Elizabeth K. Whitney is a managing principal at Meguire Whitney, a government relations firm. An earlier version of this article was published in the This Week in Hyperscaling newsletter.
Earlier this month, Gov. Greg Abbott of Texas made major waves by pausing new data center connections to the grid until regulators audit the 474 gigawatts currently in the ERCOT interconnection queue. At first, it sounded like Texas was following in New York’s footsteps and halting data center construction in the face of intense consumer backlash.
But this policy choice is an interesting one — especially in a state that had been among the top leaders in both data center development and in updating its energy policies to accommodate that growth. And while Texas’ power demand could double in the next five years, the state has been at the forefront of balancing speedy development with consumer protection with policies like SB6, which not only pioneered curtailment standards but also established the requirement that any existing generator get PUC approval before leaving the market to serve data center load.
That’s why it’s peculiar that Gov. Abbott’s approach isn’t actually a moratorium on data centers. It’s a moratorium on grid connections.
If it’s the grid connection that needs to wait, it’s not hard to surmise that Gov. Abbott’s pause is likely to shift a great deal of data center development into private, off-grid arrangements. That trend has already been underway nationwide, although it has been challenging to determine how many facilities will truly forego a grid connection and which will build a data-center-plus-power-plant and try to connect later.
The off-grid gamble
Full self-generation has its benefits, particularly with respect to speed. With no utility in the picture, concerns about cost-shifting and regional reliability can be put aside. But it can also be expensive, and inflexible: A campus running entirely on its own gas turbines has no backup if a unit malfunctions, no way to sell surplus power, and no path to cheaper or cleaner supply if the grid eventually has capacity to spare. That’s a big reason why many experts think islanding is a way to get built fast, not a way to operate forever.
That’s also why it’s likely that most of these campuses will come looking for a grid connection at some point — for reliability, for economics, or because a good market opportunity opens up. But delaying that moment isn’t necessarily a stroke of policy genius. If huge data center campuses arrive already built, already running, and negotiating from a position we can’t fully predict right now, then facilities the grid could have planned for years in advance instead show up with none of the planning runway, as well as a bunch of engineering that probably needs to be retrofitted.
Cost-causation discussions get harder. Overbuilding is virtually guaranteed. Questions about who bears responsibility for meeting emissions requirements or portfolio standards heat up. Multiply these issues by however many campuses take this path, and the state has spent a year building rules for orderly interconnection while quietly incentivizing the disorderly version instead.
A parallel pathway
Policymakers have already started laying the groundwork for what it looks like to develop powered land for data centers. But they can avoid giving mixed signals by anticipating the buildout of off-grid capacity now, instead of thinking they’ve hit “pause” while the industry reads it as “fast-forward.”
Even for off-grid facilities, there are utility concerns. For one, if utilities are going to be responsible for part of the data center buildout, we need to update the rules governing that process faster.
Cutting utilities completely out of the picture isn’t a solution either. No matter where a data center locates, it’s in some utility’s service territory, and that utility has an obligation to serve. But if more states follow Virginia’s lead in refusing to allow utilities to socialize profit on facilities built for a single customer (and I hope they do), they will need to consider what role the utility should play in scenarios where the grid connection comes last instead of first.
Here’s an idea: Let a developer building an off-grid facility pay the local utility a consulting fee to help design that project for a clean future interconnection — right-sized switchgear, a site chosen with grid capacity in mind, clarity on environmental and permitting landscapes — the same way it would pay any other engineering consultant. For an investor-owned utility, that’s straightforward revenue. For a consumer-owned utility, it’s a way to offset the cost of staff time and put downward pressure on rates instead.
Texas’ interconnection pause could be short-lived. It’s possible the governor takes a deeper look at Batch Zero and realizes that what he meant by “auditing the queue” simply will mean adding some data points to a more expansive and technical review already underway. But in the meantime, the uncertainty injected into the regulatory process may have already spooked some projects out of view.
Elizabeth K. Whitney is a managing principal at Meguire Whitney, a government relations firm. She has more than fifteen years of federal policy experience, focusing on energy policy, and writes the This Week in Hyperscaling newsletter. Her government relations practice centers on energy markets, environmental regulation, climate change, and nuclear power. The opinions represented in this contributed article are solely those of the author, and do not reflect the views of Latitude Media or any of its staff.
