Southwest Power Pool has quietly built the most complete set of large load rules of any US grid operator. In January the Federal Energy Regulatory Commission accepted SPP's High Impact Large Load study process, known as HILL, and the companion High Impact Large Load Generation Assessment, or HILLGA. On June 5 it approved a further piece, the Conditional High Impact Large Load Service, which SPP calls CHILLS. Together they give data centers in SPP's 17-state footprint, which runs from northern Texas to Montana, a path that trades guaranteed delivery for speed.
The problem SPP described
SPP's filing, summarized in FERC's January 14 order, sets out why the existing processes failed large loads. Requests from customers above 100 MW peaked in 2022 at nearly 11 GW, then fell to about 6 GW in 2024. SPP suggested the decline could be due partly to its own inability to study and interconnect large loads quickly enough to meet developers' timelines. In other words, the queue shrank not because demand fell but because customers went elsewhere. SPP argued that its market mechanisms were designed for incremental and predictable load growth, and that large loads want generation built on timelines far shorter than typical interconnection studies allow.
HILL and HILLGA: pairing load with generation
SPP submitted the HILL and HILLGA tariff revisions on October 24, 2025, and FERC accepted them effective January 15, 2026, subject to a compliance filing within 30 days. The design pairs a new large load with shovel-ready generation and studies them together, rather than forcing the load through the delivery point process and the generator through the general interconnection queue separately.
Commissioner Rosner's concurrence describes the central innovation as a new interconnection service, Load Limited Resource Interconnection Service, which limits the amount of output the paired generator can inject to what the grid can handle given the associated load. He wrote that the combined processes would allow paired generation and load to connect in less than half the time typically required to clear SPP's generator interconnection queue. He also drew parallels with the transmission services FERC had directed PJM to create for co-located load in December 2025, noting that while the mechanics differ, the practical outcome is similar: new generation needed to serve new large loads can connect faster without pushing costs onto other customers.
Rosner encouraged other transmission providers to take note of SPP's approach. He also pointed out that SPP achieved it through a section 205 filing of its own design, rather than waiting for the outcome of the federal large load rulemaking that the Secretary of Energy had initiated in October 2025.
CHILLS: interruptible service as a bridge
HILL and HILLGA speed things up, but they still require generation and network upgrades to be in place before a large load gets firm service. CHILLS addresses the gap. According to reporting on the June 5 order, the service lets data centers and other very large users connect for up to seven years on non-firm, interruptible transmission service while they line up firm generation and network upgrades. The service takes effect on July 1, 2026.
The condition is strict. SPP can curtail CHILLS customers when the transmission system is constrained or under emergency conditions, and the reporting notes that the tariff sets no cap on how often or how long curtailment can last. A data center taking the service must be prepared to drop load on SPP's command. The term is a countdown. The customer must reach permanent firm service within the window or lose the service when it expires.
For developers, that changes the calculation. Instead of waiting years for firm service, a data center can energize in months, provided it has a way to ride through curtailments, whether through on-site generation, batteries, or shifting computing work to other sites. It can then convert to firm service once its paired generation and upgrades arrive. SPP is also developing a separate Price Adaptive Load Service for customers willing to be curtailed in response to prices.
How the pieces fit
Read together, the three elements form a sequence. A developer files a large load request and, through HILLGA, nominates generation to serve it. The paired generation is studied under the limited injection service, which avoids triggering the full set of network upgrades a standalone generator would need. While that generation is being built, CHILLS lets the load connect on an interruptible basis. Once the generation and any required upgrades are in service, the load converts to firm service. Each step shifts risk to the developer in exchange for time, which is the commodity AI developers value most. The design also protects existing customers, who neither wait behind the new load nor pay for generation built to serve it.
Why interruptible service matters now
The appeal of non-firm service is a function of the alternative. In much of the country, a large load seeking firm service faces years of studies and upgrades, and in some regions, capacity shortfalls that make utilities reluctant to commit at all. Research from Duke University's Nicholas Institute in 2025 estimated that SPP could integrate about 10 GW of new load with average annual curtailment of 0.5% of the load's maximum uptime, using headroom in the existing system. CHILLS is a way to put that headroom to work, though without the curtailment caps that the Duke analysis assumed.
The absence of a cap is the main risk for customers. A facility built for continuous operation that faces open-ended curtailment needs dependable backup or a workload that can move. Hyperscalers with fleets of data centers in multiple regions are best placed to use the service. Single-site developers may find the uncertainty harder to finance.
A model for others
FERC's June 18 show cause orders asked every RTO and ISO to address, among other things, new transmission services for flexible large loads and processes to study generation serving co-located or electrically proximate loads. SPP already has both. Its approach is likely to become a reference point as the other operators respond, partly because it was designed by the operator and its stakeholders rather than imposed, and partly because it has been tested at FERC twice in six months.