northern West Virginia · Brownfield infrastructure

A grid-ready compute and energy campus in West Virginia

Silicon Foundation is redeveloping a brownfield industrial site in the northern West Virginia into a phased GPU data center and energy technology campus built around existing power, cooling, fuel, BESS, and modular power infrastructure.

10MWActive 69 kV grid connection energized today.
100MWLong-term expansion path through infrastructure buildout.
69 kVExisting high-voltage infrastructure supporting the initial deployment path.
3 mo.Target timeline for first rack online after commercial agreement.
Strategic rationale

The bottleneck is infrastructure

AI compute demand is growing faster than data center supply can respond. For new sites, the limiting factor is often grid access, power availability, cooling density, and interconnection timing.

The West Virginia site starts with the hard parts already present: energized high-voltage service, industrial zoning, river cooling potential, gas infrastructure, and a site plan that can phase from an initial 10 MW deployment toward a larger regional compute hub.

PowerActive 10 MW grid connection at 69 kV, with a defined path to 20–30 MW and a longer-term 100 MW campus.
CoolingRegional water access supports closed-loop liquid cooling for high-density GPU deployments.
MarketRegional grid-market access creates a basis for interruptible compute, capacity-style revenue, frequency regulation, demand response, and BESS arbitrage.

Built on industrial bones

The site repurposes brownfield industrial infrastructure for modern energy and compute operations. The investment thesis is focused on speed to power, high-density compute, dispatch flexibility, local industrial reuse, and monetization across both compute and grid services.

Phase 1 targets a 10 MW modular data center system with compute, BESS, on-site backup power, and EMS-controlled dispatch.

Facility infrastructure

Hard assets, measurable advantages

69 kV grid accessActive high-voltage connection reduces exposure to multi-year interconnection queues faced by new greenfield sites.
Industrial footprintExisting industrial building and industrial-zoned space support modular data halls, power equipment, BESS, and future manufacturing functions.
16" gas pipelineOn-site gas infrastructure supports backup power and behind-the-meter generation for resilience and operating flexibility.
River coolingWater access enables closed-loop liquid cooling and immersion-ready designs for 100–200 kW per rack workloads.
BESS integrationPhase 1 includes a 6 MW / 12 MWh BESS concept for UPS support, demand response, and grid-service revenue.
Fiber accessLow-latency fiber access supports compute customers and future high-density data center operations.
On-site infrastructure

Industrial capacity with a local operating footprint

The West Virginia site is being developed as more than a data center shell. The campus plan includes compute capacity, BESS deployment, power infrastructure, and a path to develop modular power systems on site.

Battery systems are part of the operating model: they support UPS functionality for compute, grid-service participation, and demand response when power prices or grid conditions require curtailment.

Longer term, the facility can support development and assembly of Silicon Foundation power pods and modular compute-and-power systems, creating a practical industrial role for the site beyond hosting GPUs.

Deployment path

Phased buildout

Phase 1

10 MW modular data center system

Compute, BESS, on-site backup power, and EMS-controlled dispatch. The objective is fast initial capacity using existing grid access, modular deployment, and a power architecture that can evolve into standardized power pods.

Phase 2

20–30 MW expansion

Grid feed upgrades, expanded cooling capacity, additional data halls, and on-site generation commissioned from the property’s gas infrastructure.

Phase 3

100 MW regional compute hub

Large-scale data center expansion with behind-the-meter generation, high-density rack capability, BESS integration, and manufacturing capacity for modular compute and power systems.

Market model

Compute revenue plus grid services

The operating model is designed around interruptible AI compute and regional grid-market participation. Revenue streams can include compute hosting, capacity-style programs, frequency regulation, demand response, and BESS arbitrage.

EMS-driven dispatch allows the facility to respond to grid signals, power price conditions, backup power requirements, and compute availability in a coordinated way.

EPC scope

End-to-end execution

The EPC scope covers electrical and civil design, interconnection work, 69 kV substation design, procurement, construction management, commissioning, EMS integration, long-term O&M, and warranty.

Infrastructure for energy-backed compute

The West Virginia campus is being developed as a practical bridge between power infrastructure, high-density compute, and market-responsive energy operations.

Back to main