PG&E Counts 5 Flex Connect Customers and 85 in the Pipeline, With Full Capacity in 90 Percent of Hours
- Customers enrolled: 5
- Prospective customers: About 85
- Full capacity available: 90 percent of hours, first two years
- Curtailed when constrained and demand high: Under 1 percent of the time
- Energization time: 4 to 8 months, against 1 to 3 years firm
- Typical project size: 2 to 5 MW, up to 10 MW
- CPUC decision: R.24-01-018, February 5, 2026
Pacific Gas and Electric has five customers on its Flex Connect program and about 85 more in the pipeline. The utility gave those figures to Utility Dive in an article published on September 2, 2026. Flex Connect trades control for speed. A customer accepts that PG&E can cut its draw when the local feeder runs out of headroom. In exchange the customer energizes in four to eight months. A firm interconnection across PG&E’s California territory takes one to three years. PG&E says enrolled customers reached their full needed capacity in 90 percent of hours over the program’s first two years. It puts the share of time it actually had to hold a customer back at under 1 percent. That count covers only the hours when the grid was constrained and demand was high together. Two customers have finished the program after about 18 months each and moved to firm service once local upgrades caught up. Most projects run 2 to 5 MW. A few data center and manufacturing prospects reach 10 MW.
The control layer is a distributed energy resource management system PG&E built with Microsoft and Schneider Electric. It reads feeder conditions in real time. When a feeder approaches its thermal limit, the system trims the enrolled customer’s draw. When headroom returns, full power returns. EV charging sites make up most of the enrolled base and most of the pipeline. The rest is smaller data centers, advanced manufacturing plants and battery storage sites. WattEV’s Fresno truck depot shows the shape of a typical customer. That site runs 15 ports at 240 kW and seven ports above 1 MW each. It can pull up to 3.6 MW during most hours of the year.
Why It Matters
The pilot arrived before the rule, and the rule has now overtaken it. On February 5, 2026 the California Public Utilities Commission decided R.24-01-018. That order tells PG&E and Southern California Edison to offer flexible service connections as a standard tariffed product, not a pilot. The utilities had 60 days to file a joint advice letter putting that standard offer in place. They owe the Commission a cost-efficiency evaluation in January 2029. So the five customers and the 90 percent figure are not a closing scorecard for a trial. They are the evidence base a permanent tariff now sits on.
The two headline numbers do not measure the same thing. Full capacity in 90 percent of hours leaves 10 percent of hours short of it. The under-1-percent figure covers only the hours when the grid was constrained and demand was high at the same time. PG&E has not published what fills the gap between those two figures. A fleet operator sizing a depot needs that answer, because 10 percent of a year is about 876 hours.
What Happens Next
Watch the pipeline convert. Eighty-five prospects against five signed customers is a 17-to-1 ratio, and a prospect costs nothing to declare. Watch the size mix as well. The program grew up around 2 to 5 MW EV depots, and a 10 MW data center on a curtailable connection carries different risk. A depot can move charging to the night. A server hall cannot. The last thing to watch is the advice letter PG&E and SCE filed under R.24-01-018, specifically what the standard offer says about curtailment limits. The pilot ran on bilateral agreements negotiated one at a time. A tariff has to state the limit up front.
Critical Perspective
A permanent tariff now rests on five customers. At the program own typical size of 2 to 5 MW, five customers is somewhere between 10 and 25 MW of enrolled load. The CPUC took that evidence and ordered a standing product for PG&E and Southern California Edison, which together serve most of California. The pilot may well be sound. The sample behind it is very small, and the decision arrived before the two-year results it is now being credited with.
Read the graduation number carefully. Two of the five finished the program, which is 40 percent, and finishing means the local upgrade got built and the customer moved to firm service. For those two the flexible connection was a bridge across a construction schedule. That is a useful product. It is not the same product as a permanent curtailable connection, and the headline results mix the two together.
The risk sits on one side of the deal and the measurement sits on the other. The customer accepts the curtailment. PG&E defers capital it would otherwise spend and rate-base. The report the Commission ordered for January 2029 is a cost-efficiency evaluation, so it will measure what the utility saved. No filing requires anyone to measure what the curtailed hours cost the customer in missed charging sessions or lost production. A fleet operator carries that number alone.
Sources
- Utility Dive: PG&E sees rising interest in ‘customer-driven’ flexible interconnection pilot
- California Regulatory: Commission Clears Path for Immediate Energization Under New Flexible Service Connection Rules
- CPUC: Internal OpFlex Pilot Report, Pacific Gas and Electric, February 28, 2025
- Latitude Media: PG&E is laying the groundwork for flexible data center interconnection