Feature Articles

The difference shows that a year after passage of S.B. 6, a Texas law requiring large loads pay the costs of connecting to the grid and implement demand response protocols, expectations among data center developers continue unabated.

Areas of Texas outside of ERCOT are exempt from these requirements. The rush to build data centers in west Texas mirrors that of other regions.

The addition of these large loads, now apparent because of regional transmission planning requirements, prompted ERCOT to tell regulators in April of “concerns” over preliminary load forecast values.

The preliminary long-term load forecast, a required filing to the Public Utility Commission of Texas, projects 278,003 MW of total demand in ERCOT by 2029 and 367,790 MW by 2032.

“For comparison, ERCOT’s current all-time peak demand is 85,508 MW,” Chad Seely, senior vice president of ERCOT, said in an April 15 filing to the commission. “ERCOT would prefer to consult with commission staff to evaluate whether it is appropriate to seek adjustment of the forecast.”

ERCOT projects that this summer’s peak load will occur in the range of 90,500 MW to 98,000 MW.

The high end of the forecast comprises a full buildout and ramp up of large load sites.

The long-term forecast based on input to transmission and distribution providers called for a peak of about 112,000 MW.

“It seems unlikely that the ramp of existing sites and new sites will cause the peak to exceed 98GW by this summer,” ERCOT told the commission.

The North American Electric Reliability Corporation, an industry standards and monitoring body, said in its most recent reliability assessment that summer demand in ERCOT is forecast to be down by 4.6% from the same period last year.

“The lower demand forecast is the result of updated load modeling that reflects the observed behavior of load during peak periods and more demand response from large computational loads,” the assessment said.

The ERCOT forecast differences widen in the coming five years. Next year, the forecast peak tops 100,000 MW, but the long-term load forecast is more than 50% higher. By 2031, the forecast peak is 108 ,000 MW and the long-term load forecast is three-and-a-half times greater.

The large loads include not only data centers but cryptocurrency miners, industrial uses and oil and gas facilities.

The commission docket on the issue does not yet suggest further action on the matter.

A consensus exists that the multiple forecasts are inaccurate, and iterative steps by ERCOT and the commission will correct the process, said James Hickey, a lobbyist for data centers in Texas.

“The point of S.B. 6 was to winnow out duplicates, and it goes a long way,” he said, while adding that ERCOT got more submissions after the bill’s adoption than before.

The batch process will by definition help make some of this happen,” Hickey said. “If loads do not make it into the batch, they don’t get on the list for the next three years.”

The ERCOT board this week approved the batch process, a plan to would replace most large load interconnection studies with a system-wide approach that evaluates many applications at once.

Several prerequisites are being considered for large load customers to be eligible to participate in the batch study process, said Christian Nagel, senior director of power supply and production at Rayburn Electric Cooperative.

“These requirements are intended to reduce speculative requests and ensure that only serious projects enter the study queue,” Nagel said.

He explained that the proposed requirements include demonstrating site control, posting financial security, providing engineering plans and project information, paying an interconnection study fee and satisfying various disclosure and certification requirements.

Rayburn has already been proactively requiring many of these assurances, Nagel said.

Outside of ERCOT, a west Texas data center developer told MAGNIFYI that his region is ideal.

Jose Ortega, principal at FOPermian, said the problem of providing remote power for drilling rigs has long been solved in the Permian Basin. “There is an entire industry that has been solving what the data centers want,” he said. We have been managing complex, high capex projects for a long time in wind, solar, oil and gas.”

He suggested that utilities and retail energy providers examine behind-the-meter (BTM) solutions for large loads.

“When your customers are coming to you and want to buy hundreds of megawatts in one location, that is a very secure long- term contract,” Ortega said. “Not a lot of BTM deals have been done, but every independent power producer is saying ‘hold my beer.’’’

Retail energy providers are well-positioned because they have customers and can source energy, he said. Aggregating supply sources which have a BTM solution can be the edge to win a deal.

Ortega said that is because many energy off-takers and independent power producers hope BTM assets, especially combined cycle turbines and large reciprocating engines, eventually can be repurposed for the grid.

“If we are doing a BTM deal, sooner or later our client is going to want the grid,” he said.

Hickey said uncertainty for data center power supply to get in the transmission queue will create increased willingness to develop assets behind the meter.

“In the short run, there will be a very real bottleneck of what gets on,” he said. “In the long run, it will be better if it ends up on the grid.”

North of the Permian Basin, one of the biggest proposed data centers, the Matador project being developed by Fermi, Inc., near Amarillo, appears to be floundering.

Fermi proposed a data center campus in the Southwest Power Pool control area with 11 GW of nuclear and natural gas generation.

Company executives said on a March conference call that they had not secured a tenant.