Attorneys
Related Practices
In the past several years, the development of massive data centers to fuel artificial intelligence and other cloud-based computing has been one of the driving forces in the energy industry and in the broader economy. While the construction of data centers continues to accelerate and the size of the data centers continues to grow, some potential risks to their business model appear on the horizon.
At the same time, recent news articles highlight a trend in which large tech companies are moving toward using more debt to finance their data center transactions, as opposed to relying on their own cash flow. The combination of risk and debt should prompt some reflection on the bankruptcy risks inherent in data center deals.
It is difficult to overstate the speed of growth of the data center market, globally and particularly in the U.S. The International Energy Agency, or IEA, has estimated that global data center consumption will grow from 415 terawatt-hours in 2024 to 945 TWh in 2030, a 128% increase in just six years.[1]
In 2024, data centers accounted for approximately 1.5% of the world's electricity usage. That usage has increased by 12% per year since 2017, which is more than four times higher than the rate of increase in total electricity consumption.[2]
The IEA estimates that data center demand will comprise approximately 10% of global electricity demand by 2030, at which time the services provided by data centers will be a $517 billion market.[3]
Part of the reason for the rapid expansion of data centers has been governmental support for the market. Both the Biden administration and the Trump administration issued executive orders seeking to encourage construction of AI data centers.[4]
In addition, both the U.S. Department of Energy and the Federal Energy Regulatory Commission have been active in attempting to foster the interconnection of large electrical loads to the U.S. electric transmission system.[5] In varying ways and to varying degrees, state governments and their regulators have also been active in the area.
Despite the proliferation of data centers and the governmental support, more business risks are emerging with the passage of time. The most obvious risks relate to the interplay between consumers and factors like increased electricity prices or electricity supply constraints caused by weather — either of which could become volatile quickly under the wrong circumstances.
Beyond those concerns, the data center boom causes other issues further up the supply chain, hindering the ability for critical supplies to keep pace with demand.
Recently, there has been stress on the supply chain for a wide variety of components, from gallium, a metal used in computing chips, supplied almost exclusively from China, to turbines for gas-fired generation facilities — currently experiencing multiyear delays.
Increased demand on water and other resources necessary for cooling the data centers also strains the existing supply and infrastructure. In 2023, data centers used 17 billion gallons of water for cooling and another 211 billion gallons indirectly for power generation.[6]
Projects in Indiana, Wisconsin and Arizona have recently been rejected due to concerns over aquifer depletion and effects on drought.[7]
The aggregate effects of AI's heavy use of electricity, water and other resources could lead to the formation of movements against data centers.
An Oct. 18 article in Barron's noted: "In Virginia, home to more data centers than anywhere else on earth, at least 42 groups have sprung up to oppose more data centers in their towns."[8] If those movements reach critical mass, the politics surrounding data centers could shift quickly as well, potentially resulting in unfavorable legislation or regulation.
Similarly, an Oct. 22 article in The Wall Street Journal quoted a former official in the Antitrust Division of the U.S. Department of Justice to predict an increased focus by antitrust enforcers on data centers' effects on power markets.[9]
Against that backdrop, we see hyperscalers increasing their reliance on debt to finance the growing number and size of their data center projects. In a Nov. 8 article titled "Debt Has Entered the A.I. Boom," The New York Times cited a McKinsey & Co. estimate that $7 trillion will be needed by 2030 for data centers to meet their projected demand.[10] That article continued:
Now, the tech giants are turning to financing maneuvers that may add to the risk. To
obtain the capital they need, hyperscalers have leveraged a growing list of complex
debt-financing options, including corporate debt, securitization markets, private
financing and off-balance-sheet vehicles. That shift is fueling speculation that A.I.
investments are turning into a game of musical chairs whose financial instruments
are reminiscent of the 2008 financial crisis.
It is most likely that the business risks noted above remain on the horizon, as data centers learn to navigate a changing landscape. Likewise, it is also likely that the hyperscalers use debt and financial instruments in mostly rational and responsible ways.
Nevertheless, there is at least some chance that the various factors combine in a way that ultimately leads to AI data centers in bankruptcy — not necessarily the large tech companies themselves, but potentially their subsidiaries and joint ventures set up for the purpose of creating and operating data centers.
Early forms of data centers were primarily bitcoin miners, which could tolerate some degree of interruption in electricity. The shift to AI data centers translates to a greater need for uninterrupted power supply. That supply may come from the grid via utility service, or it may come through negotiated power purchase agreements with electricity generating or marketing companies.
The power purchase agreements may be physical, in which the buyer takes title to the actual electricity through physical delivery, or virtual, in which the buyer obtains electricity from a utility and then pays a fixed price to a power generator that is delivering power to the market.
Often, the structure of the power supply arrangement places one or more intermediaries between the supplier and the data center customer. The intermediary may be a utility that doesn't own generation capability or an electric cooperative, as well as other market participants that are required by law — such as a qualified scheduling entity under Texas law.
The intermediaries obtain power, sometimes from a dedicated generation source, and sell it to the data center. If there is a dedicated generation source, it may be from any variety of fuel sources, including renewables, fossil fuels or nuclear energy.
In addition, many data centers have backup power available from battery systems, generators, or interconnection to the grid.
The size of the data center's load will often be the largest load served by a given power supplier. Because the outsized load acts as a force multiplier, the contracts between the supplier and the data center customer should reflect the degree of risk taken by the supplier.
A default by the data center could lead to massive market exposure and stranded costs for the supplier.
In the event of a bankruptcy filing by a data center customer, the range of outcomes for the other parties in the transaction extend from acceptable to catastrophic. In the former category, it is not hard to imagine a scenario in which a data center is forced to file for bankruptcy protection, but external economic conditions remain relatively normal.
In such conditions, the data center would likely be able to obtain debtor-in-possession, or DIP, financing, to fund its operations and its reorganization efforts, which funding would likely facilitate an ongoing relationship with the supplier and other parties. Despite some initial anxieties, such parties may experience relatively little distress in the bankruptcy
process.
On the other hand, market forces could lead to a much more turbulent bankruptcy process. A tightening of credit could lead to a scenario in which the data center cannot obtain adequate DIP financing.
Perhaps other stakeholders would supply additional funding, but perhaps not. In a scenario in which cash is especially scarce, fights with the power supplier will have elevated importance. Alternatively, a period of unusually high electricity prices could lead to higher stakes for both the supplier and the data center.
The supplier may be able to terminate its supply contract because of the bankruptcy filing, but it may not. If it cannot terminate the contract and it needs to source part of its power from the market, then it may be facing greater exposure than it ever anticipated.
The bankruptcy court may impose a deposit requirement on the data center to provide some additional security for the supplier, but whether the amount of the deposit would be sufficient to address the supplier's risk is uncertain.
Any contested issue before the bankruptcy court would be complicated by certain realities. For one, there is a risk that a court would perceive the balance of the equities to favor the data center to the extent that it had no other viable power source.
In business bankruptcy cases, a bankruptcy court generally views its role as fostering the reorganization of the debtor. If an issue arises in which ruling for the power supplier would negate the debtors' chances of reorganization, the court may try to find a way to keep those chances alive.
For another, the power supplier's right to appeal an adverse ruling may be illusory if the commercial effects of the ruling are especially harsh.
Legally, a party has a right to appeal an adverse ruling in the form of a final order. But if the effect of the order is to force the supplier to purchase large amounts of electricity at high market rates for an extended period, the supplier could face its own liquidity or solvency issues long before it can obtain an appellate ruling.
As with most contracts, the best way to address bankruptcy risk in data center deals is on the front end, with cautious and purposeful contractual language. The agreements should apportion risk carefully, to reflect the commercial realities that would exist in the event of a customer default (or a looming default).
Relevant contractual protections should address not only obvious issues like performance security and termination rights, but also more granular issues like change in law and material adverse changes. The first and most important step, however, is for the supplier and other parties in the chain to recognize that the combination of risks involved in data center transactions aggregate in such a way that bankruptcy is a possibility that must be taken seriously.
This article first appeared in Law360 on January 5, 2026 and is republished here with permission from the publication.
Mark Sherrill is a partner at Chamberlain Hrdlicka White Williams & Aughtry. The opinions expressed are those of the author(s) and do not necessarily reflect the views of their employer, its clients, or Portfolio Media Inc., or any of its or their respective affiliates. This article is for general information purposes and is not intended to be and should not be taken as legal advice.
[1] Zachary Skidmore, IEA: Data Center Energy Consumption Set to Double by 2030 to 945TWh (Dec. 8, 2025, 3:35 PM), http://datacenterdynamics.com/en/news/iea-datacenter-energy-consumption-set-to-double-by-2030-to-945twh/.
[2] Giulia Petroni, AI Boom to Fuel Surge in Data Center Energy Needs, IEA Says, Wall St. J., Apr. 10, 2025.
[3] Skidmore.
[4] See Exec. Order 14,179, 90 Fed. Reg. 8741 (Jan. 23, 2025); Exec. Order No. 14,141, 90 Fed. Reg. 5469 (Jan. 14, 2025).
[5] See Interconnection of Large Loads to the Interstate Transmission System, Notice Inviting Comments, Docket No. RM26-4-000 (issued Oct. 27, 2025).
[6] Avi Salzman, A Growing Nimby Movement Could Upend the Data Center Boom, Barron's, Oct. 18, 2025.
[7] "Press Release: The U.S. AI Data Center Fresh Water Bottleneck Could Become a National Security Crisis," Dow Jones Newswires, Oct. 21, 2025.
[8] Salzman.
[9] Edith Hancock, Data-Center Power Use to Become Major Antitrust Issue, Wall St. J., Oct. 22, 2025.
[10] Ian Frisch, Debt Has Entered the A.I. Boom, N.Y. Times, Nov. 8, 2025. See also Jonathan Weil, AI Meets Aggressive Accounting at Meta's Gigantic New Data Center; Favorable Treatment Off the Balance Sheet Hinges on Some Convenient Assumptions, Wall St. J., Nov. 24, 2025.