On October 9, 2026, the Federal Energy Regulatory Commission rejected PJM Interconnection’s plan to let new large loads connect without matching new generation if they accepted being curtailed first in emergencies. The Interim Resource Adequacy Service, or IRAS, died on jurisdictional grounds: FERC found that a federal tariff dictating which retail customers get cut, and setting up a compensation framework it read as a retail rate, crossed the line between wholesale markets, which FERC governs, and retail service, which the states govern. The diagnosis — a capacity shortfall driven by data center growth — survived. The tool did not.
Four months earlier, in the middle of the country, FERC had accepted a tool aimed at the same problem. On June 5, 2026, the commission approved the Southwest Power Pool’s Conditional High Impact Large Load Service, known as CHILLS: non-firm, conditional transmission service that lets data centers and other large loads connect faster and be interrupted when transmission is constrained, for up to seven years while firm service is arranged. The order, issued as 195 FERC paragraph 61,196 in Docket No. ER26-1323-000, called CHILLS a just and reasonable and not unduly discriminatory approach to interconnecting new large loads expeditiously while maintaining reliable operation of SPP’s system. It took effect July 1, 2026.
What CHILLS is
Start with the ordinary rule CHILLS suspends. Under SPP’s tariff, a transmission customer taking firm service must have sufficient generating resources designated to serve its load. That requirement is the friction: the resources a new data center needs may not yet be online, and the network upgrades needed to deliver firm service can take years to build. The load is ready. The paper is not. Under the old rules, the load waits.
CHILLS lets the load stop waiting. A qualifying large load can take non-firm, conditional transmission service using available transmission capacity while the designated resources come online or the network upgrades are completed. The service is implemented through a new Part VII of SPP’s tariff and a Schedule 15, with pro forma service and operating agreements modeled on SPP’s existing transmission service agreements. SPP filed the proposal on February 10, 2026, amended it on April 29 after FERC staff issued a deficiency letter in March, and won acceptance effective July 1, subject to a compliance filing within 30 days. The condition in “conditional” is the whole bargain: SPP may interrupt CHILLS service when the transmission system is constrained or under emergency conditions.
The ladder: HILL, HILLGA, CHILLS
CHILLS did not arrive alone. It is the third rung of a framework SPP built in stages, and the earlier rungs explain why the third one fit.
The first rung is HILL, the High Impact Large Load process. A HILL is a new commercial or industrial load, or an increase in such load, at a single site connected through one or more shared points of interconnection or delivery points, meeting a size threshold that depends on voltage: 10 megawatts or more if connected at 69 kilovolts or below, or 50 megawatts or more if connected above 69 kilovolts. Electric storage resources are excluded from the definition. SPP filed the HILL process in October 2025, and FERC accepted it in January 2026. HILLs register under the tariff, follow a dedicated study framework, and face stability and ride-through study requirements sized to their impact on the system.
The second rung is HILLGA, the High Impact Large Load Generation Assessment. Where HILL studies the load, HILLGA studies the generation designated to serve it — new generation or storage specifically identified for and limited to serving a nearby HILL — in a concurrent 90-day process. The studies are narrower than standard interconnection studies because they assess only the impact of the facility serving the identified load. SPP has said the streamlined HILLGA track applies when the associated generation is within two substations of the load. CHILLS is what happens when the studies are done but the firm service is not yet ready. HILL defines the load. HILLGA studies the supply. CHILLS bridges the gap between the study’s answer and the day the upgrades are built and the resources are online.
The seven-year clock
The most important number in CHILLS is seven. Conditional service lasts up to seven years: long enough to cover the construction of network upgrades or the completion of designated resources, short enough that a load cannot park on conditional service indefinitely and avoid ever paying for firm service. A conditional service with no end date would be a way for a large load to enjoy grid access without ever funding the upgrades its presence requires — cost shifting by another name. A seven-year term converts the bargain into a bridge with a far shore. The load gets speed; the system gets a date certain by which the load must hold firm service, with the designated resources and network upgrades that firm service implies.
CHILLS does not conjure generation into existence, and it does not build a single tower. The upgrades the studies identify still have to be built by transmission owners, on construction timelines measured in years, against the same workforce and supply-chain constraints that bind every other grid project. What CHILLS buys is certainty, early — a firm answer about what it will take to connect, and service in the meantime.
Why it survived: transmission, not retail
FERC’s objection to IRAS was not that curtailing data centers is wrong. Commissioner David Rosner’s concurrence to the IRAS rejection said the opposite: in an emergency, data centers not serving critical functions should be shut off before ordinary customers. The objection was to who decides, and through what instrument. IRAS, as filed, appeared to place obligations directly on a class of retail customers and to require electric distributors to curtail that specific class — a retail rate by another name, in FERC’s reading — and it would have curtailed firm load at the earliest stage of an emergency, ahead of the voluntary demand response that exists precisely for that purpose.
CHILLS never approaches that line. The distinction, drawn crisply by analysts since the IRAS order landed, is between two kinds of interruption. Conditional transmission service is interrupted when thermal or voltage limits on the transmission system are exceeded — when the wires themselves cannot carry the flow. A curtailment mandate is triggered when there is insufficient generation to meet load — when the system is short of supply. The first is a transmission matter, squarely within FERC’s jurisdiction over interstate transmission service. The second, when imposed on retail customers by a federal tariff, is a retail matter reserved to the states.
Three design choices keep CHILLS on the right side of that line. First, CHILLS is a transmission service, offered under the transmission tariff, with interruption keyed to transmission conditions. It does not sort retail customers into curtailment classes or tell distributors whom to cut in a supply emergency. Second, it is elective: a load chooses conditional service in exchange for faster access, rather than having curtailment priority imposed on it as a condition of connecting. Third, CHILLS carries none of IRAS’s secondary defects: it does not lock in capacity accreditation for favored resources, and it does not exclude a class of load from a market’s demand curve, distorting the price signal.
What FERC still asked of SPP
None of this means FERC considers SPP’s work finished. Thirteen days after accepting CHILLS, on June 18, 2026, the commission issued a show-cause order against SPP — 195 FERC paragraph 61,213, Docket No. EL26-68-000 — one of six tailored orders issued that day to each of the regional grid operators under its jurisdiction. The SPP order preliminarily found the tariff unjust and unreasonable in four respects: it does not sufficiently address evaluation of alternative transmission technologies in the large-load study process; it does not sufficiently mitigate the risk of cost shifting among transmission customers, including through transparency into network upgrade assignment and a pro forma cost recovery agreement ensuring large-load customers bear the risks and costs of the upgrades they trigger; it does not provide clarity on the rates, terms, and conditions applicable to co-located loads; and it does not include flexible transmission service for large loads willing to limit their use of the system.
The order adopted SPP’s HILL definition for its own purposes, gave SPP and its transmission owners 60 days — until August 17, 2026 — to justify the status quo or file tariff revisions, allowed up to 90 days of abeyance to work through the stakeholder process toward a filing under section 205 of the Federal Power Act, and directed an informational report within 30 days on how SPP intends to ensure adequate generation will be available to serve existing and new large loads. Being the model did not exempt SPP from the same homework FERC assigned everyone else.
PALS: the next rung
SPP is not standing still, either. The next piece is the Price Adaptive Load Service, or PALS, which SPP has described as pending before FERC. Where CHILLS bridges a load to firm service, PALS would go further: what SPP calls true connect-and-manage, in which loads are not matched with generation or transmission at all, and instead face significant curtailment risk as the price of connecting. The concept resembles ERCOT’s connect-and-manage approach for generators that agree to curtailment to get online — except applied to loads, with the curtailment risk borne openly by the customer. Until FERC acts, PALS remains a proposal, not a service.
What it means for other regions
The template is now visible, and other operators are moving toward it. PJM, in the wake of the IRAS rejection, has told FERC it intends to make filings in the separate large-load show-cause proceeding by early to mid-November, including new transmission services for flexible large loads. The most plausible shape of a revival of the IRAS idea is a voluntary version: a non-firm service a load chooses in exchange for faster or cheaper connection. That is CHILLS, with a PJM accent. The difference between a mandate and a service is the difference between rejection and acceptance, and PJM’s next filing will test whether the commission agrees.
MISO’s approach offers a useful contrast within the voluntary family. Its proposed 120-day paired study would let a load larger than 200 megawatts and nearby generation in the same local resource zone be studied together, in exchange for a 15-year pairing commitment. CHILLS asks less — no 15-year lock, no pairing requirement — but gives less too: conditional service, interruptible, with a seven-year expiration. MISO wants the load to bring its own supply and stay paired. SPP wants the load to accept interruption while its supply and upgrades catch up. Both are voluntary. Both keep the obligation in the transmission tariff rather than in a retail curtailment order.
The limits of the model
CHILLS deserves one note of caution, because models acquire a glow that outruns their evidence. The service is months old. There is not yet a track record showing how often conditional service is actually interrupted, how many loads convert to firm service within the seven-year term, or whether the compliance filing FERC required resolved the conditions attached to acceptance. The deeper limit is one CHILLS shares with every interconnection reform: it accelerates access to the grid, not the grid itself. Conditional service is a way to use existing capacity more aggressively by allocating its scarcity transparently. It does not add a megawatt of generation or a mile of wire. The seven-year clock is only as good as the construction that happens under it. If the upgrades stall — in permitting, in supply chains, in cost-allocation disputes — the loads on conditional service will face the same cliff the clock was designed to prevent, only later and all at once.
The bottom line
The IRAS rejection and the CHILLS acceptance, read together, draw the map for the next decade of large-load interconnection. The curtail-first idea is not dead; Rosner’s concurrence all but invited states to adopt it through retail orders, and voluntary versions of it will return to FERC as transmission services. But the federal path runs through the transmission tariff, not the retail meter. A load that volunteers for interruptible transmission service, keyed to thermal and voltage limits, with a term that forces it toward firm service, is a bargain FERC will accept. A tariff that assigns curtailment priority to a class of retail customers in a generation emergency, with compensation set federally, is one it will not. SPP built that bargain first, in stages, and FERC accepted it — then pointed the other operators at it and handed SPP its own list of unfinished work. The model that works is not a loophole. It is a service with a price, a trigger, and a deadline. The regions that copy it will need all three.