The data center industry has become one of the biggest supporters of new renewable projects, with tens of billions of dollars committed to renewables over the coming decade. Over the next two years, corporate procurement is expected to drive more than 40 percent of renewable expansion in the US, and year after year, the largest corporate purchasers of renewables are Big Tech, including Amazon, Microsoft, Google, and Meta.
Over the past five years, nearly all of this corporate renewable procurement has been executed via Power Purchase Agreements (PPAs) in deregulated energy markets. In a PPA, a corporate purchaser promises to purchase the power and Renewable Energy Credits (RECs) generated by the renewable project for the next 10-15 years. This firm revenue commitment allows the renewable developer to finance the construction of the new renewable project, enabling new renewables onto the grid.
While the data center industry, renewable developers, and the financing community are now a well-oiled machine for executing PPAs and using these contracts to support the construction of new renewable projects, there are new headwinds starting to impact renewable projects and corporate customers.
These new headwinds arise from evolving sustainability metrics and renewable industry challenges, and may lead to big changes in the types of structures and contracts that the data center industry will use as it continues to decarbonize its own energy usage and support new renewable projects.
How to measure sustainability impact
Much of the data industry has aligned emissions measurement and reporting frameworks with the Greenhouse Gas Protocol (GHG Protocol). The GHG Protocol, spearheaded by the World Resources Institute and the World Business Council for Sustainable Development, has provided a standard for measuring corporate GHG emissions for over 20 years.
During this time, the GHG Protocol has provided two simple ways for measuring electricity-related carbon emissions. The first way is the “location-based” method, where the average utility or grid emissions per MWh are simply multiplied by the electricity used by the data center.
The second way is the “market-based” method, where purchases of renewable energy (including REC purchases) can be used 1:1 to offset the electricity used by the data center.
So, for example, if a data center uses 1,000 MWh in a year in Arizona, the data center can purchase 1,000 (bundled or unbundled) RECs to eliminate its electricity-related emissions. Basically, the data center uses electricity in one part of the grid, purchases renewable energy in another part of the grid, and these offset each other.
However, as corporations have become more sophisticated, renewable energy has become a larger fraction of all generated, and more granular grid data has become available, the world has begun to look at better ways to measure the sustainability impact of a renewable energy purchase.
For example, under the “market-based” framework, a data center could consume 1 MWh of power in a coal-heavy state in the middle of the night, and then offset that energy usage with 1MWh of power produced in a solar-heavy grid in the middle of the afternoon.
In this scenario, the 1MWh of power that was used by the data center may have a much bigger carbon impact than the carbon “avoided” by the 1MWh of power produced in the solar-heavy grid, and so, even though the GHG Protocol shows zero emissions associated with this purchase using the market-based method, there may, in fact, be emissions due to the lack of alignment (both in time and space) between generating and consuming electricity.
The GHG Protocol is going through a revision process, and there is now a movement afoot to modify the GHG Protocol to accommodate the concept of 'emissionality': That is, what are the actual emissions impacts associated with the production of renewable energy, in order to derive a more accurate approach to emissions reductions.
If emissionality were to become part of the GHG Protocol, renewable projects located in cleaner parts of the grid will have a lower sustainability impact than renewable projects located in high emissions areas. While this new standard will drive new renewable investment into the dirtiest grids, these standards could have a negative impact on companies that have already executed PPAs for renewable projects in ”clean” grids, as they watch the value of their PPA investments erode.
Emissionality will also make it more complex for those in the data center industry to execute their next PPA.
Under the market-based method, a MWh is a MWh is a MWh, and a corporate purchaser could be sure that the MWh would continue to usefully offset electricity consumption during the lifetime of a PPA. However, through the lens of emissionality, the sustainability impact of MWhs changes over the life of the project. The nation’s grids are rapidly decarbonizing, with gigawatts of renewable generation in interconnection queues across the country.
And, as grids get cleaner, the emissions impact of renewable projects located in those grids is diluted. The first solar project in a fossil-heavy grid will have a high impact on emissions, but once the 1,000th solar project arrives, the emissions impact of that first solar project will drop significantly. This reduction and uncertainty in the long-term sustainability impact of a PPA may create additional headwinds for executing PPAs today.
Things are getting harder for renewable developers
These last few years have seen a number of headwinds for renewable developers, including long and growing interconnection queues, tough permitting environments, and supply chain volatility.
These challenges have made renewable projects riskier from a timing and pricing perspective, and developers are asking the data center industry to share in that risk. Data center professionals are being asked to become renewable market experts, to be able to understand, evaluate, and potentially mitigate these risks with renewable developers, while PPAs are becoming more challenging to negotiate and execute.
The risks above affect pricing and timing for a renewable project before it’s complete. As energy markets have become more volatile over the past few years, renewable developers are also increasingly asking for corporate purchasers to protect them against energy pricing volatility over time that is specific to their project’s specific location on the grid.
The data center industry is thus being asked to take on a financial risk that can vary over the lifetime of that asset. The energy pricing of a specific location on the grid can vary based on new projects coming online, new transmission lines getting built, and other generation capacity being retired. The complexity of energy market forecasting makes it challenging even for experts to get right, and this is well beyond the expertise of a typical data center owner.
Gone are the days when a renewable developer and corporate purchaser could agree on a 15-year fixed PPA price two years before a renewable project gets built. This makes it challenging for both parties to a traditional PPA, and as PPA negotiation timelines have drawn out in a rapidly changing market, buyers and sellers are increasingly wasting time in never-ending cycles of negotiations.
The Future of PPAs
Although long-term PPAs are likely to continue to be popular for the world’s largest companies, which can leverage large teams and internal expertise to adapt, the data center industry should also consider changing the way they green their power supply in response to these headwinds.
First, in response to potential emissionality changes to the GHG Protocol, data center owners should consider procuring renewables closer to their data center footprint. By bringing generation and load closer together, it will be less likely that the emissions impact from the renewable project will diverge from the emissions associated with their load over time. Locating generation and load together will also help with other potential sustainability requirements (such as 24/7 matching) lurking in the wings.
Second, the data center industry and renewable developers should explore ways to shorten PPA terms (e.g., to 7-9 years) or narrow PPA scope to unbundled RECs, while ensuring that the renewable projects can still be financed. Both of these changes may increase the near-term cost to purchase renewable energy, as financiers often apply a significant discount to non-contracted cash flow after the PPA term.
However, given the risk that a REC may devalue over time and the potential downside associated with a corporate customer taking on long-term energy risk, the data center industry may be better off taking a shorter tenor contract now and re-evaluating procurement priorities more frequently.
Renewable projects in the United States are also eligible for an investment tax credit worth 30+ percent of the project cost, which these projects often cannot fully take advantage of (since they often do not have sufficient tax liability). One of the features of the Inflation Reduction Act last year was to make these tax credits transferable, which will simplify the process of transferring these tax credits and should make it much easier for data center owners to partner with renewable developers to use these tax credits.
Data center owners can pair a shortened or more narrowly scoped PPA with these tax credit investments to provide additional financing support/investment to help finance new renewable construction.
Finally, for all of the reasons above, the data center industry should more than ever prioritize green tariff programs with their local utilities to green their supply. These programs can be used to match renewables directly with corporate load, to ensure that they do not take on additional energy pricing risk, and to create an investment multiplier where data centers are built as new renewable projects are built nearby.
As the PPA landscape shifts, it is important that leaders in the data center industry not fall into familiar patterns by default, and instead keep an open mind for all the ways they can meet their sustainability goals.
The data center industry can continue to support the renewable industry and further its decarbonization goals without creating friction internally by taking on more and more risk that it's not prepared to evaluate or offset.