CAISO would count home battery exports toward demand response performance, but an aggregation still may not sell power
Whether a household battery is a load or a generator decides who may sell, to whom, and on what terms, and this document answers it for California's wholesale demand response models: still a load, and a load for reasons of network modelling rather than reasons of desert. The distinction keeps being mistaken for a compensation dispute. What actually changes here is smaller than the trade coverage reported and more useful: energy that households already delivered was being discarded from the performance calculation because of a measurement convention, and aggregators were underpaid for work already done. The export right that advocates asked for was explicitly deferred. The argument worth following is the double-compensation one, because it is the same objection the state legislature has just told the CPUC to settle.
On 8 July 2026 the California Independent System Operator published the Track 1 draft final proposal of its Demand and Distributed Energy Market Integration initiative, titled "End-Use Customer Exports in Demand Response Performance Measurement". It does one narrow thing. It removes the settlement requirement that a scheduling coordinator set a customer's load meter data to zero during any interval in which that customer exports, so that authorised exports from a behind-the-meter battery are counted when the performance of the demand response resource is measured.
What it does not do is let the aggregation sell power. The proposal keeps the resource-level export limit, and states the position without hedging: individual customers inside a demand response aggregation may export if their utility has approved it, but "the aggregation as a whole may not be a net exporter."
The rule being fixed
California's wholesale demand response models, proxy demand response and reliability demand response resources, are load curtailment products under the ISO tariff. A resource may only reduce down to the load of the customers registered in it. Once there is no remaining load to curtail, it cannot provide more demand response energy.
Until now the measurement rule was stricter than that principle required. Because exported intervals were zeroed out, a battery that pushed past its household's own consumption had the excess discarded from the performance calculation entirely. The DDEMI working group's problem statement, reproduced in the proposal, puts it plainly: performance "is artificially capped at levels reflecting conservative estimates of site load", leaving behind-the-meter storage energy unused during events and unavailable to the market.
Track 1 closes that gap and nothing wider. Eligibility is limited to retail customers already approved to export onto the distribution system; for utilities under California Public Utilities Commission jurisdiction, a Rule 21 Export Interconnection Agreement satisfies the requirement. The ISO's Demand Response Registration System gains attributes marking which service accounts have exports that count. Utility distribution companies keep their existing power to approve or reject a customer registration, and disputes about that remain with the local regulatory authority rather than entering the ISO's dispute resolution process.
What the trade coverage reported, and what the document says
Utility Dive reported on 5 August that the proposal would treat DER aggregators as discrete resources, allowing them to export power within CAISO load zones until their net load hits zero, and quoted Brian Turner, senior director for western United States regulatory affairs at Advanced Energy United, estimating it could draw "upwards of 2 GW" of behind-the-meter capacity into wholesale demand response. Turner called it "a small accounting change [that] could significantly change the battery market in California."
The accounting-change description is accurate. The export description is not. The word "discrete" does not appear in the draft final proposal, and on aggregation-level export the document says the opposite: it "does not propose to allow DR resources to export beyond the resource level." Allowing exports up to the sub-load aggregation point is listed among the topics deferred to future DDEMI phases, alongside deliverability reform and device-level metering. Advanced Energy United, CALSSA, Sunrun, Tesla, Voltus, VGIC and the joint demand response parties all asked for that expansion, according to the stakeholder feedback section. They did not get it in Track 1.
The ISO's reason is a modelling one rather than a philosophical one. Demand response aggregations are represented in the Full Network Model as load inside a sub-load aggregation point, spread across network nodes using load distribution factors that approximate where demand sits. Let the aggregation become a net exporter and those same factors would be used to place injections, which the ISO says may misrepresent where the power actually enters the network.
There is already a route for an aggregation that wants to behave like a generator. It is DERA, the ISO's FERC Order 2222-compliant model, and it requires entering a generation interconnection queue. The proposal points to it explicitly.
The fight inside the document
The substantive objection came from the state, not the utilities. CPUC Energy Division staff and the Public Advocates Office argued that the change would pay the same exported kilowatt-hour twice: once as a retail export credit under net energy metering or the net billing tariff, and again as a wholesale energy payment. California's investor-owned utilities asked the ISO to address the concern.
The ISO rejected the characterisation. Its argument is that the two payments are for different things. Retail export compensation is set ex ante through the avoided cost calculator and requires nothing of the customer beyond exporting. Wholesale demand response compensation is earned ex post, only when the resource is economically selected or dispatched and delivers measurable incremental performance against a baseline. On the ISO's reading, the presence of forward wholesale price projections inside the avoided cost calculator does not convert a retail valuation methodology into a wholesale payment.
Why it matters
The recurring question about a home battery is whether it is a load or a generator, and this document answers it for California's wholesale demand response models: still a load, and a load for reasons of network modelling rather than reasons of desert. That distinction is worth holding onto, because the constraint is persistently mistaken for a compensation dispute. A household can be paid for what it stops consuming. Being paid for what it sends is a different regulatory object, and the ISO's answer is that the object already exists and has an interconnection queue attached to it.
The narrower reform is also the more useful one, which is why the gap between the document and its coverage matters. Measuring a resource honestly is worth more to a household with a battery than a theoretical export right it cannot access, because the discarded energy was real and the cap on it was an artefact. An aggregator whose measured performance was pinned to a conservative estimate of site load has been underpaid for work already done.
The double-compensation argument is the one to watch, because it is about to be handed to a different body. SB 913, passed by the California legislature in August 2026 and awaiting the governor's signature, instructs the CPUC to establish conditions for capacity credit on exports past the meter while preventing "duplicate compensation, through both the aggregated distributed energy resource and the retail bill credits", and to send the ISO recommendations on its proxy demand resource and DER aggregation models by 30 June 2028. The commission's own staff have already filed the objection the legislature is now telling the commission to resolve.
What is still unknown
Whether a final proposal exists. Comments on the draft final proposal were due on 16 July 2026. As of 20 September 2026 the initiative page lists no document later than July and gives the status as proposal development. The revised framework Utility Dive reported as expected on 19 August has not appeared there. [UNVERIFIED: absence from the initiative page is not proof no document was issued elsewhere.]
The CPUC demand response rulemaking that would settle the retail side. Utility Dive gives a February 2026 scoping order, decisions on urgent bridge-year funding in the fourth quarter of 2026, and up to February 2028 for four remaining issues. [NEEDS DATA: the docket number, and the scoping memo those dates come from.]
[NEEDS DATA: behind-the-meter capacity currently registered in CAISO demand response resources, which is the baseline the 2 GW estimate should be read against. The 2 GW is an advocate's figure, not an ISO one.]
Whether Track 2 reaches aggregation-level or sub-load-aggregation-point exports, and on what timeline. The proposal names it as a future topic without committing to one.
Whether the ISO's answer on double compensation survives contact with FERC, given that the proposal falls under the Western Energy Markets Governing Body's primary authority because the rules would apply to WEIM and EDAM entity balancing authority areas rather than to the ISO's alone.
Sources
CAISO, Demand and Distributed Energy Market Integration: Track 1 Draft Final Proposal, 8 July 2026: https://stakeholdercenter.caiso.com/InitiativeDocuments/Draft-Final-Proposal-Demand-and-Distributed-Energy-Market-Integration-DDEMI-Track1-Jul-08-2026.pdf
CAISO DDEMI initiative page, document list and status: https://stakeholdercenter.caiso.com/StakeholderInitiatives/Demand-Distributed-Energy-Market-Integration
Utility Dive, Brian Martucci, 5 August 2026: https://www.utilitydive.com/news/california-der-wholesale-market-participation-could-grow-2-gw-from-caiso/827122/
California SB 913 (Becker), bill text, adding Section 380.1 to the Public Utilities Code: https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB913
Public comments
Loading…