AI Data Centers in West Virginia: Every Facility Tracked (2026)

West Virginia is emerging as a new frontier for hyperscale computing, with two major development sites, one of them already described as a multi-billion-dollar investment. Ban the Bots tracks artificial intelligence (AI) infrastructure across the country, and the development of ai data centers in west virginia currently centers on two distinct locations: a planned multi-building campus by Google in Putnam County and the massive, actively developing Monarch Compute Campus run by Nscale in Mason County. While neighbouring Virginia remains the largest data center market in the world with over 8,300 megawatts (MW) of tracked capacity across four major clusters, West Virginia's projects are designed around extreme scale, self-generated power, and direct river access.

AI Data Centers in West Virginia: Facility by Facility

Development in West Virginia is split between an early-stage corporate hyperscale campus and an independent microgrid-powered wholesale compute site. Each site has its own ownership structure, power arrangements, and construction timeline.

Google Putnam County Data Center (Buffalo, West Virginia)

On March 27, 2026, Google announced the acquisition of a 1,700-acre parcel of land in Putnam County to develop a major cloud and AI compute hub. According to trade reporting from DataCenterDynamics, the site is located at 22272 Charleston Road at the northern edge of Buffalo, positioned directly along the Kanawha River.

The project is currently classified as planned, with Google conducting early-stage due diligence prior to breaking ground. Google's design calls for three individual data center buildings, each measuring approximately 300,000 square feet, supported by three on-site electrical substations. While Google has not disclosed the exact capital expenditure figure, company representatives have described the development as a multi-billion-dollar investment. As part of its initial development framework, Google has stated that it will fund all necessary electrical, water, and municipal sewer utility upgrades required to service the 1,700-acre property. The facilities' water source will be the Kanawha River.

Monarch Compute Campus (Mason County, West Virginia)

The Monarch Compute Campus is an industrial-scale artificial intelligence compute installation located near Point Pleasant in Mason County. Encompassing up to 2,250 acres, the site is owned and operated by Nscale, an AI infrastructure firm that acquired the campus from original developer AIPCorp, which was sponsored by Fidelis New Energy and 8090 Industries. Nscale announced the development at NVIDIA's GTC 2026 conference on March 16, 2026.

As reported by DataCenterDynamics, the site has secured Microsoft as its primary anchor tenant. Microsoft entered into a 1.35 gigawatt (GW) compute-services agreement at the campus to power deployments of NVIDIA Vera Rubin NVL72 graphics processing units (GPUs), utilizing NVIDIA's DSX AI Factory reference architecture. Across its broader global footprint, Nscale maintains over 1 GW of operating capacity, but Nscale claims the Monarch site possesses a long-term build-out potential exceeding 8 GW.

Construction is actively underway in Mason County. Nscale has established a target of generating 2 GW of dedicated operational power on-site during the first half of 2028. Hardware installations are scheduled to roll out in multiple phases, with initial GPU deployments slated to begin in late 2027. While this project is not yet entered in Ban the Bots' primary database table, it will be added to the interactive data center map as site tracking continues.

Community Pushback and Construction Damage in Mason County

While economic development agencies promote data centers for tax generation, large-scale grading and site preparation can create immediate, severe disruptions for rural communities. In Mason County, construction at the Monarch Compute Campus has resulted in documented structural and environmental damage to adjacent homes in the Meadowlands Estates subdivision.

On Saturday, May 23, 2026, severe weather dropped approximately a month's worth of rainfall across the region within a 48-hour window. As reported by local television station WSAZ, the heavy runoff overwhelmed temporary storm erosion controls installed on the Monarch property. A key section of silt fencing failed under water pressure, sending floodwater into University Lane in Meadowlands Estates. The flash flood inundated garages, entered residential crawlspaces, and damaged heating, ventilation, and air conditioning (HVAC) systems. No human injuries occurred, but property damage was substantial.

Nscale site manager Jason Bechtle publicly addressed the event through local news broadcasts, stating: "We were on site within an hour of the first call and we haven't left. We're covering all cleanup and repair costs." Nscale paid for local hotel rooms, provided meals to displaced families, and dispatched professional remediation crews to clean basements and repair waterlogged equipment. West Virginia state officials subsequently verified that Nscale began coordinating with the West Virginia Department of Environmental Protection (DEP) and the Governor's Office to re-engineer retention barriers and prevent future washouts.

Flooding is not the only issue facing neighbors. Additional reporting by WCHS-TV on May 28 and 29, 2026, revealed recurring physical property conflicts. Boundary clear-cutting caused trees to fall across private property lines, crushing residential fences. One homeowner reported losing five chickens after a felled tree destroyed an enclosure. Furthermore, nearby homeowners reported persistent ground vibrations from construction activity throughout the day.

Joshua Jeffers, a 20-year homeowner in the area, described the daily conditions to WCHS-TV reporters: "Our house just shakes and vibrates all day. You can hear [the] toilet seat just rattling on the toilet." Jeffers expressed frustration regarding property values and the possibility of leaving, adding: "We aren't even putting ours up for sale. We know it's just pointless." Although Nscale reaffirmed its commitment to pay for all direct flooding repairs and cleanup expenses, local reporting noted no broader homeowner buyout or relocation packages have been offered to the community. Broader national trends regarding local friction, zoning fights, and quality-of-life impacts are documented in Ban the Bots' guide to data center impacts and the AI backlash tracker.

Power and Water Demands of AI Data Centers in West Virginia

Hyperscale facilities require unprecedented volumes of electric power and cooling resources. The two projects in West Virginia address these requirements through vastly different technical strategies.

Microgrid Generation vs. Grid Connection

Rather than relying entirely on traditional utility substations that can strain regional transmission networks, the Monarch Compute Campus uses an independent microgrid model. According to Nscale, the Mason County facility represents the first state-certified behind-the-meter AI microgrid operating in the United States. Electric power for the complex is generated on-site using Caterpillar G3500-series natural gas generator sets. This behind-the-meter architecture insulates local residential customers on the commercial grid from the sudden, massive power draws demanded by 1.35 GW of Microsoft GPU clusters. Nscale has stated that while the primary design emphasizes complete grid independence, the microgrid may incorporate the capability to export excess electricity back into the public grid in future operational phases.

In contrast, Google's Putnam County development plans to integrate directly with regional grid infrastructure. Google intends to construct three dedicated electrical substations on its 1,700-acre site. To prevent local utility customers from absorbing the costs of this high-voltage equipment, Google has committed to fully fund all electrical utility upgrades required for its Buffalo campus.

Water Consumption and Resource Management

Water usage remains a primary point of scrutiny for industrial computing installations. At Google's Buffalo location, the data center buildings will sit adjacent to the Kanawha River, which is the site's stated water source, while Google pays for the construction of all municipal sewer and water connection upgrades required for facility operations.

At the Monarch Compute Campus, Nscale claims its system uses a high-efficiency design engineered for lower overall water consumption. Nscale asserts that the campus will not draw water from municipal utility supplies or compete with residential water systems in Mason County. In addition, Nscale public filings state that the project will incorporate planned carbon-sequestration measures designed to offset carbon emissions generated by its natural gas engine generators.

FAQ: AI Data Centers in West Virginia

How many AI data centers are planned or under construction in West Virginia?

Ban the Bots tracks two major AI data center developments in West Virginia as of 2026: Google's planned 1,700-acre campus in Putnam County and Nscale's actively under-construction 2,250-acre Monarch Compute Campus in Mason County.

What is the Monarch Compute Campus?

The Monarch Compute Campus is an artificial intelligence compute facility in Mason County, West Virginia, developed by Nscale. It features a 1.35 gigawatt anchor tenant contract with Microsoft for NVIDIA Vera Rubin GPUs and generates its own power using Caterpillar natural gas generator sets on a certified behind-the-meter microgrid.

Has the Monarch Compute Campus construction caused any damage to residents?

Yes. In May 2026, heavy rainfall overwhelmed temporary silt fencing at the construction site, causing severe runoff that flooded garages, crawlspaces, and HVAC equipment in the Meadowlands Estates subdivision. Neighbors have also reported tree falls that damaged fences and continuous ground vibrations that shake nearby homes.

How does West Virginia's data center buildout compare to Virginia's?

Virginia hosts the world's largest concentration of data centers, with over 8,300 megawatts of tracked capacity across four major operating clusters. West Virginia currently has no operational hyperscale AI facilities on that scale, but the Monarch Compute Campus alone targets 2 gigawatts of generating capacity by 2028 with a long-term potential of over 8 gigawatts.

Conclusion: Tracking West Virginia's AI Infrastructure

West Virginia's entry into the AI infrastructure sector illustrates the scale and localized disruption of modern computing facilities. As construction continues on the Monarch site and Google progresses through early due diligence in Putnam County, the state will balance massive industrial investments against local water, land, and community protections. To monitor facilities across the country, explore the interactive data center map or compare regional infrastructure in our guide to AI data centers in Virginia, and continue tracking the broader physical footprint of artificial intelligence across each state.