White House’s “BYOG” policy creates opportunities as 7 tech giants pledge to self-supply power for data centers

A deeper dive into the Ratepayer Protection Pledge

There are five prongs of the Pledge. First, the companies pledge to “build, bring, or buy the new generation resources and electricity needed to satisfy their new energy demands, paying the full cost of those resources whether by building, or buying from, new or otherwise additive power plants.” Further, but only “where possible,” the companies pledge to add more capacity for public availability. Hyperscale companies, like the parties to the Pledge, already face prices increasing tenfold in some parts of the United States,^[2] with some developers in 2025 willing to pay up to 50% more than just seven months prior to access power faster.^[3] Even if such higher rates stood to benefit power producers, interconnection queues (which can reach 72 months) ^[4] and labor shortages (some causing a 20% increase in labor costs) ^[5] have slowed down interconnected generation construction. Therefore, the Pledge addresses data center developers’ existing constraints and signals acceleration of the trend toward behind-the-meter power projects and joint ventures to service individual data centers or developers. Because the Pledge does not specify what generation resources are covered, nor whether the resources will be on-site or off-site, the potential for power producers to capitalize is broad, though the shape of such new deals is uncertain.

Second, the seven companies pledge to pay for “power delivery infrastructure upgrades” to avoid such costs being passed onto individual ratepayers, but only if those upgrades are required for their specific data center. With aging infrastructure and data centers only constituting part of the new demand, these companies’ commitments may help offset individuals’ rising energy bills without contributing additively to public grid improvements, especially if the companies are procuring their own power behind the meter. Some states and localities have partnered directly with data center developers to coordinate approvals and invest in infrastructure, such as LEADS in Wyoming^[6] and the Tahoe-Reno Industrial Center in Nevada^[7]. Though this prong faces significant uncertainty, it could head off local opposition from communities and help upgrade infrastructure that power producers need to get projects online.

Third, the companies pledge to “negotiate new, separate rate structures with their utilities and relevant State governments wherever they build data centers” and “pay these rates for the power and related infrastructure that are brought online to service their data centers, whether they use the electricity or not.” If the first prong is implemented and the companies generate most of their own energy behind-the-meter, it remains unclear whether (and how much) the companies would pay for power from the grid versus simply paying for infrastructure upgrade costs. States have begun to mandate similar structures. In Illinois, for instance, Chicago-area data centers must pay million-dollar premiums for large loads.^[8] A national commitment by several of the biggest data center developers could create more uniformity, allowing power providers to ink deals across states on common terms and control some of the uncertainty of ratemaking disputes.

Fourth, the companies pledge to hire “from within the local community” and establish “programs to develop relevant skills,” though the details and extent of such commitments remain undefined. These commitments could ease a shortage in skilled labor, especially linemen, affecting data centers and the supporting power infrastructure. Labor costs have risen to 20% annually^[9] in some parts of the country, and shortages became so severe that worker villages were constructed in North Dakota and Texas as data centers entered markets with interconnection capacity but shallow labor pools.^[10]

Fifth, the companies pledge to “coordinate with grid operators to contribute to a more reliable grid” and “whenever possible, make available their backup generation resources at times of scarcity to prevent blackouts and power shortages in their communities.” It remains unclear how this prong, which would cause data centers to interconnect their generation to the grid, will operate alongside the first prong, where much of the generation procured is likely to be behind-the-meter. Offtake agreements may prove lucrative for power producers selling to both utilities and data centers for peak load, though it remains unclear whether utilities will require data centers to cover the cost of such load, which could in turn squeeze margins for generators.

Conclusion

The Pledge portends new opportunities for collaboration between power producers and data center developers despite little certainty. While billed primarily as a protection for individual ratepayers, the Pledge addresses many of the existing constraints (i.e., labor, community resistance, and electricity cost disputes) that have stymied data center developers. Joint ventures, new offtake agreements, and public-private partnerships are set to emerge as the next wave of data center development comes online.