The company has partnered with Stokes Inc., a South Carolina firm specializing in grid-scale battery storage and energy systems, for the project
Decision Focus
In June 2026, Silicone Foundation publicly clarified that its recently acquired Centre Foundry site in Warwood, West Virginia is intended as a module manufacturing facility for the data center industry — not a large-scale compute deployment. The site carries a 10 MW grid connection, industrial crane infrastructure, river cooling access, and an on-site BESS installation. The company has engaged AEP to determine what additional utility capacity is achievable. For Global Heads of Data Center Energy, the operational signal is not this specific project — it is the emerging model of redirecting brownfield industrial power capacity toward supply chain manufacturing rather than compute deployment, and what that means for how modular infrastructure reaches the market.
90-Second Brief
In recent days, silicone Foundation acquired the former Centre Foundry site in Wheeling, West Virginia for $1.5 million and has positioned it as a production facility for modular data center components. The site’s 10 MW utility connection limits initial compute deployment to roughly 10 units but supports a manufacturing and infrastructure staging function. The company has partnered with Stokes Inc., a South Carolina firm specializing in grid-scale battery storage and energy systems, for the project. Power capacity above 10 MW depends entirely on AEP’s response to a pending inquiry; the company’s chief officer has stated that reaching 100 MW would require a full rebuild of the local electrical grid, characterizing it as theoretical rather than planned.
What Is Really Happening?
The Warwood story is being read locally as a zoning and community dispute. For data center energy operators, the more relevant read is about where modular infrastructure supply chains are anchoring.
Silicone Foundation’s published strategic rationale names the bottleneck precisely: AI compute demand is outrunning data center supply not because of capital or land, but because of grid access, power availability, cooling density, and interconnection timing. The company’s response is to treat brownfield industrial sites — those with existing high-voltage service, industrial zoning, and legacy physical infrastructure — as a shortcut past the most expensive stages of greenfield development.
The logic is structurally coherent. A site with an active 10 MW connection, industrial crane capacity, river cooling access, and existing gas infrastructure enables module pre-fabrication and staging without competing in the interconnection queue. Module manufacturing does not require the same power density as a hyperscale compute deployment; it requires reliable, already-energized infrastructure at modest scale.
The Stokes Inc. partnership adds a BESS dimension. Integrating grid-scale battery storage into a module manufacturing campus means the site can absorb renewable intermittency, manage peak demand, and potentially participate in grid services — a configuration that turns a constrained industrial footprint into a more flexible energy asset than its nameplate capacity suggests.
Why It Matters for Global Heads of Data Center Energy
The immediate procurement implication is indirect but trackable. If brownfield industrial sites are becoming anchors for modular data center supply chain manufacturing, that changes where pre-fabricated infrastructure components originate, at what lead time, and under what utility constraint profile — variables that energy and infrastructure heads routinely model for transformer and substation procurement pipelines.
The 10 MW ceiling at Warwood is not a flaw in the model. It is a defining feature. What module manufacturing at this scale requires is an energized connection, industrial footprint, and logistics access — not the interconnection capacity a compute deployment demands. Brownfield conversions that satisfy those criteria can come to market faster and cheaper than greenfield sites queuing for utility upgrades.
Geography compounds the relevance. West Virginia sits within reach of the Northern Virginia compute corridor, one of the most capacity-constrained power markets in North America. A regional manufacturing node for modular components carries delivery timing and logistics advantages that a distant supplier cannot match, regardless of nominal production scale.
The AEP engagement is also worth reading as a structural signal. Utility capacity inquiries at brownfield sites with existing connections are often resolved faster than full interconnection requests for new loads. If AEP confirms capacity above 10 MW, the site’s development pathway changes materially. If not, the model validates a 10–30 MW manufacturing niche rather than a compute deployment play — which remains a functional supply chain role.
Forward View
Three fronts are worth tracking if this model scales. First, other brownfield industrial properties — former steel, foundry, or heavy manufacturing sites — may attract similar repositioning plays, particularly in mid-Atlantic and rust-belt corridors with aging but energized utility infrastructure. Energy teams that map regional brownfield power capacity proactively may identify manufacturing supply chain partners before the market prices that access in.
Second, the BESS integration embedded in this model has downstream implications for procurement strategy. If module manufacturers combine battery storage with their own energy management infrastructure, they become capable of more flexible delivery scheduling and can potentially absorb variable energy costs that compute deployments cannot — altering the economics of behind-the-meter storage comparisons.
Third, West Virginia’s House Bill 2014 — the Power Generation and Consumption Act — is now the active regulatory framework for this type of development in the state. How it applies to manufacturing versus compute deployment will determine whether similar brownfield conversions accelerate or encounter new friction within the state.
What Is Still Uncertain
The most consequential unknown is AEP’s response to Silicone Foundation’s power capacity inquiry. No AEP position has been confirmed or publicly disclosed, and the site’s development ceiling remains unresolved. The company’s chief officer has stated explicitly that 100 MW would require a full electrical grid rebuild in the area — theoretical, not planned.
The company is early-stage in the U.S. market, registered in Wheeling and operating with leadership distributed across multiple states and countries. Stokes Inc.’s role is named but not detailed in any public filing or contract disclosure. Production capacity, module specifications, and delivery timelines for the Warwood manufacturing operation have not been published. Until AEP responds and the company submits formal development plans, the site’s operational contribution to the broader modular supply chain remains an intention, not a confirmed capability.
One Question for Your Team
Which brownfield industrial sites within 200 miles of your highest-priority constrained compute markets carry existing energized high-voltage connections, and have any been assessed as potential module manufacturing or infrastructure staging locations?
Sources
- Theintelligencer — Company Official: Warwood Facility Will Build Modules for Data Centers (Link)
