Every day, we read about data centers. Their impulsive growth because of artificial intelligence (AI); their energy demands which stress local power providers and local ratepayers; and the increasing pushback from politicians who are seeking to impose the marginal energy cost of a data center on their owner/operators. Recent months have witnessed several nine-figure asset securitization transactions secured by data center rental payments. These things are HUGE!
Therein lies an opportunity for equipment financiers: what the Monitor has termed a “gold rush” for equipment finance. Data centers require energy and support equipment to power and cool them.
Data Centers Come in Many Sizes
Hyperscale data centers include those that companies like Amazon Web Services and Meta have installed in key “metro” areas to furnish 24/7/365 responses to AI demands.
Enterprise data centers are those supporting a single organization such as a bank, insurance company or major retailer.
Edge data centers are smaller facilities geographically close to the network and data sources.
Colocation data centers are facilities which include multiple users under one roof.
Each of these data centers contain “energy and environmental” equipment needed to supply and modulate the power and cool them to keep them humming without interruption.
Internal energy equipment consists of uninterruptible power supplies, backup generators, power distribution units, transformers, switchgear, cooling equipment and so forth. This equipment could comprise 75%-80% of the total cost of a data center.
Cooling equipment consists of water-cooling equipment and internal pipes to send cool water to where the servers are located (and to re-cool the water after it has cycled through the building); without constant cooling, the servers are likely to break down and disappoint any users who are dependent on reliable data center output. HVAC equipment complements the cooling apparatus to maintain a level internal temperature, particularly in locations such as Chicago, Dallas and Phoenix, which experience temperature extremes and wide climate variations. All of these kinds of equipment typically will constitute fixtures and (depending on local real estate law) hence constitute part of the realty. But Uniform Commercial Code Section 2A-309(c)(4) permits a lessor of fixtures to have priority over the owner or mortgagee of the realty if (a) it is a purchase money lease, (b) the lessor makes a fixture filing within 10 days after the equipment becomes a fixture, and (c) the lessee is in possession of the realty. This means that lessors of the equipment can enjoy a greater ability to remarket their equipment than do lessors of the entire data center itself.
External power generation equipment is equally important, including alternative energy facilities as well as traditional power generation equipment. Increasingly, larger data centers must provide their own dedicated power sources, as their voracious appetite for electricity has stressed the local utility’s ability to provide electricity and provoked resistance from other local rate payers who have witnessed their electricity bills jump because of the utility’s need to construct increased generation capacity and the inclusion of that construction cost in each customer’s rate base. The current administration has suggested expediting the permitting process for these standalone power facilities.
Tax Profile of a Data Center
Although a data center building generally can be depreciated straight-line for 39 years, much of the related equipment, internal and otherwise, can be depreciated separately and may be written off 100% when placed in service. Assuming 20% of the overall cost is the building and 80% is equipment, the first-year aggregate tax depreciation could be 81% of the total data center cost! Thus, there can be substantial tax benefits to owning a data center or related equipment.
Financing
Given its huge costs, a hyperscale data center is financed similarly to a real estate project: the rentals and any lessee-provided residual value guarantee can be securitized based on the underlying tenant’s credit rating; using a longer (10 to 20-year) maturity with a balloon payment equal to the projected residual value; the lessee guarantees all or part of the residual value of the asset and hence is the tax owner of the asset under this synthetic lease structure. There are also variations of this structure.
A synthetic lease is a lease for book purposes but a loan for tax purposes. Under ASC 842, the lessee capitalizes the present value of the payments plus the present value of the expected residual value guarantee payment. As an example, a 15-year lease with a 75% residual financed at 5% that is partially guaranteed by the lessee may only be capitalized at perhaps 65% of the asset cost, if the lessee concludes they will owe nothing against the guarantee. Meta recently financed a $48 billion group of hyperscale data centers and assumed NO payment would be made against the residual value guarantee!
Most hyperscale leases are financed by investors which tend to be larger banks or equipment suppliers who can afford to invest $75-100 million in a single transaction. We are also seeing the introduction of the use of “fund” money for the financing.
Other data centers are financed similarly to any large piece of equipment such as a vessel or large aircraft. The structure could be a synthetic lease where the lessee owns the data center for tax purposes. If a tax-oriented lease is desired where the lessor is the owner, the challenge is that larger enterprise data centers would likely be leveraged with third-party debt and leveraged lease accounting is no longer acceptable under ASC 842.
Conclusion
The opportunities to finance data center equipment are substantial. So far, publicity has focused on large scale securitizations involving credit-tenant leases of the entire data center. But, as revealed in a recent Equipment Leasing & Finance Foundation report, financing of energy and power equipment, for both data centers and otherwise, remains the #1 growth area for equipment finance.
