Are all new loads created equally?
That is the question facing policy makers, grid operators and regulators as the tsunami of new potential electricity demand from data centres hits electricity grids across North America (and globally). The International Energy Agency estimates that data centres will drive almost 10% of global electricity demand growth to 2030, but more than 20% of demand growth in advanced countries. The impact that data centres will have on the grid will be acute, as the IEA notes that “[h]istorically, data centres have been highly concentrated in spatial terms, posing significant challenges to local grids given their substantial power draw.”
Increasingly, we are witnessing a range of different responses from all levels of oversight on how to both support integration of new large loads – and the economic benefits they can provide to the broader economy – with the potential rate impact and other cost allocation issues they can impose on existing ratepayers[1]. The responses range from requiring direct political approval where new loads are allocated increasingly scarce capacity to regulatory mechanisms designed to prevent the risk of stranded assets. The stranded asset risk arises from a grid-wide overbuild of new capacity to meet demand from loads that may have an uncertain operating life, including long-term contracts for large loads, minimum demand charges and other stipulations.
The following examples provide an overview of a number of different mechanisms that regulators and policy makers have adopted or proposed to manage both the increase in load from data centres and the cost implications for existing and future customers. This list is not exhaustive, as many jurisdictions are introducing new tariffs, specific rate design changes or other stipulations in response to the rapid increase in demand from data centres. Nonetheless, the list highlights a number of common responses seen across North America.
Power Advisory Commentary
There is no doubt that the world is entering an age of significant electrification, both to support energy security and decarbonization targets. Supporting growing electrification requires system expansion and potentially increased reliability investments – while also requiring a more proactive buildout of the grid to encourage fuel-switching and other load growth objectives. Yet, significant system expansion does not come without risks.
The potential for a rapid increase in demand from an emerging sector can be both a blessing and a curse for utilities and ratepayers, depending on how regulators, policymakers and utilities/system operators respond to that growth and protect ratepayers from the risk of stranded assets or other cost allocation concerns.
On one hand, many electricity grids across Canada and the United States have experienced years of falling or flat demand growth. This has often resulted in grids that were “overbuilt” from a capacity or energy perspective, resulting in upward pressure on rates to recover a growing amount of costs over a flat or declining demand base – essentially more costs needed to be recovered over a similar or shrinking number of units sold. Adding new demand to the existing grid – when there is spare capacity – allows for those costs to be spread over a wider rate base and can result in a lower per unit cost of electricity.
Conversely, a rapid increase in the build-out of the electricity grid – both from a generation and wires (transmission and distribution) perspective – can result in a material increase in the total cost of the electricity system in each province (or state). If the new load driving grid-wide expansion is fully allocated those costs, then existing or small-volume consumers face little impact in the form of higher rates. But if those new loads are diminished or disappear altogether over the medium and long-term (i.e. greater than 10 years), then the costs that were incurred to support new load growth will flow to ratepayers that did not cause the system-wide expansion and/or be passed on to a utility’s shareholders. If the cost implications become severe enough, it is likely to result in a difficult intervention by the energy regulator to protect the utility’s remaining customers from rate and service-related repercussions, while also maintaining the financial viability of the utility.
Putting cost implications to the side for a moment, an emerging and rapidly evolving sector could bring significant economic benefits to provincial economies – many of which are facing economic headwinds from the ongoing trade dispute with the United States. In particular, access to computing power is being increasingly being viewed as an essential piece of a modern economy for security purposes and to maintain competitiveness. The IEA has noted that AI data centres offer the potential of significant productivity gains to the electricity system through optimization of operation, maintenance, planning and resilience (including demand response and the integration of renewables) as well as the economy more broadly through optimizing industrial production processes and building design. For example, the IEA observed that “AI in transport can enhance vehicle operation and management, which could cut energy consumption by up to 20%.” This positive feedback loop could lead to a further increase in electrification and potentially the significant productivity gains in the economy that have been expected that would further justify the rapid system expansion.
Many of the investments required to expand the grid – including new capacity and transmission and distribution assets – are typically recovered over a long period, such as 40years or more. If large load customers, such as data centers or other new and emerging large customers, are not obligated to cover a portion – or all – of the grid-wide costs incurred to serve them regardless of their final connected capacity, there is a potential that these costs will need to be recovered from existing customers. In a province like Ontario where many fixed generation costs are recovered through the Global Adjustment (GA), which is predominantly allocated to small-volume customers, the risk of stranded assets is even more severe.
One approach that is being both considered and implemented in a number of jurisdictions is the requirement that new large loads sign a long-term contract of 15 years for their demand and capacity. This policy can maintain the “non-discriminatory access” policy that Open Access Transmission Tariff (OATT) designs established as a critical part of deregulation to eliminate the negative economic impact of unequal treatment or short-term political intervention. The policy also provides some form of protection to small-volume consumers. Additionally, the policy may allow large load customers to “sell” the contract to another party if their individual load projections fall short of realized demand (either through increased efficiency or stunted growth). The 15-year – or whatever length policy makers view as reasonable – contract has the potential to create a secondary, liquid market that both offers protection to existing customers and new large loads alike. But this is merely one example of how the investment and productivity potential of data centres can be integrated in a way that does not impose significant financial risk on existing customers. As noted above, there are a variety of different tools and policies being considered and implemented.
In any case, while policy makers and provincial, state and federal lawmakers are quick to highlight the potential economic benefits of rapid investment and demand from data centers (and other large loads), the current cost allocation and rate design in many jurisdictions must be updated or reviewed. These reviews should focus on providing some form of protection from the risk of stranded assets without unduly curtailing the potential upsides for local communities and the broader economy. Data centers, in particular, add a new complexity of how the current regulatory and policy environment incorporates the potential benefits of AI data for all economic sectors into the assessment of potential ratepayer impacts.
The future is inherently uncertain and there will be uncertainty on the benefits that data centers and other large industrial loads will ultimately bring. The predictions range from data centres providing a transformational productivity gain for the entire economy to a short-term boom-to-bust. Regardless, what is clear is that policy makers cannot let not-built-for-purpose regulatory constructs and/or process-related inefficiencies to choke off the potential benefits from an emerging sector. Stagnant demand over the last decade largely removed incentives to investigate innovative and adaptive regulatory and policy frameworks to address rapid load growth. The demand environment is changing and reform is needed now to better manage the balancing act of ensuring long-term economic growth while prudently managing ratepayer impacts.
In our view, the opportunity and challenge points to the pressing need to consider new processes and approaches. Without critically examining the status quo and need for, and impact of, regulatory changes, it is not clear how a jurisdiction can effectively manage to integrate large load additions and the associated acceleration in demand growth in a way that will protect ratepayers and support system expansion at the pace required.
[1] In addition to society, environmental, land-use and other community concerns raised with new large loads which are not discussed in this client note.