California Reopened Non-Residential Rate Design, and Utility Billing Systems Will Set the Timeline

The California Public Utilities Commission instituted Rulemaking 26-04-009, the Order Instituting Rulemaking on California Advanced Electric Rate Design, on April 9, 2026, and issued it the following day. The proceeding is assigned to Administrative Law Judge Perez-Green and Commissioner J. Reynolds. Its stated object is to “update residential and non-residential rate structures and their underlying cost inputs.” A prehearing conference was held on June 18, 2026.

The non-residential half of that sentence is the part with consequences for behind-the-meter storage. Commercial demand charges are the single largest determinant of whether a battery installed on a customer’s side of the meter pays for itself, and California is where that market sits: the state accounted for 77 percent of US commercial and industrial storage installation volume in the first quarter of 2026.

The venue shift. Marginal cost and revenue allocation for non-residential classes have historically been resolved inside General Rate Case Phase 2 proceedings, and the Commission’s own explanation for opening a standalone rulemaking is that those proceedings stopped producing a legible record. Repeated settlements of marginal-cost and revenue-allocation issues in recent PG&E, SCE and SDG&E Phase 2 cases have, in the OIR’s language, made it “difficult to assess the underlying system cost assumptions and whether the resulting revenue allocation remain equitable across customer classes.” A settlement resolves a dispute without adjudicating the inputs. Enough consecutive settlements, and the inputs stop being examined at all. R.26-04-009 moves them into a proceeding with a public docket.

The Commission also names wildfire-related cost increases among its reasons for reopening rate design. That places two questions in the same forum: how much revenue non-residential classes are asked to produce, and through which charges.

Two statutory riders. The rulemaking carries two legislative assignments that arrived independently of the rate-structure question. Senate Bill 57 (Padilla, Stats. 2025, ch. 647) requires an assessment of data-center-driven cost impacts by January 1, 2027. Assembly Bill 2109 (Carrillo, Stats. 2024, ch. 700) requires implementation of an exemption from nonbypassable and departing-load charges for industrial customers using qualifying process heat recovery. Neither is a demand-charge item, but both put large-load cost allocation inside the same record as commercial rate structure, which is where the arguments about who is subsidizing whom will be made.

The proceeding additionally inherits the unresolved demand-flexibility items left over from R.22-07-005, including whether to widen time-of-use differentials. Wider differentials increase the arbitrage spread a storage asset can capture. That item is now live in a docket that is also examining whether commercial rates accurately reflect cost of service.

Direction unresolved. The OIR commits to re-examining the cost inputs that produce commercial demand charges. It does not commit to raising or lowering them. Storage economics in California depend on demand charges being both high and avoidable, and a rulemaking of this scope can move either property. A conclusion that commercial demand charges over-recover marginal cost would compress the savings a battery produces. A shift of recovery toward demand-based charges would expand them. A shift toward fixed customer charges would move revenue permanently outside the reach of any customer-side asset, because a fixed charge does not respond to load management, efficiency or on-site generation. The recovery mechanism is therefore a more consequential variable for behind-the-meter economics than the overall level of the bill, and it is the variable the record has not yet resolved.

The billing layer. Whatever the Commission approves, the date it reaches a customer bill is set by infrastructure that rate proceedings rarely discuss. Writing in Utility Dive on August 7, GridX Chief Commercial Officer Scott Engstrom argued that legacy customer information systems built for analog meters, not regulatory appetite, are what actually gate rate design. The number of approved US electricity rates has grown roughly an order of magnitude in five years to more than 50,000. Moving a complex rate into production routinely takes 18 to 36 months.

The cost of fixing that is not marginal. Full CIS replacements routinely exceed $100 million. One West Coast utility’s billing modernization request exceeded $700 million, and regulators rejected portions of it for insufficient justification. Where rate logic lives in both the meter data management system and the CIS, it is programmed twice and can fail in twice as many places. The rate designs Engstrom identifies as blocked by this are precisely demand charges, critical peak pricing and time-of-use restructuring.

Engstrom also counts at least 13 states that now tie allowed return on equity partly to time-varying-rate enrollment. That arrangement pays a utility for enrollment outcomes while leaving the billing capability that produces them to be funded through a separate and contested capital request.

The arithmetic. The practical consequence is that an approved rate and a billed rate are different objects separated by a period measured in years. Add an implementation lag of 18 to 36 months to whatever R.26-04-009 concludes, and a revised non-residential demand-charge structure reaches a commercial customer’s bill well after the decision that authorized it. Any payback model treating a newly approved rate as a present-day input overstates near-term economics by that interval. The defensible input is the rate structure currently in production in a given territory, with pending proposals held separately.

What the record will show next. The scoping memo following the June prehearing conference is the document that determines whether non-residential rate structure proceeds alongside residential reform or waits for a later phase. Two related signals sit outside California and point the same way on cost pressure: Maryland’s Office of People’s Counsel has estimated that PJM’s proposed reliability backstop procurement could cost that state’s customers more than $500 million over 15 years, and opponents in Nevada are contesting NV Energy’s $411 million three-year natural disaster protection plan as too expensive. Those costs flow into commercial delivery rates through mechanisms that do not themselves touch rate structure. The question R.26-04-009 will answer is which charges on a commercial bill carry them, and that answer determines how much of the bill a customer-side asset can still address.

Sources