Data Center Feasibility Study in Ohio

Fifth in the country, with small-ticket facilities

Ohio ranks 5 of 52 among states for data center loans, with 39 loans from FY2017 to FY2026 worth $17.3 million, but the class sits 22 of 39 in Ohio's own list and the location quotient is 0.79x. These are small operations: the median loan of $200,000 is 0.97x the national figure, with $100,000 to $400,000 covering most requests. Hosting, managed services and small colocation rooms fit that scale far better than a purpose-built campus.

A Huntington story centered on Columbus

SBA 7(a) carried 38 loans and SBA 504 just 1; no loan is franchised, at 0.00%. The Huntington National Bank dominates, with 21 loans worth $5.6 million, more than half of the file and an average ticket of only a few hundred thousand dollars. No other lender reaches five loans. Franklin County has 17 loans, Cuyahoga 4, Medina 3, and Hamilton, Lorain, Stark and Warren 2 each, so the Columbus area is the center of gravity.

Low cancellation, no published charge-off rate

Slippage is mild in this cell: 3.3% at the floor and 6.7% at the ceiling that includes undisbursed loans, for approvals dated FY2017 to FY2024. Lenders in Ohio have mostly seen approvals through. No charge-off figure is printed since the group of funded loans falls under the 30-loan minimum; that is a blank and not a zero. Approvals have been lumpy, from 7 loans in each of FY2017 and FY2019, down to 1 loan in FY2021 and FY2023, and up to 8 loans worth $3.4 million in FY2025.

Questions an Ohio data center study should answer

Which customers will pay, and under what contract term, comes first, since revenue from a handful of tenants decides repayment. Power price and redundancy, cooling and network carriers are the main cost lines. A study should test what happens to coverage if the largest tenant leaves. Specialised buildings have thin comparables, so the appraisal should be conservative and should show an alternative use.

In this data, colocation space is blended with hosting services, because the entry also takes in data processors who may own no building at all.