A giant new load can promise the moon. The tariff fight starts when it changes its mind. This is Power Grid Daily. A massive Georgia contract, a reliability miss already on the books, and a Texas transmission case that just got very awkward. Here's RTO Insider:
Georgia Power can move forward with a 3.2-GW contract to power an OpenAI data center after satisfying state regulators’ concerns about the potential impacts on the utility’s existing ratepayers. Following a nearly two-week delay to hammer out the final details, the contract cleared the Georgia Public Service Commission without objection at an Aug. 27 meeting.
Georgia PSC approved 3.2 gigawatts for OpenAI and secured the clause that matters: if OpenAI exits early, Georgia Power can’t send the lost revenue to ordinary customers. That puts a real contractual backstop behind the forecast. Good protection—on paper. But phased service starts in 2028, so transformers, substations, and interconnection work are already on the clock. The commission approved a 25-year power deal; it didn’t conjure up the equipment. Georgia Power also has to publish a portfolio-performance report every six months, plus a public contract summary within 10 days. With the full OpenAI agreement redacted, those reports are where we’ll see whether the promised $180 in annual residential savings for 2029 survives contact with the actual load. A $20 billion facility near Savannah can be a terrific customer. But if its ramp from 2028 to 2032 slips, it can become a very expensive half-built assumption. Those public reports need megawatts actually taken, not just megawatts reserved. From RTO Insider:
SPP stakeholders have endorsed a priority tariff change to address tight capacity margins that have resulted in three energy emergency alerts this summer in the RTO’s new Western balancing authority area. The Markets and Operations Policy Committee approved the proposed change ( RR801 ) with 78.24% support during a virtual meeting Aug. 25. Three stakeholder working groups also have approved the measure.
SPP expected 6,480 MW available in the West BAA and has averaged 4,700 since April. A tariff vote won’t close a 1,780-MW hole. It does show who was counting capacity that isn’t showing up. RR801 at least puts money on the distinction. Asset owners short of adequate supply pay, and owners with adequate megawatts share the proceeds. After an EEA 3 on July 20, an EEA 2 four days later, and another alert Aug. 9, SPP had reason to act. And Yasser Bahbaz says half the West BAA’s 9,204 MW of registered nameplate is unavailable on a typical day. Nameplate is a very comforting number right up until the control room needs actual output. The 78.24% stakeholder approval matters because this is an interim market rule. It won’t put a capacity resource online by next summer. Watch whether the charges change day-ahead behavior—or simply put a price on a shortage everyone can already see. RTO Insider writes:
MISO South’s network of load pockets has not been operating to the one-day-in-10-years reliability standard and could face more capacity shortcomings by 2030, despite utilities’ generation additions and transmission upgrades. That’s the crux of MISO’s recently conducted load pocket risk assessment, the first step in the grid operator’s quest to draw up a long-range transmission plan for MISO South.
MISO is saying the problem is already here: all four South load pockets missed the one-day-in-ten-years standard in 2025/26. Amite South alone showed a retroactive 914-megawatt winter shortfall. And the regional total for Local Resource Zone 9 can clear the benchmark while a pocket inside it is short. Power doesn’t teleport through a constrained interface because the spreadsheet says Louisiana and southeast Texas balance in aggregate. The long-range plan starts with measured loss-of-load risk: 0.4 to 1.1 days a year in Amite South, up to 0.8 in SETEX. Planned generation and wires may improve that by 2030, but MISO has already identified pockets where the margin stays tight. Which puts a hard edge on that 2030 date. A transmission plan can choose projects this year; transformers, substations, outages, and construction crews still have to arrive before the shortages do. Here's RTO Insider:
Based on forecast capital investments, the New England transmission owners expect ISO-NE ’s Regional Network Service rate to grow from $177.63/kW-year in 2026 to about $237/kW-year in 2031, an increase of about 33%. The companies filed with FERC a 3.6% increase to the current rate for 2027 to reflect new projects, increased revenue requirements and their expectation for a higher return on equity ( ER20-2054 ).
New England’s Regional Network Service rate is forecast to reach $237 per kilowatt-year in 2031, up from $177.63 this year. That’s a 33% increase in the bill, mostly tied to asset-condition work. Seventy-five percent of the $2.31 billion through 2027 goes to replacing aging gear. A 101-mile, 230-kV rebuild at 345-kV specifications is a major construction program with a very real supply chain behind it. The 2027 filing seeks 3.6% immediately, then assumes a 6.6% annual pace after that, partly on a higher return on equity. Burnham calls the forecast indicative; ratepayer advocates will call it the opening exhibit in every formula-rate challenge. I’d rather see the rebuild than watch a deteriorated line fail in a bad winter. But FERC should make the owners show the commissioning logic, the equipment plan, and why the 345-kV scope is the least-cost durable fix—not just approve a five-year curve. San Saba News & Star writes:
The PFD is the ALJs recommendation to the PUC. The PUC will make the final decision at a Hearing that will likely be scheduled near the end of September. The SOAH ALJs are telling the PUC they should deny the Bell County East and Sand Lake 765kV CCN applications.
A 765-kV line is a massive piece of steel-and-transformer infrastructure, and SOAH says Oncor and LCRA never proved Import Path 2 was needed. The load forecast has to establish that need before you build the physical system. The judges called ERCOT’s Permian Basin Reliability Plan forecasts unverified, overstated, flawed, and unsubstantiated. That goes straight to the CCN standard: proponents need an independently defensible need case, not a planning label attached to it. The process has its own problem: roughly 1,400 landowners across 400 miles of newly added route links were caught up in public-meeting violations. For a project sold as reliability planning, the paperwork is looking remarkably unreliable. The PUC gets the final call near the end of September. If it overrides this recommendation, it should explain exactly what evidence establishes need for Bell County East and Sand Lake—and who bears the cost if those projected loads fail to materialize. If you’re finding Power Grid Daily useful, take a moment to subscribe or leave a review wherever you’re listening. Reviews help other people find the show, and they help us keep these briefings coming.
We’re watching for Georgia Power’s public summary of the OpenAI contract, due to the Georgia PSC within 10 days, and for the Texas PUC’s expected late-September hearing on the Bell County East and Sand Lake 765-kV CCN applications.
Links to every story are in the show notes if you’d like to dig deeper. That’s Power Grid Daily for today. This is a Lantern Podcast.