Gridlocked

Market Memo

Columbus

A rising Midwest market with a credible land-and-power story, but utility timing and infrastructure delivery will determine scalability.

Gridlocked's directional read-through on this market's ability to convert AI infrastructure demand into buildable capacity.

Overall

79

Constraint

Utility delivery timing

Saturation

Medium

Executive Summary

Columbus is emerging as a credible Midwest AI infrastructure market because it offers land availability, business climate, connectivity, and a power story that could support campus development at lower saturation than tier-one markets. The key question is whether utility delivery and infrastructure timelines can keep pace with demand.

Market Scorecard

Overall

79

Power

80

Fiber

73

Land

84

Water

78

Policy

77

Demand

75

Saturation

Medium

0–100 directional Gridlocked framework scores. Higher means a stronger relative buildability signal, not guaranteed investment upside. Saturation is shown as a qualitative constraint signal.

Constraint Stack

Primary constraint

Utility delivery timing

Saturation risk

Medium

A rising Midwest market with credible land and power story if delivery timelines hold.

Saturation is a constraint signal — it reflects how much friction incremental growth faces, not whether the market is good or bad on its own.

Bull Case

Land availability, Midwest connectivity, business climate, power potential, and lower saturation than tier-one markets create a credible runway for early site-control strategies.

Bear Case

Utility delivery timing, construction capacity, and less mature hyperscale ecosystem depth can limit how quickly the market scales.

Real Estate Read-Through

Columbus may reward early site control where utility commitments, infrastructure timelines, and campus-scale development plans are credible before the market fully institutionalizes.

Real Estate Market Data

Data center fundamentals

H2 2025 / Midwest

CBRE North America Data Center Trends H2 2025

Inventory

520 MW

Vacancy

3.10%

16.1 MW available

Available

16.1 MW

Under Construction

310 MW

Preleased

84.00%

of capacity under construction

Net Available Pipeline

49.6 MW

Asking Rent Midpoint

$143

$/kW/month

Power Delivery Timeline

18–24 months

Primary Constraint

Fiber

Additional metrics

Net Absorption

118.4 MW

Asking Rent Low

$135

Asking Rent High

$150

Prior Period Rent Midpoint

$138

Rent Growth

3.60%

Secondary Constraint

Permitting

Major Demand Driver

Enterprise

Key Takeaways

Vacancy is 3.10% with 16.1 MW available against 520 MW of inventory.
310 MW is under construction and 84.00% is preleased, leaving 49.6 MW of net available pipeline.
Accelerated hyperscale campus site footprints breaking ground and expanding beyond traditional clusters into New Albany, Lancaster, and Pataskala sub-markets.

Market data sourced from CBRE North America Data Center Trends H2 2025 inputs entered into Gridlocked's market data workbook.

Source ID: CBRE_H2_2025_001

Key Risks

AEP delivery timelines slipping past announced schedules
Anchor hyperscale commitments failing to materialize
Construction labor competition from other Midwest projects
Infrastructure commitments that lag site-control activity

Public Market Read-Through

Public Exposure names that touch this market's constraint stack — potential beneficiaries and constrained exposures. Exposure can be positive, constrained, regulated, second-order, or mixed; this is read-through, not a buy list.

AEP

American Electric Power

69

American Electric Power is a regulated utility exposure to the grid delivery side of the AI infrastructure buildout. If data center and industrial load growth increase demand for transmission, distribution, and generation investment, AEP sits close to the utility infrastructure required to serve that load. The exposure is meaningful, but it is regulated and execution-heavy rather than a simple AI winner story.

ETN

Eaton

73

Eaton is one of the cleaner supplier-side ways to express the power bottleneck thesis, but Gridlocked treats the exposure as equipment-cycle sensitivity rather than automatic upside.

GEV

GE Vernova

69

GE Vernova sits close to the power infrastructure side of the AI buildout. It does not own data centers, but it provides generation, grid, and electrification solutions tied to the question of whether large-load demand can be served reliably. The exposure is meaningful because AI infrastructure ultimately needs generation capacity, grid equipment, and transmission reliability.

VRT

Vertiv

71

Vertiv is one of the cleaner public-market ways to express the physical data center infrastructure bottleneck. As AI workloads increase rack density and cooling intensity, the limiting factor is not just compute demand but whether facilities can deliver power, reject heat, and operate reliably. Vertiv sells into that constraint through power, thermal, integrated infrastructure, and service offerings.

PWR

Quanta Services

73

Quanta is a public-market way to express the grid construction side of the AI infrastructure bottleneck. If data center demand forces utilities and power markets to upgrade transmission, distribution, substations, and interconnection infrastructure, Quanta sits close to the physical work required to turn load growth into deliverable power.

Data Status

Market memo content and scores are directional Gridlocked framework inputs based on public infrastructure, utility, site-control, water, demand, and saturation indicators. They are screening signals, not investment recommendations.