SBA 7(a) Data Center Feasibility Study in California

The top state in a small class

California ranks 1st among states for SBA 7(a) data center loans, with 103 loans and $53.2 million gross. Within California, though, the class is only 23 of 39, so California leads the program while the class is small within the state. The median loan, $250,000, is 1.25x the national median, and the central loans lie from $122,650 to $375,000. No brand is involved: the franchise share is 0.00%.

BayFirst National Bank leads with 9 loans. Next, with 8 each, come Newtek Bank, then Northeast Bank, then Wells Fargo; Bank of America, National Association has 7. No lender dominates, and the counts only reflect today's holders. Los Angeles County supplies 23 loans, San Diego 12, Orange and Santa Clara 10 each, and Alameda 6. Annual volume peaked at 20 loans and $13.5 million in FY2024, and the dollar totals bounce widely, with $9.4 million in FY2021 against $1.3 million in FY2018.

Floors

Cancellations account for at least 6.3% of approvals, or 7.6% counting undisbursed loans. In the funded cohort of 44 loans, charge-offs have reached at least 6.8%, with 4.5% inside the first five years; unresolved loans may push both higher. None of these differs from California overall and data centers nationally by the required margin, yet a loss floor of this size deserves attention.

Facility or hosting business

Colocation buildings and hosting services cannot be told apart, because the industry code also covers data processing and hosting firms. A California study should say which business is being financed, a facility or a service. For a facility, it must test power supply and cost, cooling, network connectivity, seismic and fire design, tenant concentration and contract length. For a hosting firm, customer churn and equipment refresh matter more. The cost estimate must separate building from specialized equipment that depreciates quickly, and debt service coverage should be run with the loss of the largest customer. An appraiser will need leases and utility contracts to support value, and a lender should expect the study to state which customers, contracts and equipment would still be there if the largest tenant left.