Data centers, transmission, water, and energy. South Carolina has become one of the country's fastest-growing destinations for AI build-out — and the state's choices on how to host that growth may matter more than any single policy decision in front of the General Assembly.
Most chapters of this report describe what AI is doing inside South Carolina institutions. This chapter describes what AI is doing to South Carolina infrastructure — and what infrastructure decisions are now in front of the state's utilities, regulators, and elected officials. It is, in the editor's view, the chapter most likely to shape the state's economic trajectory through 2035.
The AI applications described in earlier chapters — manufacturing-AI deployments, clinical AI, agricultural decision support, port logistics, generative-AI office tools — share an underlying dependency that is invisible to end users: the compute infrastructure those applications run on. Training and serving large AI models requires datacenter-scale electricity, cooling water, transmission capacity, and physical land. As global AI demand has accelerated, so has demand for the places to host it.
Several of those places are in South Carolina.
On April 25, 2025, the SC Daily Gazette reported that Santee Cooper, the state-owned utility, had voted to enact a new experimental electricity rate structure for large customers — defined as those drawing 50 megawatts or more from the grid. The rate is structured as a four-year pilot, designed to ensure that data centers and other large new loads contribute fairly to the cost of new generation capacity rather than have those costs flow through to existing residential and small-business customers.
The threshold of 50 MW is significant. It captures most modern AI-focused data center campuses. It also captures certain large EV-and-battery manufacturing facilities, where loads in the 100–150 MW range are now common. The rate is one of the more substantive utility-side responses to AI-driven load growth in any U.S. state and is being watched by utility regulators in neighboring jurisdictions.
The fundamental question Santee Cooper's pilot tries to answer is one being posed in essentially every state with serious data-center growth: who pays for the new generation, transmission, and grid investments that AI infrastructure requires? National analysts have estimated that AI data centers could account for up to 44 percent of U.S. electricity load growth through 2028. Without a price-signal mechanism, the cost of meeting that growth flows through to ratepayers broadly. With one — like Santee Cooper's — the cost is more directly aligned with the customers driving the demand.
This is not a settled debate. Reasonable arguments exist on multiple sides; some peer states have chosen the opposite policy posture (subsidizing rather than surcharging large loads to attract investment), and economic-development considerations cut both ways. The point for purposes of this report is that South Carolina has chosen to pose the question explicitly and structurally, in advance of most peer states. That deserves attention.
Latitude Media reported in August 2025 that Meta is building an $800 million AI data center in Orangeburg County, anchored to a 100-megawatt solar farm developed by Silicon Ranch. The project uses what Meta has been calling its "co-op" energy model: Silicon Ranch builds and operates the solar facility, Santee Cooper coordinates the local grid integration, Central Electric Power Cooperative sells the power to its 19 member cooperatives, Aiken Electric Cooperative directly serves Meta, and Meta receives the renewable-energy credits associated with the solar generation.
The arrangement is a meaningful template. It threads several needles: the data center gets renewable-attributed power, the rural cooperative ecosystem gets a major industrial customer, the utility coordinates without owning the generation asset, and the developer (Silicon Ranch) builds permanent solar capacity in the state. Whether the model scales to additional facilities is, at the time of writing, an open question — but the Orangeburg deal is structurally interesting beyond its size.
SC Daily Gazette reported in March 2025 on a $2.8 billion computing center proposed in Spartanburg County, codenamed "Moc-1." According to the reporting, the project would be built by an aeronautics-and-engineering firm, would generate its own electricity on-site (rather than drawing from the regional grid), and is intended to serve the company's own internal operations rather than function as a data-center-for-lease. The facility is expected to employ 27 full-time staff. Spartanburg County is offering substantial tax incentives — a reduced 4% property tax rate for 40 years (versus the standard 6%) plus tax credits totaling approximately $1.5 million annually for the first 20 years.
The "generates its own electricity" feature is significant. It avoids many of the grid-impact and rate-structure issues raised by the Santee Cooper pilot, and it represents a different model for AI-infrastructure development — one where the load and the generation are designed and financed together rather than negotiated across customer-utility boundaries. Whether this becomes a more common model is worth watching.
The Post and Courier reported in September 2025 that South Carolina has emerged as a regional data-center hotspot, driven by competitive electricity rates, robust transmission capacity, tax incentives, and a business-friendly siting environment. The reporting also flagged a notable structural issue: the state does not maintain a comprehensive public registry of data-center projects in development. Specific facilities are visible when they make individual public announcements; the cumulative picture is harder to assemble. Chapter 8 includes a recommendation that the state consider building such a registry.
The Virgil C. Summer Nuclear Power Station near Jenkinsville, in Fairfield County, has two unfinished reactor units that have been idle since 2017. The project's collapse — costs that grew from $9.8 billion to $25 billion before being suspended — was one of the largest infrastructure failures in the state's recent history, contributed to Westinghouse's bankruptcy, and resulted in years of subsequent litigation.
In January 2025, TechCrunch reported that Santee Cooper has been seeking partners to potentially finance and complete the two unfinished reactors, framed explicitly as a response to AI-driven baseload power demand. Nuclear's appeal in the AI context is structural — large, reliable, low-carbon, geographically dense generation, well-suited to the kind of stable 24-hour load profile data centers produce.
Whether V.C. Summer's two unfinished reactors can be revived is uncertain. The project's history is contested, the financial path forward is complicated, and the regulatory environment for restarting partially-completed nuclear construction is unprecedented. But the fact that the question is being asked at all is meaningful: it reflects the scale at which AI infrastructure demand is reshaping baseload-energy policy thinking. SCAIO will track the V.C. Summer question in subsequent editions.
Electricity dominates the infrastructure conversation, but it is not the only constraint.
Modern AI data centers can be water-intensive — particularly those using evaporative cooling. South Carolina's water resources are generally robust by national comparison, but specific siting decisions interact with watershed conditions in ways that vary across the state. Public visibility into the water-use disclosures of announced data-center projects in SC has been uneven, in keeping with national patterns. As data-center water use has become a more contested issue elsewhere (Arizona, Texas, Virginia), it is reasonable to anticipate similar scrutiny in SC. Chapter 8 includes a recommendation on water-use disclosure as part of a broader siting-transparency framework.
Adding new generation is one challenge; getting that generation to where the data centers are sited is a separate challenge. Several SC data-center projects depend implicitly on transmission upgrades that are themselves multi-year, multi-stakeholder undertakings. The interplay between Santee Cooper's transmission planning, Dominion Energy's SC service territory, and the state's electric cooperatives is not as publicly legible as the rate-structure conversation but is structurally important.
Data centers are land-intensive at scale. Siting decisions interact with local land-use patterns, agricultural land, conservation considerations, and community-impact questions. South Carolina's planning posture across counties is variable; Spartanburg County's posture (substantial incentives, openness to large industrial siting) differs meaningfully from other counties' postures.
Broadband is the substrate that connects AI applications, not the infrastructure that hosts AI compute — but the two are linked. South Carolina's substantial broadband investment over the past several years, including federal-program-funded rural buildouts, is what makes AI applications usable across rural SC in the first place. Without it, much of the AI-application picture described in earlier chapters would be unevenly accessible across the state.
The infrastructure picture above can be read in two ways.
Read pessimistically, it describes a state that has become attractive to AI infrastructure capital because of its competitive rates and favorable siting environment, and that is now confronting the costs that wave brings — utility rate pressure, transmission strain, water-use questions, and uneven local readiness. The Santee Cooper rate pilot, on this reading, is partly a defensive response to growth that is happening faster than the state's regulatory and planning posture was designed to handle.
Read optimistically, it describes a state that has positioned itself for one of the largest infrastructure investment cycles in the U.S. economy in a generation, and that has the energy assets (Santee Cooper's flexibility, the V.C. Summer question, the cooperatives' willingness to host large industrial customers, the Silicon Ranch model), the transmission baseline, and the industrial-siting posture to convert that positioning into durable economic gain. The Santee Cooper rate pilot, on this reading, is a thoughtful structural response that lets the state grow without socializing the cost — a posture peer states are now studying.
Both readings are partially right. The strategic question for SC is whether the state can use the next 24 months — when many of these decisions will harden into precedent — to choose which aspects of the optimistic reading to make true and which aspects of the pessimistic reading to mitigate. That is not a question for any single agency or institution. It is a question for the state's collective infrastructure leadership, which currently does not convene around AI as a single topic. Chapter 8 returns to that.