[datacenter_tag_image]
By Peter Asmus, Guest Author
https://www.linkedin.com/in/pathfindercommunications/
The United States’ electricity sector is at an important inflection point. For the first time in several decades, demand for electricity is growing at an unprecedented pace. Tremendous opportunities lay ahead—as well as daunting challenges. The latter is what has grabbed the interest of state governors, federal and state regulators, industry (both suppliers and consumers) and ratepayer advocates. The goal remains a reliable yet affordable electricity system. But how to get there?
While the electrification of homes and transportation is fostering a jump in electricity demand, the chief culprit and current focus of scrutiny is AI and data center growth. Consider this: forecasts for data center electricity consumption are projected to quadruple annually across the U.S.
The epicenter of the U.S. debate on how best to accommodate data centers in a capacity constrained power market is the Pennsylvania, New Jersey and Maryland (PJM Interconnection) control area. Its long-term forecast shows peak load growth of an astounding 32 GW between 2024 and 2030. A remarkable 30 GW of this future growth—or almost 94%—is attributed to data centers. Meeting this surge in AI-driven electricity demand will not only require new transmission-dependent generation beyond what is in the current queue but no attrition to currently planned future new power plant capacity. Even then, a shortage of electricity supply is still quite possible.
Co-location and Flexibility are the Name of the Game
Being a fan of both microgrids and flexibility resources such as demand response (DR), I’ve become intrigued by a proposal by Enchanted Rock to the Federal Energy Regulatory Commission (FERC) to encourage a new, flexible asset class: “Load Co-located with Dispatchable Generation” or “LCDG.” LCDG would effectively be a large load (such as a data center) that would have dispatchable generation on site that could be actively controlled by a grid operator such as a utility or RTO. (For more details on what is also being characterized as a “speed-to-power” concept, please see my recent white paper done in partnership with Enchanted Rock.)
This LCDG concept is being fleshed out as an option for PJM during its current Critical Issue Fast Path (CIFP) proceeding, which, among other things, seeks to treat data centers over 50 MW in size as a “non-capacity backed load” or NCBL. This concept was dropped and replaced by a new PJM proposal with a voluntary pathway where Price Responsive Demand (PRD) and DR are essentially the same but with one important difference. DR can earn capacity credits and is directly dispatched by PJM. In contrast, PRD cannot earn capacity credit and operates under provider supervisory control, yet PJM can still monitor performance in real time. The shortcoming of either one of these alternative approaches is the need for adequate financial incentives to drive adoption. Energy market revenues alone will not draw enough participation in PRD or DR.
Data centers traditionally never showed much interest in DR, due, in part, to their need for ultra-reliable electricity. In today’s new market dominated by a lack of adequate supply for new large electricity loads, the chief barrier to creative demand management is the lack of monetary rewards being offered for flexibility services—including DR—rather than the basic concept. AI itself has improved the surgical capabilities of a broad spectrum of flexibility services, moving away from an on/off paradigm to a more sophisticated throttling down of consumption based on real-time market conditions. Yet increasing payments to these large energy users for DR could threaten efforts to keep rates down for smaller customers, including residents.
Speed-to-power may serve as a sufficient incentive to attract voluntary investments in flexibility.
Instead, speed-to-power may serve as a sufficient incentive to attract voluntary investments in flexibility. If done right, a construct that gets new loads online while leveraging existing infrastructure would add new MW and MWh to the denominator of the cost allocation equation, mitigating concerns about data centers increasing costs for other electricity consumers. Ultimately, the large electricity loads that have driven fear of high energy costs could actually help keep costs flat or even reduce rates across the board. Interestingly enough, a proposal from the Data Center Coalition includes required financial commitments from utilities and other large load customers to support the buildout of new generation or to prevent retirements of existing resources.
Can PJM Serve as a Model for the Nation or the Entire World?
The speed-to-power solution would offer multiple benefits to grid planners. It could help defray transmission costs needed to serve new large loads coming online. It could also help ensure that more dispatchable generation is brought online in tandem with new loads at a time when much of the country is facing, or will soon face, a dramatic rise in resource adequacy challenges.
Data center companies are looking for solutions to get power to their facilities as quickly as possible and cannot afford to wait for a host of landmark regulations on dockets in multiple states and control areas to finally take effect. They want and need power right now.
To remain competitive globally, data centers and other large loads are seeking commercially available solutions to secure reliable power supply on schedules that are consistent with their needs. The “speed-to-power” solution encompassing both onsite generation and non-firm flexible interconnections to the grid fits the bill.
Enchanted Rock first proposed such an option in the PJM stakeholder process on large load issues and was shortly followed by a fuller “Bring Your Own Generation” pathway proposal by Data Center Coalition with support from the governors of Pennsylvania, Virginia, Maryland, and New Jersey. Enchanted Rock has had success with this approach for a new data center in the neighboring Southwest Power Pool (SPP) region under the rationale that onsite generation can serve as prime power for the data center in the near term, and then switch to backup power and grid services once the local utility is ready to serve the data center and provide interconnection.
The concept could also dovetail with proceedings at other regional control areas, including the Southwest Power Pool with its so-called conditional high impact large load service (CHILLS) proposal. The CHILLS program offers large loads willing to take a long-term curtailable transmission service to come online even though their electricity demand cannot be served by a designated grid resource or the transmission system.
Utilities such as Portland General Electric are also putting forth innovative answers to the data center growth dilemma with a pioneering program leveraging AI to find existing sources of flexibility via GridCARE, a new company pioneered by virtual power plant innovator Amit Narayan. Jigar Shah has also endorsed the same basic non-firm flexibility concept for data centers and large electricity loads.
Combining onsite generation, flexibility services and a fast-tracked and nimble interconnection process could offer models for data center deployments around the globe. Unprecedented growth in data center deployments is certainly not limited to the U.S. By 2026, data centers globally are forecasted by the International Energy Agency to consume 1,000 TWh, roughly equivalent to Japan’s total electricity consumption.
PJM and its stakeholders still have an opportunity to show the world new market-based models to maintain reliability, keep electricity affordable and allow data centers to play an interactive role in meeting the needs of all energy ecosystem stakeholders. Here’s how I would sum up the benefits of the Enchanted Rock proposals for data centers:
In short, flexibility and speed-to-power models represent the practical path forward for balancing grid reliability, affordability, and growth. By enabling large users like data centers to co-locate dispatchable generation and participate actively in grid operations, organizations can relieve transmission bottlenecks while accelerating time-to-power for critical new loads. Enchanted Rock’s approach proves that pairing onsite generation with flexible interconnections isn’t just a bridge — it’s a blueprint for how utilities and data centers can work together to build a more resilient, efficient, and globally competitive energy future.
Peter Asmus has almost 40 years of experience as a journalist, analyst, consultant and book author. This article was developed in partnership with Enchanted Rock.
This article was originally published on LinkedIn.
