannot accurately describe every campus.

Dorchester County estimates obtained by The State showed the possible scale. One planned Google site could use 978.2 million gallons annually in its final development phase. That is a planning estimate, not a verified record of operational consumption.

The distinction matters. Estimated demand helps utilities design infrastructure, while operating data shows the resource use that actually occurred. South Carolina historically lacked a comprehensive public system for collecting both figures across the industry.

Lawmakers responded through water-reporting proposals and the state budget. One proposal covered commercial data centers with peak demand of at least 100 megawatts. Facilities consuming three million gallons monthly would report their sources and annual use.

A fiscal-year budget provision went further by directing the Department of Environmental Services to collect monthly water information. Covered facilities must report the source, prior-year monthly volume, and expected consumption.

The reporting threshold does not resolve every concern. A facility using less than three million gallons during a month may fall outside that requirement. Public access, enforcement, and consistent definitions also shape whether the figures become useful.

Still, regular reporting creates a baseline that residents and planners previously lacked. It allows officials to compare projected requirements with actual use and examine cumulative pressure across several campuses.

Water concerns have a history in Berkeley County. Google previously sought permission for additional groundwater withdrawals at its existing campus. Local critics argued that drinking-quality groundwater should not be prioritized for computer cooling.

Google has supported community energy-efficiency programs and said sustainability guides its operations. Those efforts provide benefits, but they do not replace site-level disclosure. Communities need to know the source, quantity, and seasonal pattern of local water demand.

The proposed Google South Carolina data centers also show why cooling technology must be discussed carefully. A closed-loop system recirculates cooling fluid, reducing routine withdrawals. Yet it can still require replacement water and supporting infrastructure.

Evaporative systems can reduce electrical demand under some conditions but consume more water. Air-cooled systems generally use less water but can require additional electricity. The relevant comparison is a local tradeoff, not a universal ranking.

South Carolina’s humid summers add another variable. Cooling equipment must handle sustained heat while maintaining reliability. Operators can combine technologies, so published design details matter more than generic labels.

The useful public question is not whether data centers use water. Every major industrial facility uses resources. The question is whether developers disclose enough information for communities to plan responsibly.

For Google, transparent operating data would support its claim that the campuses can grow without undermining local water security. For regulators, it would reveal whether reporting thresholds capture the facilities placing the greatest pressure on supply.

For residents, disclosure creates an enforceable reference point. Without it, debates depend on developer projections, incomplete county records, and comparisons with facilities using different cooling methods.

The Real Conflict Is Who Pays for New Power

The Lowcountry argument turns on cost allocation, not on whether cloud computing has economic value.

Utilities build generation and transmission assets years before those investments are fully recovered. Customers then repay approved costs through electricity rates over long periods.

A hyperscale campus can transform a utility forecast. It may require a new substation, transmission upgrades, generation capacity, or long-term power purchases. If the campus arrives late, reduces demand, or closes early, other customers could inherit part of the cost.

Developers and utilities can reduce that risk through special contracts. These agreements may include minimum payments, long commitments, exit fees, and requirements covering dedicated infrastructure.

The public rarely sees every commercial term because companies treat portions as confidential. Regulators must therefore test whether a contract protects households and small businesses without revealing security-sensitive or proprietary information.

South Carolina lawmakers proposed a Data Center Siting Act that addressed this problem directly. The bill would require state certification before a data center began operating. It also called for infrastructure assessments, environmental review, water standards, and rate protections.

The proposed siting requirements would direct regulators to consider cost allocation between utilities, data centers, and other customers. The measure remained in a Senate committee, showing that a comprehensive statewide framework had not yet cleared the legislature.

South Carolina also considered moratorium legislation. One joint resolution would stop state and local bodies from acting on data center applications until lawmakers established an oversight process. That proposal defined covered facilities as those with peak demand of at least five megawatts.

Governor Henry McMaster resisted calls for a statewide pause in September 2026. He favored continued development while several counties used local moratoriums to review their rules.

This split creates an uneven policy map. One county can welcome a project into an established commerce park, while another can suspend applications. Developers gain flexibility, but regional power and water systems still cross county boundaries.

Google’s sites are less controversial than the proposed Colleton campus because they occupy planned employment areas. However, their electricity demand still contributes to broader utility decisions.

The difference between site compatibility and system impact is important. A well-sited campus can avoid placing industrial buildings beside rural homes. It can still require generating and transmission investments affecting customers across a larger territory.

South Carolina approved a major natural gas plant in Colleton County during 2026. Dominion Energy and state-owned Santee Cooper proposed sharing its output and development cost.

Utilities said population, industrial activity, and large customers drove their need for more power. Opponents argued that data center development increased the scale and urgency of those investments.

It would be inaccurate to assign an entire power plant to Google or any single company without a public allocation. Electricity planning combines multiple demand forecasts. Yet data centers clearly occupy a large and growing share of the debate.

The policy test is straightforward. If a facility triggers dedicated infrastructure, its contract should recover those costs from the facility. If infrastructure benefits the broader system, regulators should document how they separated shared benefits from project-specific expenses.

That approach does not require rejecting AI infrastructure. It requires matching financial responsibility with the party creating the demand.

Colleton County Shows What Happens When Trust Breaks

The strongest Lowcountry resistance emerged where residents believed industrial development bypassed normal public safeguards.

Eagle Rock Partners proposed an 859-acre data center campus south of Walterboro in Colleton County. The site sits near the ACE Basin, a nationally recognized landscape of rivers, wetlands, forests, and protected properties.

The proposal was not a Google project. Its importance comes from what it revealed about public tolerance for data center development across the region.

Colleton County changed its zoning ordinance to allow data centers as a special exception in a rural development district. Residents argued that the change conflicted with the county’s comprehensive plan and received inadequate public notice.

Two nearby landowners, represented by the Southern Environmental Law Center, filed a lawsuit in January 2026. Their complaint challenged the zoning change rather than a final operating permit.

The legal challenge said the proposal would place an industrial operation near homes, wetlands, and conserved land. It also focused attention on noise, light, wildlife disruption, and resource use.

The proposed campus reportedly included nine data center buildings. Court filings said construction could disturb more than 400 acres of forest and some wetlands.

Developers disputed the scale of the environmental threat. A project representative told a public hearing that the campus would affect 1.5 acres of wetlands. Eagle Rock also said it planned a closed-loop cooling system and would pay for the energy required by the site.

Those statements define the project’s central disagreement. The developer emphasized limited direct wetlands impact, recirculated water, tax revenue, and private payment for electricity. Residents focused on land-use transformation, cumulative infrastructure, noise, and proximity to protected landscapes.

Colleton County adopted a six-month moratorium in July 2026, the first data center pause in the Lowcountry. The environmental law center then agreed to pause its lawsuit during the review period.

A moratorium does not permanently reject development. It gives local officials time to define appropriate zones, setbacks, noise limits, water disclosures, and review procedures.

The episode offers Google a practical lesson. Communities judge a project partly by its technical design, but also by the credibility of the approval process. Early disclosure can matter as much as an efficiency claim.

Google’s Berkeley and Dorchester campuses benefit from established industrial locations. That advantage will hold only if local institutions maintain trust around water, electricity, construction impacts, and public costs.

South Carolina’s experience also challenges a familiar economic-development strategy. Counties have often competed for capital investment by offering fast approvals and favorable tax treatment. Data centers require a wider calculation because their permanent employment can remain modest.

The two Google campuses in Dorchester County promise 200 operational jobs against billions of dollars in investment. Those positions may be well-paid and technically significant. Still, job totals alone do not justify every infrastructure commitment.

Counties should measure property-tax revenue, service costs, water obligations, road use, emergency response needs, and the lifespan of specialized equipment. They should also disclose incentives in terms residents can evaluate.

The Colleton conflict is therefore not proof that every Lowcountry data center will face rejection. It is evidence that communities no longer accept investment totals as a complete public-interest argument.

Three Signals Will Decide Whether Expansion Keeps Its Support

The next phase will be judged through measured resource use, enforceable utility contracts, and local siting decisions.

The first signal is actual water reporting from operating campuses. South Carolina’s Department of Environmental Services now has a basis for collecting information from facilities exceeding the reporting threshold.

Those reports should clarify which cooling sources data centers use and how consumption changes across seasons. They can also show whether actual demand remains below county planning estimates.

If Google publishes site-level results, confidence in its South Carolina expansion will strengthen. If data remains inaccessible or definitions vary between facilities, public skepticism will deepen.

The second signal is the treatment of large-load customers in utility proceedings. Regulators should examine contract length, minimum payments, early termination protections, and responsibility for dedicated infrastructure.

A strong framework would protect other customers if projected AI demand fails to materialize. It would also give developers clearer expectations before construction begins.

This issue will intensify as utilities pursue new generation and transmission. The relevant question is not whether electricity demand grows. It is whether regulators can trace major costs to the customers driving each investment.

The third signal is what counties do after their moratoriums expire. Colleton County must decide whether data centers belong in rural districts near homes and sensitive landscapes. Other counties face similar zoning choices.

Clear rules could preserve development in industrial areas while restricting incompatible sites. Requirements might cover setbacks, sound studies, lighting, water disclosure, backup generators, and decommissioning plans.

The result will influence more than one proposed campus. Developers watch neighboring counties for evidence about approval times, public opposition, and infrastructure requirements.

Google South Carolina data centers occupy the stronger side of this emerging distinction. Their locations are established business parks, and Google has operated in Berkeley County for nearly two decades.

That history does not give the company a permanent exemption from scrutiny. Its enlarged investment dramatically increases the importance of verifiable local performance.

Developers often describe data centers as the infrastructure behind familiar digital services. That statement is true but incomplete. These campuses also belong to physical communities with finite resources and elected governments.

Technology customers should care because local resistance can affect the pace, location, and cost of AI infrastructure. Delayed power projects or stricter resource limits can eventually influence cloud availability and computing costs.

Enterprise buyers should also treat infrastructure concentration as a resilience issue. A service may appear entirely digital while depending on permits, water systems, power markets, and community acceptance in specific regions.

Knowledge workers can track these complex dependencies through a structured AI knowledge base. The useful task is connecting corporate announcements with permits, utility filings, and later operating data.

South Carolina does not have to choose between all data centers and no data centers. It can distinguish compatible industrial locations from environmentally sensitive sites. It can also require major users to disclose resource needs and cover project-specific costs.

That model would let the Lowcountry host important computing infrastructure without treating residents as passive recipients of its consequences.

The next announcements will attract attention because of their investment totals. The more decisive evidence will appear later in water reports, utility contracts, zoning decisions, and monthly bills.

Watch those records closely. They will reveal whether South Carolina’s AI infrastructure strategy creates durable local value or transfers too much risk to the communities hosting it.