Where Data Centers Get Built?
Institutional Friction and the Spatial Logic of Compute Infrastructure in the United States
- Wu, Shaoyuan
Global AI Governance and Policy Research Center, EPINOVA LLC
https://orcid.org/0009-0008-0660-8232
Description
This policy brief analyzes why large-scale data center development in the United States is increasingly shaped by institutional feasibility rather than purely technical or market-based conditions. It introduces the Infrastructure Friction Boundary (IFB) as a diagnostic framework for understanding where compute infrastructure can be deployed with lower procedural resistance and what governance risks follow from such spatial concentration.
Abstract
Large-scale data center development in the United States is increasingly shaped not by purely technical or market-based considerations, but by differences in institutional feasibility across jurisdictions. While recent public discourse often describes a southward shift of data centers, this framing overstates relocation and obscures underlying structural drivers. The brief argues that current development patterns are better understood as institutional site selection: data centers are built where governance, permitting, and utility coordination allow rapid, low-friction deployment. It introduces the Infrastructure Friction Boundary (IFB), a diagnostic framework for identifying where hyperscale compute infrastructure can be accommodated with minimal procedural resistance. Understanding institutional friction is essential for anticipating the long-term governance consequences of AI- and compute-intensive infrastructure expansion.
Files
| Name | Type | |
|---|---|---|
| Where Data Centers Get Built.pdf Full-text PDF of the publication | application/pdf | Download |
Keywords
- data centers
- compute infrastructure
- AI infrastructure
- institutional friction
- Infrastructure Friction Boundary
- IFB
- permitting
- utility coordination
- energy infrastructure
- governance capacity
- land-use governance
- infrastructure-governance asymmetry
- United States
Subjects
- {'subject': 'AI infrastructure governance'}
- {'subject': 'Data center siting'}
- {'subject': 'Institutional feasibility'}
- {'subject': 'Compute infrastructure policy'}
- {'subject': 'Energy and utility coordination'}
- {'subject': 'Public policy and governance'}
Recommended citation
Wu Shao-Yuan (2026), Where Data Centers Get Built? Institutional Friction and the Spatial Logic of Compute Infrastructure in the United States, Policy Brief No. EPINOVA–2026–PB–04, Global AI Governance and Policy Research Center, EPINOVA LLC, https://doi.org/10.5281/zenodo.18592482. DOI: To be assigned after Crossref membership approval.
APA citation
Wu, S.-Y. (2026). Where data centers get built? Institutional friction and the spatial logic of compute infrastructure in the United States (Policy Brief No. EPINOVA–2026–PB–04). Global AI Governance and Policy Research Center, EPINOVA LLC. https://doi.org/10.5281/zenodo.18592482. DOI: To be assigned after Crossref membership approval.
Alternate identifiers
| Scheme | Identifier | Description |
|---|---|---|
| EPINOVA publication number | EPINOVA–2026–PB–04 | Publication identifier printed in the PDF |
| DOI | 10.5281/zenodo.18592482 | Zenodo/DataCite DOI printed in the PDF recommended citation |
| DOI | 10.5281/zenodo.18592481 | Earlier DOI value from early ORCID-derived metadata record; retained for reconciliation |
| ORCID put-code | 205251164 | ORCID Public API record identifier from early metadata |
| Series number | Policy Brief No. EPINOVA–2026–PB–04 | Policy Brief series number |
Related works
| Relation | Identifier | Type | Description |
|---|---|---|---|
| Related EPINOVA policy brief on structural governance and AI infrastructure risk. | |||
| Related working paper on data center siting and the spatial logic of AI infrastructure. | |||
| Related working paper on AI infrastructure, governance lock-in, and siting politics. |
References
No references listed.