DER Weekly Digest — Week Ending November 13, 2026


This week: the demand side finally published depth, and a state regulator priced the fleet it had already paid for. On November 12 the New Jersey Board of Public Utilities unanimously approved “PowerUp New Jersey,” an interim virtual power plant program targeting a 3% peak-demand reduction at each of the state’s four electric distribution companies — 547.1 MW in total — and found that up to 82% of each utility’s target is reachable from devices already funded through state energy-efficiency and EV-charger programs if about three-quarters of eligible devices enroll. In the same week, three vendors put delivered demand-flexibility numbers on the record for the first time this quarter: Oracle reported a record ~1 GW of peak reduction across 73 events and 4.3 million customers; Uplight reported 15,629 MWh shifted over 1,807 events with a 735 MW peak dispatched the week of August 30; and Octopus Energy US reported 50–60% demand-response participation among smart-device owners against a 5–10% historical benchmark. Every one of those figures is self-reported and none carries third-party measurement and verification. And a targeted search closed the longest-open tracked item in this series: the energy assurance standard — the instrument that will define “dependable energy under stress” — is NERC Project 2024-02, titled “Long-Term Energy Reliability Assessments,” and in its October 2025 initial draft Demand Response appears exactly once, as item R2.2 in a list of inputs a Planning Coordinator’s methodology must address. DERs, VPPs and load flexibility do not appear at all.


⚡ Virtual Power Plants & Demand Flexibility

New Jersey stopped being the “undecided” entry in this series’ enrollment-channel comparison, and the number that matters is not the 547 MW. On November 12 the New Jersey Board of Public Utilities unanimously approved an interim virtual power plant program and the next block of the state’s storage program, branded together as “PowerUp New Jersey.” Customers will be paid to let their utility remotely curtail smart devices — thermostats, EV chargers — or discharge home batteries and solar during peak demand. The program targets a 3% peak-demand reduction at each of the state’s four electric distribution companies (PSE&G and Atlantic City Electric among them), a combined 547.1 MW. Interim programs must launch by October 1, 2027 and run through September 30, 2029 before more permanent versions take over, with enrollees paid either directly by the utility or through a distributed energy resource manager. The companion Phase 2, Block 1 storage incentive pays $200/kW for new capacity and $100/kW for existing capacity — targeting 150 MW new plus 100 MW existing, with payments trued up to delivered output — against a state goal of 2,000 MW of storage by 2030.

The finding to carry out of this order is the 82%. The BPU found that devices already funded through the state’s energy-efficiency and EV-charger programs could meet up to 82% of each utility’s target if roughly three-quarters of eligible devices enroll. That is the first time a commission in this series has quantified what the daily log and the Brattle work have been arguing from the other direction: the marginal cost of residential flexible capacity is enrollment, not hardware. Ratepayers in New Jersey have already bought most of the fleet. What has not been bought is the enrollment, the telemetry and the settlement path — which is the entire reason the “DER manager” role appears in the order at all. Read the 82% beside two independent confirmations already in this series: Brattle’s finding that OEM-led enrollment of installed devices is the lever that takes national residential VPP capability from ~10 GW to 35 GW, and California’s ChargeWise Phase 2, where the CEC is subsidizing user acquisition and onboarding rather than hardware. Three different institutions, in three different postures, have now located the binding cost in the same place. If your own DSM business case prices new devices, re-price it against the installed base and report the enrollment gap in MW.

The second thing this order does is resolve the four-way enrollment-channel comparison this series has carried since July. New Jersey was the undecided entry against Virginia’s dual enrollment, Minnesota’s utility-only and Nevada’s utility-only with a two-year deferral. It has now landed on dual channel: paid directly by the utility, or through a DER manager. That is a win for the platform specification this series has recommended indefinitely — IEEE 2030.5 plus dual-counterparty settlement as pass/fail — because New Jersey has institutionalized a third-party orchestration layer between customer devices and the utility, which is precisely the function a DERMS fills and precisely the arrangement a utility-only design cannot accommodate without a rebuild. Specify for two counterparties even where your current order names one; two of four states have now gone dual and the third changed its mind by commission order.

The caveats are real and they are the usual ones. The 82% is conditional on ~75% enrollment, which no residential program in this series has ever achieved — Octopus’s 50–60% figure below is the high-water mark and it covers smart-device owners only. No per-utility MW split was disclosed, and the order does not address PJM wholesale-market participation, so whether these resources will also earn capacity revenue — and under what accreditation — is open. Note the timing against PJM’s backstop: registration for a qualifying offset runs through June 1, 2032, and New Jersey’s interim programs expire September 30, 2029, so an interim program’s enrollments have to survive a program transition to be worth anything in that auction. (Source: New Jersey Monitor, November 12 — November 13 entry; the BPU’s VPP proposal has appeared five times in this series since July 27 and is reported here as the decision, not as a new proposal)

Three vendors published demand-flexibility depth in the same week, and that is the news whatever you think of the numbers. This series has carried, for a month, the observation that PJM publishes call frequency — nine demand-response days in 2026 against eight across all of 2015–2024 — but not delivered depth, and that the only depth claim on the public record was an aggregator’s unauditable “~105% of committed.” That changed this week, from the vendor side.

Oracle reported (November 11) that its Opower platform helped North American utilities achieve a record ~1 GW of peak-demand reduction in 2026 through behavioral demand response and peak-time rebates, across 73 winter and summer events — up from 49 in 2025 — reaching 4.3 million customers, up 30% from 3.3 million, at an average 16,592 kW per event. The named examples are the useful part: Puget Sound Energy’s Flex Events delivered up to 16.6 MW per event and introduced a $0.50/kWh bill credit in 2026, while CPS Energy in San Antonio enrolled more than 350,000 customers — over a third of its base — and cut nearly 19 MW across 12 August events.

Uplight reported (November 11) that its behavioral-DR, connected-device and VPP programs shifted 15,629 MWh across summer 2026 over 1,807 flexibility events, with resources called roughly five times more often in July and August than in prior periods, residential dispatchable capacity up more than 75%, and enrollments past 1 million as participation more than doubled. The capacity figure is a 735 MW peak dispatched the week of August 30, nearly double the 2025 peak week, with battery dispatch up 70% and EV participation up 2×. The low-cost line is the one to note: behavioral demand response, whose enrollments rose more than fivefold, delivered over 20% of the total load shift with no customer hardware at all, and an unnamed Midwestern utility delivered 800 MW across nine events through opt-out behavioral DR. Uplight restated goals of 20 GW under management and $1B in customer savings over five years.

Octopus Energy US reported (November 6, updated November 8) that 50–60% of its customers who own a smart device participate in demand response, against a frequently cited 5–10% historical utility benchmark — a result CEO Nick Chaset, who also leads Uplight after the two combined, attributes to “all-carrot,” opt-out program design that leads with customer benefit rather than mandatory-style event calls, with about a 10% peak-to-off-peak household shift as the savings mechanism. Uplight GM Hannah Bascom supplied the operative claim: utilities could call enrolled distributed assets dozens of times a year rather than a handful of seasonal emergency events.

Read the three together and the shape of the argument is clear, as is the hole in it. Taken at face value, the cluster says that opt-out behavioral enrollment and device orchestration together now deliver gigawatt-scale, repeatedly dispatched peak capacity, and that the cheapest megawatts in the stack come from households with no hardware — Oracle’s entire ~1 GW and over a fifth of Uplight’s load shift. That is the direct answer to the frequency-without-depth problem, and it lands in the same week PJM’s winter outlook says reliability holds at its high-stress case only with emergency demand response in the stack. But every figure in the cluster is the vendor’s own. No third-party measurement and verification accompanies any of it; the two utilities Uplight cites are unnamed; Uplight discloses no total GW under management; Octopus’s 50–60% covers smart-device owners only, with no disclosed measurement method, opt-in/opt-out split or time period; and Oracle’s per-event kW is a platform average across heterogeneous programs. Treat all of it as vendor claims pending filed program evaluations, and note the asymmetry that makes this week instructive: the demand side can now produce depth numbers at will and still cannot produce an audited one. The unresolved question Chaset himself names is the right one to carry: does participation hold as call frequency rises? If it does, distributed devices begin to compete with utility batteries in resource plans; if it falls, programs revert to seasonal emergency tools, and the 2026 figures above will have been a peak rather than a baseline. (Sources: Interesting Engineering, November 11 — November 12 entry; Uplight via GlobeNewswire, November 11 — November 13 entry; MarketScale, November 6 — November 11 entry. Three distinct vendors and datasets, reported here as one cluster because the analytic point is the cluster)

Sunrun’s Puerto Rico fleet ran 18 consecutive days of dispatch, which is a stress test of the one DERMS function everyone assumes. Sunrun’s Puerto Rico residential-battery VPP has roughly doubled from about 4,000 customers in 2024 to more than 60,000 — a fleet of nearly 70,000 home batteries providing 45 MW of dispatchable power over four-hour windows — and has been dispatched more than 60 times since April and for 18 consecutive days beginning October 13 during unusually warm autumn weather, to address shortfalls and help avoid rolling blackouts. Running parallel to it, Walmart’s October 30 strategic investment in Base Power extends a provider-owned home-battery-and-electricity retail model to Walmart and Sam’s Club customers beginning in Texas, with Base installing, owning and maintaining the battery and a 20% discount on the standard installation fee.

The operating record is the point, and it is a different kind of evidence from the vendor numbers above. Forty-five megawatts of callable, four-hour residential storage behaving as a repeatable grid-stress resource is deferred peaker value delivered entirely from behind-the-meter assets the utility neither owns nor financed. And eighteen consecutive days is the closest thing in this series to a published endurance test of fleet-scale state-of-charge management, dispatch and performance verification across tens of thousands of endpoints — the core DERMS function, exercised not for an hour but for most of three weeks. Ask any DERMS vendor for the multi-day case: what happens to available capacity on day twelve, how state of charge is rationed across a fleet when recharge windows are compressed, and what the utility sees when it is not met. On the Walmart side, take the channel rather than the investment: provider-owned hardware attacks the single largest barrier to residential DER enrollment — customer up-front cost and asset ownership — and is the scaling path the DOE’s 80–160 GW-by-2030 VPP target depends on, against today’s ~30–60 GW. Carry it beside Base Power’s Central Georgia EMC deal, where the co-op holds the dispatch right, which remains the cleanest structure in this series for crediting distributed capacity.

The standing caveat applies and has now been earned twice over: repeated dispatch shifts wear onto customer-owned equipment, which raises the battery-degradation-versus-compensation question, and per-battery kWh ratings and customer compensation amounts are undisclosed. Eighteen consecutive days of discharge on customer hardware with no published degradation accounting or compensation adjustment is a liability question before it is a capacity question. (Source: MGRID, November 9 — November 9 entry. First appearance in the tracked record for the Puerto Rico fleet; the September and October California dispatch figures carried in this series — 580 MW, 601 MW, 745 MW — are a different fleet and are not comparable)


🔌 DERMS & Grid Integration Technology

The seventh consecutive week produced no DERMS deployment item — but it produced the first standalone ADMS item in that run, and one sentence in it is an RFP line worth copying verbatim. Hitachi Energy was named a Leader in the 2026 IDC MarketScape for Worldwide Advanced Distribution Management Systems (Doc #US54133826, October 2026), an assessment that scored eight ADMS providers on capability and strategy criteria. IDC analyst Jean-François Segalotto credited a deep bench of grid-automation and control-system engineering talent concentrated in Europe and inherited from ABB Power Grids, a SCADA foundation shared across transmission and distribution, and — the sentence that matters — a distribution state-estimation capability that “handles high-DER conditions without requiring telemetry from every resource.” Network Manager is described as built over 30-plus years on more than 130 years of utility heritage.

That capability statement is the inverse of the problem this series has been specifying around for three months, and both framings are correct. This series’ standing instruction is that metering and telemetry fidelity is the primary DERMS specification line, and that per-asset settlement behind a single service point is pass/fail — because, under California’s SB 913, device-level telemetry became a statutory precondition for resource-adequacy qualification, and because the settlement quantity is per-asset while a net meter measures the net. Nothing in a state estimator changes that, and the distinction is the one to hold: as feeders fill with distributed solar, storage and VPP-enrolled devices, the binding constraint shifts from control to observability, and a state estimator that can infer feeder conditions without full device telemetry directly lowers the metering and communications burden an ADMS operating specification must carry. It does not lower the burden a settlement specification carries. So write both lines into the RFP and keep them apart: inferred feeder-level state for operations and hosting-capacity assessment, measured per-asset telemetry for accreditation and settlement. A vendor that offers the first as a substitute for the second is offering you an operating tool and a disallowance.

This is the same grid-edge-visibility bottleneck the Carnegie Mellon synthesis below identifies as the real constraint, arriving from the vendor side. The caveats are substantial and they bound how far this can be used: the release names no utility deployments, no market-size or share figures, and no DERMS product — it describes DER-aware operation rather than a distinct DERMS module — and an IDC “Leader” placement is a point-in-time analyst judgment, not evidence of a procurement. Use it as context for how the ADMS field is being graded and for the one capability claim worth testing, not as a deployment signal. (Source: Hitachi Energy via PR Newswire, November 11 — November 13 entry)

Arizona approved the hardware layer underneath the orchestration policy this series tracks, and it added a sixth public price point for residential flexibility. Arizona Public Service (APS), Salt River Project (SRP) and Tucson Electric Power (TEP) approved new residential-storage and interconnection technologies that lower the time and cost of bringing home batteries into virtual power plants. On the program side: SRP’s Battery Partner Program rewards customers for letting SRP discharge FranklinWH or Tesla batteries during high-demand periods, and APS’s Storage Rewards Pilot pays up to $110/kW based on a participant’s average contribution across demand-response events — both coordinated through EnergyHub’s DERMS, with leasing through Palmetto’s LightReach lowering the up-front barrier. On the hardware side: ConnectDER’s IslandDER meter-socket adapter, now approved by all three utilities, is installable in 15–30 minutes without a main-panel replacement or service upgrade and is compatible with FranklinWH, Lunar Energy, SolarEdge and EcoFlow; ConnectDER’s solar and EV adapters were already approved across the three. The battery is FranklinWH’s 15 kWh aPower2 LFP system.

Two things to take, one of them a direct continuation of last week’s lead. First, the price point. APS’s up to $110/kW on average contribution across events joins a stack this series can now quote in full: the ~$66/kW-year avoided-cost anchor; Alliant’s $100 per average kW discharged; DTE’s $105/kW-year battery pilot; APS at up to $110/kW; Entergy Louisiana’s $125 per average kW/year; and now New Jersey’s $200/kW for new storage capacity and $100/kW for existing. Every utility-set price in that stack sits above the generic avoided-cost benchmark, and the spread runs from roughly 1.5× to 3×. That is either deliberate acquisition buy-down or evidence that local avoided costs are materially above the generic figure — and in any proceeding where you are defending the lower number, you now have to say which, because six jurisdictions have priced it higher. Note also that APS pays on average contribution across events, the same availability-weighted basis Entergy uses and the opposite of how a capacity market accredits depth — which makes the basis, not just the level, the thing to reconcile.

Second, the binding constraints at the grid edge are enrollment friction and physical interconnection, not platform capability. A DERMS books flexible capacity in an avoided-cost filing only once devices are actually and administratively connected, so a utility’s VPP ceiling is set as much by adapter approvals and financing as by software. The IslandDER approval is the concrete version: in older homes whose meter-main sockets lack space for storage equipment, a 15–30-minute install with no panel swap removes a cost and a scheduling bottleneck that no amount of orchestration logic can route around. Put utility-side equipment approvals on the DERMS deployment critical path alongside interconnection and local permitting — this series has already moved local permitting there, on New York’s evidence, and this is the same lesson one layer down.

Caveats: no MW or customer-count totals were disclosed, so the programs’ current contribution to peak cannot be sized, and the supporting figures — a ~47% Arizona battery attachment rate (ConnectDER estimate) and 8% quarter-over-quarter US residential-storage growth in Q4 2025 (Wood Mackenzie/ACP) — are vendor and analyst estimates rather than utility-filed data. Read this beside APS’s own IRP filing carried last week, which projects behind-the-meter storage rising from 69 MW to 1.2 GW by 2045 while rooftop solar runs about 30% below forecast: these approvals are the mechanism by which the storage half of that forecast either happens or does not. (Source: Energy-Storage.news, early November — November 12 entry. Distinct from last week’s FranklinWH item, which covered Entergy Louisiana and Alliant Energy — different utilities, different programs, same vendor)

A Carnegie Mellon panel named the adoption barrier in a form a utility can act on: one device, nine or ten internal departments. A synthesis of Carnegie Mellon University’s Energy Week 2026 argues DERs can help utilities absorb data-center, reshoring and electrification load faster than new transmission can be built, and quantifies both sides of the gap. On the demand side: a single planned upstate-New York semiconductor facility will need as much electricity as New Hampshire and Vermont combined, against a NERC 2025 finding that peak-demand growth in several US regions is now outpacing new generation for the first time in years. On the transmission side: the 3,500 MW SunZia line in New Mexico took 17 years to approve. The demand-side evidence is directly on this series’ thread — utility-scale battery storage and demand response are credited with keeping Texas wholesale prices from spiking during a recent winter storm, against Winter Storm Uri in 2021, when ERCOT prices ran roughly 400× above average for four consecutive days.

But the diagnosis is the contribution, and it has two numbers in it. First: connecting a single behind-the-meter device can involve nine or ten internal utility departments. That is a process cost, it is invisible in every program evaluation this series has read, and it is the mechanism behind the enrollment gap New Jersey just quantified from the other end. Count the departments in your own interconnection-to-enrollment path and treat the count as a program cost line, because it is the one line a larger device incentive cannot buy down. Second: roughly 800 distribution cooperatives and 2,000 municipal utilities mostly lack virtual-power-plant programs at all. This series has repeatedly found that co-ops and munis move faster than IOUs when they do act — Base Power’s Central Georgia EMC deal with the co-op holding the dispatch right, Budderfly’s Cadillac, Michigan contract, FMPA specifying connectivity-model requirements for 32 member utilities — and the 2,800-utility number is the size of that under-served channel stated plainly. The throughline for practitioners is that assets behind the customer’s meter give utilities little visibility into actual performance, which is precisely the gap a DERMS is procured to close, and that for smaller utilities the remaining work is analytics, valuation and dispatch rather than hardware. Working models offered: Danville Utilities in Virginia, using distribution-level batteries to cut transmission peak charges, and six Massachusetts municipal utilities.

The caveat is the format: this is panel-synthesis commentary, not a filed program or dataset, so its value is as a framing of adoption barriers and as two citable counts, not as a quantitative input. (Source: Environment+Energy Leader, updated November 11 — November 11 entry)


🔋 Storage, Distributed Solar & DER Procurement

PG&E deferred $2B and stalled 15 GW of renewable and storage interconnection — and the driver is a balance sheet, not a grid constraint. PG&E has postponed about $2 billion of work planned for 2027, cutting its 2027 capital plan from $13.4B to $11.4B — roughly 15% — in a reduction announced September 2. The deferral puts more than 60 renewable and storage projects representing at least 15 GW that were targeting 2030 operation into open-ended delay for grid access, on top of large-load connections delayed beyond an initial 1.8 GW tranche. The driver is a push to strengthen the balance sheet and reach investment grade while PG&E lobbies for wildfire-liability changes: SB 492 advanced without those protections, shares fell 20%, and Fitch moved the outlook from stable to negative while affirming BBB-.

This is the ownership-and-financing-risk dimension of the avoided-cost argument, made concrete, and it cuts in the demand side’s favor for an uncomfortable reason. Interconnection access and supply-side buildout hinge on a utility’s capital allocation and credit health, not only on queue throughput or hosting capacity — which is exactly why demand-side resources, VPPs and non-wires alternatives that defer rather than require utility capex are a hedge when a balance sheet forces deferrals. Pair it with this series’ standing instruction to run the ownership screen per jurisdiction: a restructured state bars the wires company from owning generation, and a credit-constrained utility cannot fund it even where it may. Both roads lead to third-party and customer-owned distributed capacity, and the second road is the one nobody plans for. The article notes developers’ partial workarounds — competitive transmission developers taking over delayed projects, and solar and storage owners building connecting infrastructure through build-transfer agreements — which are the supply side’s version of the same adaptation.

Two limits on how far to push it. The article does not address distribution hosting capacity directly, so the DER-specific read is inferential rather than reported. And a 15 GW interconnection delay is not a 15 GW capacity shortfall — these were 2030 targets in a queue whose historical completion rate, on Goldman’s finding carried in this series, runs 50–60%. The citable fact is the mechanism, not the megawatts: a utility’s credit rating is now an input to DER deployment timing in the largest DER market in the country. (Source: MGRID, November 8 — November 9 entry)

New York put $13M into long-duration chemistries, and one award is a clean statement of the substation-deferral use case. New York’s Power Generation and Storage Innovation Program, funded by the state’s Clean Energy Fund, awarded nearly $13 million across 15 projects developing long-duration and advanced battery storage rated for more than 10 hours at full power. The largest single award — over $2.4 million to Inlyte Energy for an iron-sodium battery — is explicitly targeted at substation peak-load growth and distribution resilience, which is precisely the non-wires, substation-deferral use case a DERMS is procured to orchestrate and the locational value an avoided-T&D business case captures. The rest of the portfolio spans chemistries aimed at congestion, limited hosting capacity and lithium-ion siting constraints: Standard Potential sodium-ion portable station ($1.8M), Urban Electric Power zinc-manganese dioxide (~$1.5M), Sprocket Power pairing DC fast chargers with an underground vanadium redox flow battery ($1M+), and Aestus compressed-air (a 20 MWh target funded here only as a ~35 kWh subscale heat-storage test).

Read it as a forward indicator and nothing more, but read the ten-hour threshold. Most of these are prototype, subscale or demonstration stage, with power ratings, energy capacities, sites and operating dates undisclosed, so near-term dispatchable capacity is minimal — and the source’s own critical note argues that spreading ~$13M across 15 projects risks laboratory-stage rather than field-deployable results. What is worth carrying is the specification New York chose to fund against: more than ten hours at rated power. That is a duration band no current four-hour residential or utility battery serves, and it maps onto the 5.3–7-hour curtailment durations the Currence projection carried in this series attaches to “last-in-first-curtailed” interconnection regimes. The rule that sets curtailment duration is written before the asset is sized, and New York is funding the chemistry that a long-duration rule would require. Note also the disclosure discipline problem this award set repeats: an energy rating without a kW rating cannot be converted into a capacity contribution, which remains a standing requirement in this series for every DER disclosure relied on. (Source: MGRID, November 8 — November 9 entry)


🏗️ Data Centers & Large Load Growth

ERCOT’s CEO wants to curtail data centers before scarcity begins — and the authority he already has has never been used. ERCOT President and CEO Pablo Vegas said on November 11 that he plans to ask the Texas Legislature for authority to require large data-center loads to reduce consumption earlier — at a Level 1 condition or even before the first sign of scarcity — because the operational realities of curtailing very large customers “really requires us to think about pulling that up a little bit earlier,” rather than waiting until the grid reaches the second step of a three-stage emergency, as current rules require. The request would extend the toolkit created by Senate Bill 6 (2025), which already allows ERCOT to order large loads to cut consumption on 24 hours’ notice — an authority ERCOT has not yet exercised. For the coming winter, Vegas framed data centers as a low-single-digit-gigawatt pool of flexible supply that voluntarily scaled back during last winter’s severe weather. The stated goal is to shield residential and small-business customers during scarcity.

This answers a tracked question in this series and the answer is the interesting part: no, SB 6 curtailment has still never been exercised — and the operator’s response to that is to ask for more of it, earlier. Read it as a statement about where mandatory curtailment sits in the hierarchy of useful instruments. Moving curtailment “up” the emergency ladder is exactly the regulatory lever that converts large-load flexibility from an emergency-of-last-resort measure into a planned, dispatchable resource — the firmer and earlier the curtailment obligation, the more avoided-capacity value a verified demand-side commitment can carry, and the more directly it competes with building new peaking capacity. But this series’ standing instruction holds and is reinforced rather than weakened here: count mandated large-load curtailment as emergency risk reduction, never as DR capacity with an avoided-cost value. Texas Docket 59220 established that SB 6 curtailment is uncompensated and uncapped; a Level 1 trigger makes it earlier and uncompensated. The customer response to an earlier uncompensated obligation is to buy its way out of it — with on-site storage, on-site generation, or a flexibility-priced contract — which is where the addressable market is, not in the mandate itself.

Note the cross-market pattern, because it is now three jurisdictions in two weeks. ERCOT is seeking earlier curtailment authority; Pennsylvania’s PUC has a live proceeding on the emergency load-control order of operations for large loads; and PJM’s winter outlook below places emergency demand response and DOE-authorized large-load backup generation at the center of its reliability plan. The common movement is away from treating large flexible loads as exempt “critical load” and toward treating them as curtailable. That single reclassification is what gives firm, M&V-verified demand-side flexibility its avoided-capacity value, because a large customer able to self-curtail or bring on-site DER moves out of the first-to-be-shed tier. Screen every large-load tariff and emergency operating procedure in your territory on whether it places enrolled DR ahead of or behind unsupported large load, and get the dispatch-order language into both documents with control-room verification.

The caveats: this is a stated legislative intention, not a bill, rule or filed tariff; no MW threshold or trigger definition was specified; and the inference that data centers will respond with more on-site generation or volatility-priced contracts is the reporter’s, not an ERCOT plan. (Source: Bloomberg, November 11 — November 12 entry)

Oncor has 293 GW of data-center requests against a 33 GW peak, and only about 45 GW of it has cleared even a conditional classification. Oncor reports about 293 GW of data-center requests in the ERCOT transmission queue as of September 30 — nearly nine times its new 33 GW system peak, itself up 6% — and expects its 2027–2031 base capital plan to rise roughly 20% from the current $47.5B, with 2026 spend alone running to about $9B ($6.3B through September), its largest annual program. ERCOT’s “Batch Zero” process, approved June 2026, groups large users seeking at least 75 MW for joint study rather than one-by-one, and has conditionally classified ~24 GW of Oncor’s submitted load as base (10 GW already energized) and ~21 GW as studied, with applicant review due in December and a final Batch Zero transmission plan in fall 2027. Oncor holds about $6B in customer collateral, including ~$2B tied to Batch Zero that the current base plan excludes.

The business-case signal is the ratio, and it is the sharpest version of the forecast-risk argument this quarter. Only about 45 GW of the 293 GW queue has cleared even conditional classification — roughly 15%. The gap between speculative requests and firm load is enormous, and it is precisely the “risk of unrealized loads” that makes utility transmission capex a stranded-cost hazard and strengthens the case for flexible, deferrable demand-side resources that do not commit ratepayers to decade-long wires builds. This series has found that every financial-commitment filter collapses a large-load pipeline by roughly an order of magnitude — AEP Ohio’s 30 GW to 13 GW is the cleanest prior instance — and Oncor’s 293-to-45 is an order of magnitude from the other direction, applied by the operator rather than the utility. Use the ratio, not the headline: a capital plan sized against a queue is sized against an artifact, and the deferrable alternative is the lower-regret purchase at every point on that curve.

The second citable element is the collateral. $6B held, with ~$2B tied to Batch Zero and excluded from the base plan, is the financial-assurance mechanism that cost-causation advocates across this series’ large-load items have pushed for, shown working at scale. It joins SRP’s 60% pre-payment requirement, Xcel’s Clean Energy Accelerator Charge and OpenAI’s self-funded upgrades as the fourth utility-designed cost-causation instrument in six weeks. Ask the question this series has asked of each of the others: can the instrument fund flexibility and storage, or only wires and generation? A collateral or charge mechanism built on demand-side cost causation that can only be spent on supply-side assets writes the demand side out of a mechanism designed around it.

The limit on this item: it does not address demand response or distribution hosting capacity directly, so the DER read is inferential. Its value is as a hard, utility-level data point on the load-growth and stranded-cost pressures that frame the avoided-cost argument. (Source: MGRID, November 9 — November 10 entry)

PJM’s winter outlook says reliability holds at its high-stress case only with emergency demand response in the stack. PJM’s 2026–27 winter assessment studies a high-stress cold peak of up to 152,200 MW against the record winter peak of 143,700 MW set January 22, 2025, expects adequate supply under normal conditions, but concedes margins are tightening as demand outpaces new generation — and its studies show reliability can be maintained even at the 152,200 MW stress case only with pre-emergency and emergency demand response in the stack. The Department of Energy’s standing authority for PJM to order data centers and other large loads to run on-site backup generation sits one rung below that as a last-resort demand-cutting tool. The assessment flags localized delivery risk in the transmission-constrained BGE, Pepco and Dominion zones (Maryland, Virginia, D.C.), and the National Weather Service forecasts a slightly warmer winter across the footprint.

This is a live, operator-authored validation that emergency demand response offsets generation a utility would otherwise have to build or procure — the same logic behind the ~$66/kW-year capacity-value anchor and the 70 MW → ~$130M net-benefit arithmetic this series carries. It is also the most useful sentence available against the problem named in this week’s EPRI Spotlight: the energy assurance standard treats Demand Response as a methodology input, while the operator running the largest market in North America says its winter case does not close without it. Those two documents describe the same resource and credit it completely differently. Put PJM’s sentence in the record of any proceeding where demand-side credit is being set, and attribute it to the operator rather than to an advocate.

Two operational reads. First, the BGE, Pepco and Dominion call-out is a textbook case for locational demand-side and non-wires resources, because imports from elsewhere cannot relieve a local delivery constraint — which is the same locational logic behind Toronto Hydro’s station-by-station procurement carried in Benchmark Hygiene below and behind FERC’s objection to backstop capacity delivered anywhere rather than in the constrained zone where the load sits. Second, for DERMS and ADMS procurement, the implication is the telemetry, dispatch-signal and M&V capability needed to actually call and verify emergency DR and large-load backup at scale — and this series’ standing instruction applies directly: build the measured winter record now (morning-ramp delivery, multi-hour sustained output, state of charge at call time, cold-weather derate, per event and per site), because winter is the specific attack on demand-side credit and a modeled record will not answer it.

Caveats: no reserve-margin percentage is disclosed, and the slightly-warmer-winter forecast tempers the probability these tools are called — which is itself the problem, because a resource whose credibility depends on being called is poorly served by a mild winter. (Source: MGRID, November 7 — November 9 entry. Distinct from MISO’s 2026–27 Winter PRA carried in last week’s digest: different operator, different proceeding, different source)

One supply-side marker, carried for the gap it illustrates rather than for the project. EDP Renewables North America broke ground November 4 on the 250 MW Cypress Knee solar project in Chicot County, southeastern Arkansas, under a long-term PPA with Meta — the pair’s third deal, 545 MW combined — with completion expected in 2028 and projected county tax revenue of about $25M/year. The telling detail is the source’s own caveat: solar does not run around the clock, and Meta is separately backing new gas plants to meet growing demand. A 24/7 AI data-center load matched to an annual-energy renewable PPA still leaves an hourly firm-capacity shortfall that a utility must plan for — the gap emergency DR, curtailable large-load agreements and dispatchable storage are positioned to fill, and the reason a 100%-annual-match pledge is not a resource-adequacy statement. No storage, interconnection detail, serving utility (Entergy operates the region) or total investment disclosed; direct DER relevance is low. (Source: MGRID, November 7 — November 10 entry)


📋 Regulatory & Policy

A Senate bill would convert FERC Order 2222 from a tariff-by-tariff compliance process into a statutory mandate with an 18-month deadline — and names capacity accreditation as the real battleground. Two readings of the Bipartisan American Affordability and Jobs Act (BAAJA, S. 5653, introduced September 30 by Sen. Capito with Lee, Whitehouse and Heinrich, now before the Senate Energy and Natural Resources Committee) entered the log this week, and together they cover both halves of the bill.

On the DER and VPP provisions: Section 2108 would create a new Federal Power Act Section 229 requiring RTOs and ISOs to admit “qualifying” aggregations to wholesale markets within 18 months of enactment, with compliance filings due in 180 days and any blocking rule deemed unjust and unreasonable — six years after Order 2222 issued and with several grid operators still not fully compliant. A new Section 230 would bar distribution utilities from using control of the distribution system or retail sales to block participation or exports, and require them to share the same interval metering and telemetry the wholesale market uses (safety and reliability actions preserved). The bill avoids the terms “DER” and “VPP,” instead defining a “qualifying behind-the-meter distributed energy system facility”: behind a retail meter, dispatchable within 15 minutes, at least 5 kW, with an aggregation requiring two or more such facilities totaling at least 100 kW. It preserves state primacy — utility obligations yield to state law, FERC gains no authority over distribution-level interconnection rates, and state regulators decide retail-bill treatment — permits dual state and wholesale participation so long as the same service is not paid twice in an interval (VPP providers self-certify, with FERC penalties up to three times the amount), and funds a voluntary DOE local-permitting program at $20M/year for FY2026–2031 covering rooftop solar, batteries of at least 2 kWh, and EV chargers.

On the permitting provisions: more than 300 solar and storage companies signed a letter urging passage. The bill would amend NEPA, the Clean Water Act and the Endangered Species Act to shorten permitting timelines and give FERC authority to permit some transmission projects. SEIA projects the reforms could unlock 167 GW of additional solar between 2028 and 2035; signatories cite $120B of permitting-stalled clean-energy investment and note solar-plus-storage supplied about 70% of new US grid capacity added so far in 2026.

The analytic payoff is the authors’ own judgment, and it confirms the central theme of this series: capacity accreditation — not interconnection — is flagged as the real battleground, since the economic value of an aggregation in any value stack turns on how its capacity is credited. That is the same finding this series reached from three other directions in the last month: PJM’s narrowed offset definition, MISO’s stiffer LMR non-performance penalty, and the ~8% default winter accreditation now live in Missouri’s Ameren docket. A statutory right of entry does not settle credit. The authors’ second judgment is the one to plan around: they expect the data-sharing mandate to be the provision “honored in form but slow in practice.” If that is right, then the operative asset is not the statutory right but the metering and telemetry record you already hold — which is exactly what PJM’s qualifying-offset criteria, recovered last week, make dispositive. Two instruments, one federal and one market, both reduce to the same question: can you document the asset?

Two design details worth lifting into your own advocacy regardless of the bill’s fate. The thresholds are an eligibility screen in disguise — 5 kW per facility, 100 kW per aggregation, dispatchable within 15 minutes — and a 15-minute dispatch requirement writes out slow thermal loads while writing in batteries, EV chargers and smart thermostats. And the double-payment rule with self-certification and treble penalties is the first federal attempt in this series to solve the dual-enrollment problem that four states have each solved differently; New Jersey’s dual-channel order this week makes it a live question rather than a theoretical one.

Caveats, and they are substantial on both halves: this is draft legislation with uncertain prospects — a parallel Senate permitting effort has stalled before, and the House’s own 100 MW bill stalled the day after passing 417-3; open questions remain on whether standalone rooftop solar, smart-thermostat flexible load or uncapped on-site gas generators would qualify; the thresholds are stated in kW, kWh and dollars with no GW or MW system figures; the permitting scope covers natural gas, oil and coal alongside clean technologies; and FERC transmission-permitting authority could trigger disputes with state commissions. The 167 GW is a contingent projection by a trade association, not a forecast. (Sources: Foley Hoag, November 6 — November 11 entry, and MGRID, November 10 — November 10 entry. Two sources, two halves of one bill, consolidated here as one item — deduplicated)

California is about to adopt guiding principles for the calculator every DSM program in the state is tested against, on the same agenda that tightens the terms on a 250 MW Google load. Two DER-relevant items sit on the California PUC’s November 19 voting meeting. In R.22-11-013 the Commission would adopt guiding principles — accuracy, consistency and transparency — for the Avoided Cost Calculator (ACC) used in DER cost-effectiveness analysis, together with a companion Distributional Equity Analysis. In Resolution E-5455 it would approve, with modifications, PG&E’s agreement to energize a new 250 MW Google load in the San Jose area, adding a $600,000/MW capital advance and capping annual refunds at 100% of the net revenues PG&E receives from the customer, explicitly to add ratepayer protections.

The ACC item is the one with money attached to it, and this series has a standing instruction that this agenda directly engages. Because the ACC monetizes the avoided generation, capacity, T&D and ancillary-service value that every DSM, DR and storage program is measured against, changes to its governing principles ripple into which programs clear and how much capacity value a demand-side resource can claim in a California filing. This series already requires auditing every avoided-cost figure for ACC vintage — now Resolution E-5476, the final 2026 ACC — and for which test the commission actually applies (TRC/PAC/RIM/PCT), a question California sharpened in October when it authorized ~$608M of two-year DR bridge funding and exempted it from the Societal Cost Test in the same decision. A principles order is the upstream document for all of that: “accuracy, consistency and transparency” are the criteria under which a later methodological change gets justified or resisted, so this is the proceeding in which to establish, now, that a principle of accuracy requires locational and hourly granularity rather than a system-average avoided cost. Get in before the methodology fight, because the principles will be cited back at you during it.

The pairing on one agenda is the week’s cleanest illustration of the series’ second theme. The same regulator tightening the valuation rules for demand-side resources is simultaneously tightening the cost-assurance terms on a hyperscale load — both expressions of the “growth pays for growth” and “verified value” principles. The $600,000/MW capital advance is also a benchmark: it joins SRP’s 60% pre-payment, Oncor’s $6B of collateral and FirstEnergy’s ~$250M/GW transmission multiplier as a quotable, utility- or commission-set number for what a gigawatt of new load is made to post. At $600,000/MW, a 250 MW load advances $150M — which is the same order as the avoided-transmission figure a 500 MW demand-side deferral would claim at FirstEnergy’s multiplier. Put the two numbers side by side in any cost-allocation docket.

Caveats: these are agenda items on draft or forthcoming decisions rather than final orders — either could be held or revised; the source is a law-firm regulatory digest rather than the primary CPUC decision; and the ACC principles are stated at a high level without the specific methodological changes that would move a given program’s benefit-cost ratio. Get the primary documents in R.22-11-013 and Resolution E-5455 after the November 19 meeting. (Source: Stoel Rives Energy Regulatory Update, November 11 — November 13 entry)

Pennsylvania narrowed its data-center curtailment proceeding because FERC rejected the PJM product it was built on. On November 5 the Pennsylvania PUC issued a clarification order updating the comment schedule on its October 1 tentative order in Docket M-2026-3064961 — the emergency load-control order-of-operations proceeding carried in this series since late September — after FERC’s decision rejecting parts of PJM’s proposed Interim Resource Adequacy Service (IRAS). Because the tentative order’s pre-emergency curtailment provisions were tied to IRAS, the PUC is no longer seeking comment on them, narrowing the proceeding to the electric-distribution-company emergency load-control changes that govern how utilities manage large loads when demand threatens to exceed supply while protecting critical facilities and essential public services. Comments are due 20 days after the clarification order is published in the Pennsylvania Bulletin, with reply comments 30 days after publication, expected “in the coming weeks.” The commission has also set a November 17 technical conference specifically on how the costs of large computational loads should be allocated, inviting utilities, consumer advocates, large-load customers and other stakeholders.

The significance is twofold and the second half is where the money is. First, the surviving question — whether data centers can be classified as curtailable rather than exempt “critical load” — is exactly the regulatory lever that gives firm, M&V-verified demand-side flexibility its avoided-capacity value. Second, the November 17 cost-allocation technical conference is where “growth pays for growth” gets tested in practice, determining whether large-load-driven capacity and upgrade costs fall on the data centers or on other ratepayers. That conference is four days after this digest’s window closes and is the single most actionable date on the board.

The structural lesson is about venue dependency, and it is worth generalizing. Pennsylvania’s pre-emergency curtailment provisions did not fail on the merits — they were dropped because the federal product they referenced was rejected on jurisdictional grounds. This series has already noted that FERC’s rejection of IRAS was jurisdictional, not conceptual, and that Commissioner Rosner’s concurrence put states in charge of the curtailment order. The consequence now visible: a state rule that incorporates a federal market product by reference inherits that product’s jurisdictional risk. Draft state curtailment hierarchies to stand alone on state authority rather than referencing an RTO service, because the RTO service is the part that gets remanded.

Caveats: this is a procedural clarification opening a comment window, not a final order, and the IRAS-linked provisions were dropped on a federal-jurisdiction development rather than resolved on the merits — so the durable curtailment hierarchy still depends on the final order, targeted for January 28, 2027, and on the revised federal design PJM must refile. And the Pennsylvania Bulletin publication date that starts the comment clock remains unverified after five runs — see the Method note. (Source: PA Environment Daily, November 5 — November 11 entry. The October 1 tentative order was reported in the late-September/early-October digest; this is the distinct November 5 clarification order, reported as movement)

FERC held, for the second time, that the cost of keeping a plant online under an emergency order falls on the load inside the emergency footprint. In an October 19, 2026 order FERC again rejected — without prejudice — TransAlta’s proposal to recover the fixed costs of keeping the 729.9 MW Centralia Unit 2 coal plant available under two DOE emergency orders, affirming that TransAlta may recover its compliance costs but faulting how it proposed to allocate them. DOE’s Section 202(c) orders of December 16, 2025 and March 16, 2026 directed TransAlta to keep Centralia — scheduled to end coal operation December 31, 2025 ahead of a gas conversion — available at the direction of Gridforce Energy Management (the balancing authority) and the Southwest Power Pool (the reliability coordinator). The unit was never dispatched, so TransAlta sought only a fixed charge of about $6.4 million ($6,395,348), split equally between Gridforce and SPP. FERC found the equal split violates cost causation because the two entities’ service areas do not match the emergency’s WECC-Northwest geographic scope, and directed the costs be spread across load within that area — the same reasoning by which a separate October 1 order rejected roughly $19.9 million for the first order’s period.

For this series’ cost-allocation thread the precedent is portable and it runs in the demand side’s favor. FERC is establishing that the costs of holding units online under emergency orders — a cost driver increasingly tied to load growth and thinning reserve margins — must fall on the load inside the emergency footprint. That is the identical principle at issue in data-center curtailment and large-load tariff fights across this series, arriving through a generation docket. Cite it where a utility proposes to socialize reliability-must-run or emergency-order costs across a whole system: FERC has now twice required the footprint to match, and has done so on a cost-causation rationale that applies to any emergency cost, not only this one. It also puts a price on the alternative this series keeps making: $6.4M plus $19.9M to hold one never-dispatched 729.9 MW unit available across two winters is about $26M for zero delivered energy — the expensive, last-resort version of what a planned, compensated demand-side program delivers with a settlement record attached. The 2026 Section 202(c) docket now exceeds fifty orders and has accredited nothing.

Caveats: this concerns a generation unit rather than a DER or demand-side resource, and FERC’s rejection is procedural, deciding allocation only and expressly not whether the costs themselves are just and reasonable. (Source: Troutman Energy Report, November 11 — November 12 entry)


🏢 Benchmark Hygiene & Carried Context

NERC’s registry petition is filed, and it is the Rules of Procedure track this series told you to watch rather than the standards track. On October 21, 2026 NERC submitted a petition to FERC in Docket RR27-1-000 seeking approval of Rules of Procedure revisions adding two functional entity types to the NERC Compliance Registry: Computational Load Owner (CLO), the entity that owns and maintains a Computational Load Site, and Computational Load Operator (CLOP), the entity that operates one. Registration would apply at a total connected load of at least 50 MW supplied at 100 kV or greater from one or more points of connection — the two-part test FERC itself declined to name — and NERC estimates this captures about 68,288 MW, roughly 86% of the computational load projected to be in service by end-2027. NERC requested expedited action so it can engage potential registrants sooner, and will file the proposed Reliability Standards and implementation plans later. The petition responds to FERC’s July 16, 2026 order on Bulk-Power System reliability risks from computational loads.

This closes a tracking item carried since September and confirms the sequencing this series predicted: the registry binds first, the standards follow. The standards NERC’s board adopted on October 26 bound nobody until this track landed. The number to carry forward is 68,288 MW at ~86% coverage — that is the first official sizing of the regulated computational-load population, and it is the denominator for every argument about how much large load is observable. Note what it does and does not do: registration creates observability obligations for large loads and credits none of their flexibility. Neither the petition as reported nor the registration categories mention demand response, load flexibility or DERs. (Targeted gap-fill — Troutman Energy Report, November 2026, on NERC’s October 21 petition; resolves the tracked “NERC early-November FERC filing” item. The anticipated filing date on NERC’s own Legal Update was November 4; the petition went in October 21)

A locational demand-response benchmark, dated rather than reported. Toronto Hydro’s Local Demand Response (LDR) program secured 12 MW of local flexible capacity through a competitive auction held in October 2025 — its largest single-year procurement since the program launched in 2018, against a historical 4–8 MW/year — expanding from two transformer stations to six (Finch, Leslie, Manby, Horner, Strachan, Dufferin) for dispatch during the 2026 summer peak, with a target of up to 30 MW across the six stations by 2029. The utility positions LDR explicitly as a non-wires alternative to defer or reduce distribution upgrades, against more than $600M of 2025 distribution capital. This entered the research log on November 9 under the “most recent relevant piece” rule from a page refreshed that day, but the underlying procurement is from October 2025 and the original publication is February 2026 — so it is carried here as a dated benchmark and is not reported as a new development. Its value is as the clean, utility-run instance of the station-by-station locational procurement that PJM’s BGE/Pepco/Dominion constraint call-out above argues for, and of the avoided-T&D logic behind this series’ 70 MW → ~$130M arithmetic. No clearing prices, participant payments, program cost or on-site DER types were disclosed, so no benefit-cost ratio can be derived, and 12 MW is a locational-deferral instrument rather than a system-capacity play. (Source: Environment+Energy Leader, published February 22, 2026, page updated November 9 — November 10 entry)

Three ERCOT large-load queue figures must stay reconciled; citing the wrong one misstates the pipeline by a factor of two. Oncor’s ~293 GW is data-center requests in the ERCOT transmission queue attributable to Oncor as of September 30, 2026. ERCOT’s ~474 GW is the total interconnection queue ERCOT was auditing, roughly 90% data centers, carried in this series since September. ERCOT’s 191.8 GW is large-load studies, the figure CAISO’s board materials used for comparison in October. They are three different denominators — one utility’s share of a transmission queue, a system-wide queue under audit, and a study count — and they are not interchangeable. Oncor’s ~45 GW conditionally classified (24 GW base, of which 10 GW energized, plus ~21 GW studied) is the only one of the four that reflects any screening at all.

Four PJM demand figures to keep straight. 162.6 GW is PJM’s 2026 summer peak (EIA-930, July 2), the anchor for summer derivations — trade reporting circulates 168.2 GW and a derivation on the wrong one is off by 5.6 GW. 143,700 MW is the record winter peak, January 22, 2025. 152,200 MW is the 2026–27 winter high-stress study case, not a forecast. 112,447 / 113,312 MW were the forecasts in the October 13–16 shoulder-season event. Winter and summer anchors are not comparable and the stress case is not a peak.

The residential-flexibility price stack, now six public utility-set points against one benchmark. ~$66/kW-year avoided-cost anchor · $100 per average kW discharged (Alliant, Iowa/Wisconsin) · $105/kW-year (DTE battery pilot) · up to $110/kW on average contribution across events (APS Storage Rewards) · $125 per average kW/year (Entergy Louisiana) · $200/kW new and $100/kW existing, one-time, trued up to delivered output (New Jersey Phase 2 Block 1) · and up to $350,000/MW-year participant-facing in the Google–Voltus arrangement, which is not a utility price. Every utility-set figure exceeds the generic anchor. Three of them — Alliant, APS, Entergy — pay on average contribution or average kW across a season rather than depth per event, which is the opposite of how a capacity market accredits. Reconcile the basis before the level.

Carried benchmarks, cited rather than re-reported. The $325/MW-day 2028/2029 PJM clear (~$118,625/MW-year) against PJM’s $555/MW-day backstop cap and the 6,831 MW shortfall; PJM’s recovered qualifying-offset criteria — executed EDC agreement covering metering, telemetry and baseline measurement, three-year minimum registration history, binding fifteen-year site-level commitments, letters of intent not accepted, and no location that supported a Demand Resource in a prior Delivery Year unless the increase comes from new generation or storage investment there; FERC’s October 23 MISO LMR penalty order effective June 1; the CPUC’s ~$608M two-year DR bridge funding and its Societal Cost Test exemption; R.21-06-017’s orchestration framework, CAISO-visible telemetry mandate and 2027–2028 Load Flexibility Scenario; SB 913’s June 2028 resource-adequacy deadline and its device-level telemetry precondition; NERC’s October 26 Board adoption of the computational-load standards; the ~8% winter capacity accreditation in Missouri’s Ameren docket; Google–Voltus >1 GW to 2040; SRP’s 60% pre-payment; FirstEnergy’s ~$250M/GW transmission multiplier; Brattle’s 4% benchmark (≈40 MW of residential flexible capacity per GW of system peak ≈ $2.6M/year per GW at ~$66/kW-year); Ember’s ~3% shiftable data-center share and the 0.55-vs-2.7 GW Texas worked case; CAISO’s four FILI instrumentation requirements; the ~$66/kW-year and ~$20.70/MWh avoided-cost anchors; and OpenAI’s 1,000 MW curtailable at Effingham County for 25 years with no published measurement method — still the largest flexibility commitment on the North American record and still unaccreditable.


🔬 EPRI Research Spotlight

No EPRI publication dated inside the window — the eighteenth consecutive week — but this week a targeted search finally located the standard that will decide whether demand-side resources are credited at all, and EPRI’s own framework is again the instrument it is missing. A direct check returned no EPRI publication dated November 9–13. What the check did resolve is the single highest-value tracking item in this series, open for three runs.

The energy assurance standard is NERC Project 2024-02, “Planning Energy Assurance,” and the standard it will produce is titled “Long-Term Energy Reliability Assessments.” The number is still “[Standard Name and Number TBD]” in the posted draft. This is the instrument NERC CEO Jim Robb named at FERC’s October 2026 Reliability Technical Conference as the first of four remedies — forecasting anchored on dependable energy under stress rather than nameplate — in a statement whose six risks and four remedies mentioned demand response, VPPs and load flexibility nowhere, and DERs only as an integration risk. This series has spent three runs trying to establish whether demand-side resources are in scope. They are, in exactly one place, and the place matters.

The initial draft (formal posting 1, October 2025) applies to Planning Coordinators, Resource Planners and Transmission Planners. R1 requires the Planning Coordinator to document a process for a biennial assessment covering one year from planning years one through five and one year from years six through ten. R2 requires a documented methodology covering supply and demand, whose required elements include demand forecasts, resource capabilities, imports and exports, and transmission constraints — and, as R2.2, “Demand Response programs, capabilities, and operational limitations.” R2.3.5 requires the methodology to address “Energy storage resource constraints.” R3 requires documented criteria for unacceptable reliability risk, covering the frequency and duration of unserved-load events, maximum unserved load and total unserved energy. R4 through R8 run the assessment, distribute results to regulators and planners within 90 days, require Corrective Action Plans within 90 days of receipt, submission to regulators within 30 days after that, and distribution back to the Planning Coordinator within 120 days. Milestones: a 51-day formal comment period with initial ballot, October 21 – December 10, 2025; anticipated Board of Trustees adoption August 2026; and NERC’s 2027–2029 Reliability Standards Development Plan lists the project among planned completions, with the goal “completed by October 2026” — about 12 subject-matter experts and ~2,496 work hours.

So here is the answer, stated precisely, because the precision is the whole point. Demand Response is an input to a methodology. It is not a resource with a credit. R2.2 obliges a Planning Coordinator to address DR “programs, capabilities, and operational limitations” in how it builds its assessment. It does not say how capability is to be quantified, what evidence establishes it, how operational limitations are to be derated, or what a demand-side resource must demonstrate to count toward avoiding an unserved-load event. DERs, VPPs and load flexibility do not appear in the draft at all. And R3 — the requirement that actually defines risk — is written entirely in terms of unserved load and unserved energy, which is the correct frame and the one on which a dispatchable demand-side portfolio should score well, if the methodology in R2 credits it. The entire question of demand-side credit has been delegated, without criteria, to a methodology each Planning Coordinator writes for itself.

Which is exactly the gap EPRI’s FlexMosaic framework was built to fill, and the second week running that this Spotlight has to point at it. FlexMosaic — announced at CERAWeek in March 2026, built with 65-plus utilities, system operators, regulators and hyperscalers — defines five flexibility classes: Class A, support during infrequent extreme system stress; Class B, management of daily or weekly demand peaks; Class C, response to long-lasting energy shortages; Class D, response to sudden supply or demand swings; Class E, frequency stabilization. Class membership is determined by notification time, response duration and frequency, and the depth and speed of the response. Those four parameters are precisely the “capabilities and operational limitations” R2.2 requires a methodology to address and declines to define. Class C — response to long-lasting energy shortages — is a direct match for the multi-day, energy-limited risk R3 measures as total unserved energy, and it is the class a four-hour battery fleet does not serve while a behavioral or curtailable-load program does.

The deliverable is therefore specific, cheap and available now: file FlexMosaic class definitions into your Planning Coordinator’s R2 methodology as the quantification scheme for R2.2. Do it before the methodology is written rather than after, because R2 methodologies will be documented once and inherited for years, and a methodology that treats demand response as a single undifferentiated input will derate a Class C resource to a Class A assumption and never show the difference. Three supporting moves follow. One: the evidence you need is the measured winter record this series has asked for since October — morning-ramp delivery, multi-hour sustained output, state of charge at call time, cold-weather derate, per event and per site — because R3 is scored on duration and total unserved energy, not on peak coincidence. Two: pair the filing with PJM’s own winter sentence from this week — reliability holds at 152,200 MW only with pre-emergency and emergency demand response in the stack — which is an operator saying in a seasonal assessment what the standard’s input list does not say. Three: comment on the project itself. The draft is at or past its planned completion, so the open channel is the additional comment period and final ballot, both still listed as TBD, plus the Corrective Action Plan provisions in R6 through R8, where demand-side mitigations are eligible as “mitigating actions” a planner can take or recommend — including those needing regulatory approval. R6 is the sentence that lets a Corrective Action Plan be a demand-side procurement, and nobody in the demand-side industry appears to be arguing for it. (Targeted gap-fill — NERC Project 2024-02 initial draft, October 2025, and the NERC Standards Committee October 21, 2026 agenda package; EPRI FlexMosaic is a carried framework reported in this series through the spring and summer, applied here. No in-window EPRI publication)


🚩 Utility-Sector Relevance Flags

⚑ The Energy Assurance Standard Names Demand Response Once, As A Methodology Input — And Delegates The Credit, With No Criteria, To Each Planning Coordinator
Topic: Resource Accreditation / Reliability Standards
Relevance: The energy assurance standard is NERC Project 2024-02, “Planning Energy Assurance,” producing a standard titled “Long-Term Energy Reliability Assessments” (number TBD). Applicability: Planning Coordinator, Resource Planner, Transmission Planner. R2 requires a documented supply-and-demand methodology whose elements include, at R2.2, “Demand Response programs, capabilities, and operational limitations” and, at R2.3.5, “Energy storage resource constraints.” R3 defines unacceptable reliability risk in terms of frequency and duration of unserved-load events, maximum unserved load and total unserved energy. R6–R8 require Corrective Action Plans within 90 days, submitted to regulators within 30 days and returned to the Planning Coordinator within 120 days, including mitigating actions a planner can take or recommend to third parties. DERs, VPPs and load flexibility appear nowhere in the draft. Initial draft posted October 2025 with a 51-day comment period and initial ballot October 21 – December 10, 2025; anticipated Board adoption August 2026; RSDP goal “completed by October 2026”; further comment period and final ballot still TBD. ~12 SMEs, ~2,496 work hours.
Action Signal: Engage — This is the proceeding this series has called the highest-value tracking item for a month, and the answer is that the credit question has been delegated without criteria. File FlexMosaic’s five classes — set by notification time, response duration and frequency, and depth and speed of response — into your Planning Coordinator’s R2 methodology as the quantification scheme for R2.2, before the methodology is documented rather than after, because an R2 methodology is written once and inherited for years. A methodology that treats demand response as a single undifferentiated input will derate a Class C long-duration-shortage resource to a Class A rare-stress assumption and never show the difference. Then argue R6 explicitly: a Corrective Action Plan may consist of demand-side mitigation, including measures requiring regulatory approval — that sentence makes a CAP a potential demand-side procurement, and no one in the demand-side industry appears to be arguing for it. Bring the evidence R3 actually scores: duration and total unserved energy, not peak coincidence — which is the measured winter record (morning-ramp delivery, multi-hour sustained output, state of charge at call time, cold-weather derate, per event and per site). Use the open comment period and final ballot; both are still TBD.

⚑ New Jersey Found 82% Of A 547 MW Peak-Reduction Target In Devices Ratepayers Have Already Paid For
Topic: VPP Program Design / Enrollment Economics
Relevance: The New Jersey BPU unanimously approved “PowerUp New Jersey” on November 12 — an interim VPP program targeting a 3% peak-demand reduction at each of the state’s four EDCs, 547.1 MW combined — and found that up to 82% of each utility’s target is reachable from devices already funded through state energy-efficiency and EV-charger programs if about three-quarters of eligible devices enroll. Customers are paid directly by the utility or through a distributed energy resource manager to allow remote curtailment of smart devices or discharge of home batteries and solar. Interim programs launch by October 1, 2027 and run to September 30, 2029. The companion Phase 2, Block 1 storage incentive pays $200/kW new and $100/kW existing (150 MW new + 100 MW existing, trued up to delivered output) toward 2,000 MW of storage by 2030. No per-utility MW split and no PJM wholesale-market participation addressed.
Action Signal: Implement — Re-price your residential DSM business case against the installed base rather than new devices, and report the enrollment gap in MW. This is the first commission finding in this series to quantify what the marginal cost of residential flexible capacity actually is, and it is enrollment, telemetry and settlement — not hardware. It is the third independent confirmation from a third kind of institution: Brattle located the lever in OEM-led enrollment of installed devices; California’s ChargeWise Phase 2 subsidizes user acquisition and onboarding rather than hardware; New Jersey has now put a percentage on the already-purchased fleet. Two further actions. Specify for two settlement counterparties even where your current order names one — New Jersey’s dual channel makes it two of four tracked states going dual (with Virginia) against Minnesota and Nevada utility-only, and New Jersey got there by changing its mind, so IEEE 2030.5 plus dual-counterparty settlement stays pass/fail. And stress-test the enrollment assumption honestly: 75% device enrollment has never been achieved by any residential program in this series — Octopus’s 50–60% among smart-device owners is the high-water mark — so model the 82% at 40% and 55% enrollment and show the commission the sensitivity before someone else does. Note the term mismatch for PJM purposes: interim programs expire September 30, 2029 while qualifying-offset registration runs to June 1, 2032.

⚑ Three Vendors Published Demand-Flexibility Depth In One Week And Not One Figure Is Independently Verified
Topic: M&V / Portfolio Performance Claims
Relevance: Oracle (November 11): a record ~1 GW of 2026 peak reduction via behavioral DR and peak-time rebates across 73 events (up from 49 in 2025) and 4.3 million customers (up 30%), averaging 16,592 kW per event; Puget Sound Energy up to 16.6 MW/event with a new $0.50/kWh credit; CPS Energy 350,000+ enrolled, ~19 MW across 12 August events. Uplight (November 11): 15,629 MWh shifted over 1,807 events; resources called ~5× more often in July–August; residential dispatchable capacity +75%; 735 MW peak dispatched the week of August 30 (~2× the 2025 peak week); batteries +70%, EVs +2×; behavioral DR enrollments +5×, delivering >20% of total load shift with no customer hardware; an unnamed Midwestern utility delivered 800 MW across 9 events; goals of 20 GW under management and $1B in savings. Octopus Energy US (November 6, updated November 8): 50–60% DR participation among smart-device owners against a 5–10% historical benchmark, on “all-carrot,” opt-out design and a ~10% peak-to-off-peak household shift. All figures are vendor-reported with no third-party M&V; Uplight’s two cited utilities are unnamed and its total GW under management is undisclosed; Octopus’s rate covers smart-device owners only with no disclosed method, opt-in/opt-out split or period.
Action Signal: Implement — This is the answer to the frequency-without-depth problem and it arrived in the wrong currency, so convert it before using it. For each claim demand the four things that make a depth number citable: the measurement point (grid interface, service point or per asset), the baseline method, delivered-versus-committed MW per event rather than aggregate MWh or a percentage of commitment, and the named utility and program whose filed evaluation will confirm it. Until those exist, carry these as capability with a timeline, not capacity — the same treatment this series applies to every aggregator target. Two things in the cluster are usable now. Behavioral DR delivering over 20% of a portfolio’s load shift with no customer hardware is a cost structure, not a performance claim, and it is the cheapest line in any DSM stack — pair it with Brattle’s 4% benchmark (≈40 MW per GW of system peak) and ask what share of your own target could come from opt-out messaging before a single device is subsidized. And Chaset’s open question is the right acceptance test to write into a contract: does participation hold as call frequency rises? PJM called DR on nine days in 2026 against eight across all of 2015–2024, so re-price every program on nine calls a year and put a measured attrition clause in the enrollment terms.

⚑ A Senate Bill Would Make Order 2222 Compliance Statutory In 18 Months — And Its Own Authors Say Accreditation, Not Entry, Is The Fight
Topic: Wholesale Market Access / Federal Legislation
Relevance: BAAJA (S. 5653), introduced September 30 by Sen. Capito with Lee, Whitehouse, Heinrich, now in Senate ENR. Section 2108 / new FPA §229: RTOs and ISOs must admit qualifying aggregations to wholesale markets within 18 months of enactment, compliance filings due in 180 days, any blocking rule deemed unjust and unreasonable. New FPA §230: distribution utilities may not use control of the distribution system or retail sales to block participation or exports, and must share the same interval metering and telemetry the wholesale market uses (safety/reliability actions preserved); obligations yield to state law; FERC gains no distribution-interconnection-rate authority. “Qualifying behind-the-meter distributed energy system facility”: behind a retail meter, dispatchable within 15 minutes, ≥5 kW; aggregation = ≥2 facilities totaling ≥100 kW. Dual state/wholesale participation permitted but not paid twice per interval (self-certified; FERC penalties up to 3×). DOE voluntary local-permitting program at $20M/year FY2026–2031 for solar, ≥2 kWh batteries, EV chargers. Separately, 300+ solar and storage companies backed the bill’s permitting title (NEPA/CWA/ESA amendments plus FERC authority to permit some transmission); SEIA projects 167 GW of additional solar 2028–2035 against $120B said to be permitting-stalled. The authors flag capacity accreditation as the real battleground and expect the data-sharing mandate to be “honored in form but slow in practice.”
Action Signal: Engage — Comment on the accreditation gap now, in the venues that are already open, rather than waiting on a bill whose House counterpart stalled the day after passing 417-3. A statutory right of entry settles nothing about credit, which is the same conclusion this series reached from PJM’s narrowed offset definition, MISO’s stiffer LMR non-performance penalty and the ~8% default winter accreditation live in Missouri — and now from the bill’s own analysts. Treat the thresholds as an eligibility screen and test your portfolio against them today: 5 kW per facility, 100 kW per aggregation, dispatchable within 15 minutes. A 15-minute requirement writes out slow thermal load and writes in batteries, EV chargers and thermostats, so inventory which of your enrolled assets could certify and which could not. And take the authors’ prediction seriously as a planning assumption: if the data-sharing mandate is slow, the operative asset is the metering and telemetry record you already hold — the identical thing PJM’s qualifying-offset criteria make dispositive. Two instruments, one federal and one market, reduce to the same question: can you document the asset? Caveat every citation: draft legislation, uncertain prospects, open questions on rooftop solar, thermostat load and uncapped on-site gas, no system MW figures, and the permitting title covers gas, oil and coal alongside clean technologies.

⚑ California Is Adopting Guiding Principles For The Calculator Every DSM Program In The State Is Tested Against
Topic: Avoided-Cost Methodology / DER Cost-Effectiveness
Relevance: On the CPUC’s November 19 voting meeting, in R.22-11-013, the Commission would adopt guiding principles — accuracy, consistency and transparency — for the Avoided Cost Calculator used in DER cost-effectiveness analysis, plus a companion Distributional Equity Analysis. The ACC monetizes the avoided generation, capacity, T&D and ancillary-service value every DSM, DR and storage program’s benefit-cost test runs against. The same agenda carries Resolution E-5455, approving with modifications PG&E’s agreement to energize a new 250 MW Google load in San Jose, adding a $600,000/MW capital advance and capping annual refunds at 100% of net revenues received from the customer for ratepayer protection. Both are agenda items on draft or forthcoming decisions, not final orders; the source is a law-firm digest, and the principles are stated at a high level without the methodological changes that would move a given program’s benefit-cost ratio.
Action Signal: Engage — A principles order is the upstream document for every later methodology fight, so establish now — while the criteria are being written — that a principle of “accuracy” requires locational and hourly granularity rather than a system-average avoided cost. The principles will be cited back at you during the methodology proceeding; get them pointed the right way first. Keep this series’ two standing audits running against it: check every avoided-cost figure for ACC vintage (now Resolution E-5476, the final 2026 ACC) and for which test your commission actually applies (TRC/PAC/RIM/PCT) — a distinction California sharpened in October by authorizing ~$608M of two-year DR bridge funding and exempting it from the Societal Cost Test in the same decision. And use the agenda pairing as a cost-allocation exhibit: at $600,000/MW, a 250 MW load advances $150M, the same order of magnitude as the avoided transmission a 500 MW demand-side deferral claims at FirstEnergy’s own ~$250M/GW multiplier. Put the two numbers on one page. Get the primary documents in R.22-11-013 and E-5455 after November 19.

⚑ ERCOT Wants To Curtail Data Centers Before Scarcity — And Has Never Once Used The Authority It Already Has
Topic: Large-Load Curtailment / Emergency Hierarchy
Relevance: ERCOT President and CEO Pablo Vegas said on November 11 he will ask the Texas Legislature for authority to require large data-center loads to reduce consumption earlier — at a Level 1 condition or before the first sign of scarcity — rather than only at the second step of a three-stage emergency, as current rules allow. SB 6 (2025) already lets ERCOT order large loads to curtail on 24 hours’ notice, and ERCOT has not yet exercised it. Data centers are framed as a low-single-digit-GW pool of flexible winter supply that voluntarily scaled back during last winter’s severe weather. No MW threshold or trigger definition specified; this is a stated legislative intention, not a bill, rule or tariff. Texas Docket 59220 established SB 6 curtailment as uncompensated and uncapped.
Action Signal: Watch — Do not revalue your portfolio on this, and keep the standing rule intact: count mandated large-load curtailment as emergency risk reduction, never as DR capacity with an avoided-cost value. The addressable market is the customer response to an earlier uncompensated obligation — on-site storage, on-site generation, flexibility-priced contracts — not the mandate. What is actionable now is the cross-market pattern, three jurisdictions in two weeks: ERCOT seeking earlier curtailment authority, Pennsylvania’s live emergency-load-control proceeding, and PJM’s winter outlook leaning on emergency DR plus DOE-authorized large-load backup. The common movement is away from treating large flexible loads as exempt “critical load” and toward treating them as curtailable, which is the single reclassification that gives verified demand-side flexibility its avoided-capacity value by moving a self-curtailing customer out of the first-to-be-shed tier. So act on the sequencing rather than the statute: screen every large-load tariff and emergency operating procedure in your territory on whether enrolled DR is called before or after unsupported large load, and get the dispatch-order language into both documents with control-room verification.

⚑ PJM Says Its Winter Case Does Not Close Without Emergency Demand Response — Quote The Operator, Not The Advocate
Topic: Winter Reliability / Demand-Side Credit
Relevance: PJM’s 2026–27 winter assessment studies a high-stress cold peak of up to 152,200 MW against the record winter peak of 143,700 MW (January 22, 2025), expects adequate supply under normal conditions, and concludes reliability can be maintained even at the 152,200 MW stress case only with pre-emergency and emergency demand response in the stack. DOE’s standing authority lets PJM order data centers and other large loads to run backup generators as a last resort. Localized delivery risk is flagged in the transmission-constrained BGE, Pepco and Dominion zones (MD/VA/DC). NWS forecasts a slightly warmer winter; no reserve-margin percentage disclosed.
Action Signal: Implement — Put this sentence, with the operator’s attribution, into the record of every proceeding where demand-side credit is being set. It is the sharpest available counterweight to the asymmetry this week exposed: the energy assurance standard treats Demand Response as one line in a methodology input list, while the operator of the largest market in North America says its winter stress case does not close without it. Two operational follow-ons. The BGE/Pepco/Dominion call-out is a locational procurement case — imports cannot relieve a local delivery constraint — which is the same logic behind Toronto Hydro’s station-by-station LDR auction and behind FERC’s objection to backstop capacity delivered anywhere rather than in the constrained zone where the load sits; so site and bid demand-side resources by constrained zone, and price the locational premium separately. And build the measured winter record now — morning-ramp delivery, multi-hour sustained output, state of charge at call time, cold-weather derate, per event and per site — because winter is the specific attack on demand-side credit and a mild winter is a bad outcome for a resource whose credibility depends on being called.

⚑ Six Utilities Now Price Residential Flexibility Above The Avoided-Cost Anchor, And Half Of Them Pay On Availability Rather Than Depth
Topic: Incentive Design / Benchmark Reconciliation
Relevance: APS’s Storage Rewards Pilot pays up to $110/kW based on a participant’s average contribution across demand-response events, and SRP’s Battery Partner Program rewards customers for letting SRP discharge FranklinWH or Tesla batteries at peak — both coordinated through EnergyHub’s DERMS, with Palmetto LightReach leasing lowering the up-front barrier, and APS, SRP and TEP all approving FranklinWH’s 15 kWh aPower2 and ConnectDER’s IslandDER meter-socket adapter (15–30-minute install, no panel swap; compatible with FranklinWH, Lunar, SolarEdge, EcoFlow). The stack now reads: ~$66/kW-year anchor · $100 per average kW discharged (Alliant IA/WI) · $105/kW-year (DTE pilot) · up to $110/kW on average contribution (APS) · $125 per average kW/year (Entergy Louisiana) · $200/kW new and $100/kW existing one-time (New Jersey Block 1). No MW or customer-count totals disclosed for the Arizona programs; the ~47% Arizona battery attachment rate is a ConnectDER estimate and the 8% QoQ Q4 2025 residential-storage growth is Wood Mackenzie/ACP.
Action Signal: Implement — Reconcile the basis before the level. Alliant, APS and Entergy all pay on average kW across a season or across events rather than depth per event, which is the opposite of how a capacity market accredits — so model payment and accreditation separately and never assume they move together. On the level: every utility-set figure in the stack exceeds the generic anchor, by roughly 1.5× to 3×, which means in any proceeding where you defend ~$66/kW-year you now have to say whether the gap is deliberate acquisition buy-down or a genuinely higher local avoided cost, because six jurisdictions have priced it higher. Then take the deployment lesson: put utility-side equipment approvals on the DERMS critical path alongside interconnection and local permitting. A 15–30-minute meter-socket install with no panel swap removes a cost and scheduling bottleneck that no orchestration logic can route around, and a DERMS books capacity in a filing only once devices are actually and administratively connected — the VPP ceiling is set as much by adapter approvals and financing as by software. Read it beside APS’s own IRP projecting BTM storage from 69 MW to 1.2 GW by 2045 while rooftop solar runs ~30% below forecast: these approvals are the mechanism by which the storage half of that forecast happens or does not.

⚑ “Handles High-DER Conditions Without Requiring Telemetry From Every Resource” Is An Operating Claim, Not A Settlement Claim — Write Both Lines And Keep Them Apart
Topic: DERMS/ADMS Specification / Observability vs Settlement
Relevance: Hitachi Energy was named a Leader in the 2026 IDC MarketScape for Worldwide ADMS (Doc #US54133826, October 2026), which scored eight ADMS providers. IDC credited ABB Power Grids-inherited grid-automation engineering, a shared transmission/distribution SCADA foundation, and distribution state estimation that “handles high-DER conditions without requiring telemetry from every resource.” No utility deployments, no market-size or share figures and no DERMS product are named — the release describes DER-aware operation rather than a distinct DERMS module — and a “Leader” placement is a point-in-time analyst judgment, not a procurement. This is the seventh consecutive week with no DERMS deployment item, and the first standalone ADMS item in that run.
Action Signal: Implement — Put two separate requirements in the RFP and do not let a vendor collapse them: inferred feeder-level state for operations and hosting-capacity assessment, and measured per-asset telemetry for accreditation and settlement. The state-estimation claim is genuinely useful on the operating side — as feeders fill with distributed solar, storage and enrolled devices the binding constraint shifts from control to observability, and inference lowers the metering and communications burden an operating spec must carry. It lowers nothing on the settlement side. Under SB 913 device-level telemetry is a statutory precondition for resource-adequacy qualification in California from June 2028, and the settlement quantity is per-asset while a net meter measures the net — so metering and telemetry fidelity remains the primary DERMS specification line, and a demonstration of two aggregators settling independently behind one service point remains pass/fail. A vendor offering inference as a substitute for measurement is offering you an operating tool and a disallowance. Use the IDC report for how the ADMS field is being graded and for the one capability claim worth testing on your own feeders, not as a deployment signal.

⚑ One Behind-The-Meter Device, Nine Or Ten Internal Departments — And 2,800 Co-ops And Munis With No VPP Program At All
Topic: Program Delivery Cost / Under-Served Channel
Relevance: A Carnegie Mellon Energy Week 2026 synthesis finds that connecting a single behind-the-meter device can involve nine or ten internal utility departments, and that roughly 800 distribution cooperatives and 2,000 municipal utilities mostly lack virtual-power-plant programs at all. Context given: a single planned upstate-New York semiconductor facility will need as much electricity as New Hampshire and Vermont combined; NERC’s 2025 finding that peak-demand growth in several US regions is outpacing new generation for the first time in years; the 3,500 MW SunZia line took 17 years to approve; and storage plus demand response are credited with preventing price spikes in a recent Texas winter storm, against Uri in 2021 when ERCOT prices ran ~400× above average for four consecutive days. Working small-utility models: Danville Utilities (VA), using distribution batteries to cut transmission peak charges, and six Massachusetts municipal utilities. Panel-synthesis commentary, not a filed program or dataset.
Action Signal: Implement — Count the departments in your own interconnection-to-enrollment path and carry the count as a program cost line, because it is the one cost a larger device incentive cannot buy down. It is also the mechanism behind the enrollment gap New Jersey quantified from the other end: ratepayers may already own the devices, but a nine-department process is what stands between owning them and dispatching them. Then work the channel. 2,800 co-ops and munis without VPP programs is the size of the under-served market, and this series has repeatedly found that co-ops and munis move faster than IOUs once they act — Base Power’s Central Georgia EMC deal with the co-op holding the dispatch right, Budderfly’s Cadillac, Michigan contract, FMPA specifying connectivity-model requirements across 32 member utilities. For those utilities the remaining work is analytics, valuation and dispatch rather than hardware, which argues for engaging the joint-action agency rather than the member utility — this series’ standing instruction where public power sets GIS or connectivity-model requirements. And treat grid-edge visibility as the binding constraint it is named as here: assets behind the customer’s meter give the utility little visibility into actual performance, which is exactly the gap a DERMS is procured to close and exactly what an unfunded analytics layer leaves open.

⚑ 18 Consecutive Days Of Dispatch On 70,000 Customer Batteries Is The Multi-Day DERMS Test Nobody Publishes
Topic: VPP Deployment / Fleet Orchestration
Relevance: Sunrun’s Puerto Rico residential-battery VPP has grown from ~4,000 customers in 2024 to more than 60,000 — nearly 70,000 home batteries delivering 45 MW of dispatchable power over four-hour windows — dispatched more than 60 times since April and for 18 consecutive days beginning October 13 during unusually warm autumn weather, to address shortfalls and help avoid rolling blackouts. Walmart’s October 30 strategic investment in Base Power extends a provider-owned home-battery-and-electricity retail model to Walmart and Sam’s Club customers beginning in Texas, with a 20% discount on the standard installation fee and Base installing, owning and maintaining the battery. Per-battery kWh ratings and customer compensation amounts are undisclosed.
Action Signal: Implement — Ask every DERMS vendor for the multi-day case, and use this as the benchmark: what happens to available capacity on day twelve, how is state of charge rationed across a fleet when recharge windows are compressed, and what does the utility see when the commitment is not met? Eighteen consecutive days is the closest thing in this series to a published endurance test of fleet-scale state-of-charge management, dispatch and verification across tens of thousands of endpoints — the core DERMS function exercised for most of three weeks rather than an hour. The capacity read is real: 45 MW of callable, four-hour residential storage behaving as a repeatable grid-stress resource is deferred peaker value from assets the utility neither owns nor financed. On Walmart, take the channel, not the investment: provider-owned hardware attacks customer up-front cost and asset ownership, the largest barrier to residential enrollment, and is the scaling path the DOE’s 80–160 GW-by-2030 VPP target depends on against today’s ~30–60 GW. And raise the liability question before the capacity question: eighteen straight days of discharge on customer-owned equipment with no published degradation accounting and no compensation adjustment is the battery-wear-versus-compensation problem this series has flagged since the summer, now with an operating record attached.

⚑ Oncor Is Sizing A Capital Plan Against A 293 GW Queue Of Which 15% Has Cleared Any Screen
Topic: Large-Load Forecast Risk / Stranded-Cost Exposure
Relevance: Oncor reports ~293 GW of data-center requests in the ERCOT transmission queue as of September 30, against a new 33 GW system peak (up 6%), and expects its 2027–2031 base capital plan to rise ~20% from $47.5B, with 2026 spend ~$9B ($6.3B through September), its largest annual program. ERCOT’s Batch Zero (approved June 2026; ≥75 MW grouped for joint study) has conditionally classified ~24 GW as base load (10 GW already energized) and ~21 GW as studied — ~45 GW of the 293 GW, roughly 15%. Applicant review due December; final Batch Zero transmission plan fall 2027. Oncor holds ~$6B in customer collateral, including ~$2B tied to Batch Zero and excluded from the base plan. Does not address demand response or distribution hosting capacity.
Action Signal: Watch — Use the 293-to-45 ratio, not the headline. A capital plan sized against an unscreened queue is sized against an artifact, and this series has found that every financial-commitment filter collapses a large-load pipeline by roughly an order of magnitude (AEP Ohio’s 30 GW to 13 GW is the cleanest prior case). The deferrable, cancellable demand-side alternative is the lower-regret purchase at every point on that curve — make the “risk of unrealized loads” the argument, in the utility’s own words where a filing supplies them. Then press the collateral. $6B held, ~$2B tied to Batch Zero, is a cost-causation instrument shown working at scale — the fourth utility-designed one in six weeks alongside SRP’s 60% pre-payment, Xcel’s Clean Energy Accelerator Charge and OpenAI’s self-funded upgrades — so ask of it what this series asks of each: can the instrument fund flexibility and storage, or only wires and generation? A mechanism built on demand-side cost causation that may only be spent on supply-side assets writes the demand side out of its own logic. And keep the three ERCOT queue figures reconciled: Oncor’s 293 GW transmission-queue share, ERCOT’s ~474 GW audited queue and the 191.8 GW of large-load studies are three different denominators, not one number.

⚑ FERC Twice Ruled That Emergency-Order Costs Fall On The Load Inside The Emergency Footprint — $26M For A Unit That Never Ran
Topic: Cost Causation / Section 202(c) Allocation
Relevance: In an October 19, 2026 order FERC again rejected without prejudice TransAlta’s proposal to recover the fixed costs of keeping the 729.9 MW Centralia Unit 2 coal plant available under DOE Section 202(c) orders of December 16, 2025 and March 16, 2026, at the direction of Gridforce Energy Management (balancing authority) and SPP (reliability coordinator). TransAlta’s right to recover was affirmed; the allocation was rejected. The unit was never dispatched, so only a fixed charge of $6,395,348, split equally between Gridforce and SPP, was sought. FERC held the equal split violates cost causation because the two entities’ service areas do not match the emergency’s WECC-Northwest geographic scope, and directed the costs be spread across load in that area — the same basis on which a separate October 1 order rejected ~$19.9M for the first order’s period. The rejection is procedural, deciding allocation only and expressly not whether the costs are just and reasonable. Concerns a generation unit, not a demand-side resource.
Action Signal: Watch — Cite it wherever a utility proposes to socialize reliability-must-run or emergency-order costs system-wide: FERC has now twice required the cost footprint to match the emergency footprint, on a cost-causation rationale that reaches any emergency cost, not only this one. That is the identical principle contested in data-center curtailment and large-load tariff fights across this series, arriving through a generation docket where it is easier to win. And carry the arithmetic as the standing comparison: roughly $26M across two orders to hold one never-dispatched 729.9 MW unit available over two winters, for zero delivered energy. That is the expensive, last-resort version of what a planned, compensated demand-side program delivers with a settlement record attached — the 2026 Section 202(c) docket now exceeds fifty orders and has accredited nothing, which remains the cleanest statement of why directive flexibility builds no asset while compensated flexibility does.

⚑ NERC’s Registry Petition Is In And Expedited — 68,288 MW, About 86% Of Projected End-2027 Computational Load
Topic: Large-Load Registration / Observability
Relevance: NERC petitioned FERC on October 21, 2026 in Docket RR27-1-000 for Rules of Procedure revisions adding two functional entity types to the NERC Compliance Registry: Computational Load Owner (CLO) and Computational Load Operator (CLOP). Registration applies at a total connected load of at least 50 MW supplied at 100 kV or greater from one or more points of connection — the two-part test FERC declined to name — and NERC estimates ~68,288 MW, roughly 86% of the computational load projected in service by end-2027. NERC requested expedited action; proposed Reliability Standards and implementation plans will be filed later. The petition responds to FERC’s July 16, 2026 order. Neither the petition as reported nor the categories mention demand response, load flexibility or DERs.
Action Signal: Watch — This closes a tracking item carried since September and confirms the sequencing this series predicted: the registry binds first and the standards follow, so the Rules of Procedure track was the one to watch. The standards NERC’s board adopted on October 26 bound nobody until this landed. Carry 68,288 MW at ~86% coverage as the first official sizing of the regulated computational-load population — it is the denominator for any argument about how much large load is observable, and it belongs beside the queue figures above precisely because it is the one that reflects connected reality rather than requests. And note the asymmetry one more time, because it is this week’s throughline: registration creates observability obligations for large loads and credits none of their flexibility, exactly as the energy assurance standard lists demand response as an input and credits none of its capability. Watch the comment docket and the follow-on standards filing, and read the implementation plan for whether a registered CLO’s flexibility obligations ever acquire a measurement standard.


📌 Sources

November 9, 2026 Entry
– MGRID — PJM Winter Outlook Projects Narrow Reserve Margins and Relies on Emergency Demand Response (November 7, 2026) (first appearance; distinct from MISO’s 2026–27 Winter Planning Resource Auction carried in last week’s digest — different operator, different proceeding, different source)
– MGRID — Walmart Backs Base Power as Sunrun’s Puerto Rico Battery VPP Doubles to 60,000 Customers (November 9, 2026) (first appearance in the tracked record for the Puerto Rico fleet; the September–October California dispatch figures carried in this series — 580 MW, 601 MW, 745 MW — are a different fleet and are not comparable, and Base Power’s Central Georgia EMC co-op deal reported in the October 23 digest is a different arrangement)
– MGRID — PG&E’s $2 Billion Grid Cut Puts California Renewable Connections on Hold (November 8, 2026) (first appearance)
– MGRID — New York Awards Nearly $13 Million to 15 Energy Storage Projects (November 8, 2026) (first appearance; distinct from the New York storage-permitting and 576-MW-against-6-GW material carried in the October 30 digest)

November 10, 2026 Entry
– MGRID — Oncor Expects 20% Capital Plan Increase as Texas Data Center Requests Reach 293 Gigawatts (November 9, 2026) (first appearance; the ERCOT ~474 GW audited queue and the 191.8 GW of large-load studies carried in prior digests are different denominators and are reconciled in Benchmark Hygiene)
– MGRID — Solar and Storage Companies Back Permitting Bill Tied to 167 GW of Additional Solar (November 10, 2026) (first appearance; consolidated with the Foley Hoag analysis below as one BAAJA item — two halves of one bill, reported as one item)
– MGRID — EDP Renewables Breaks Ground on 250-MW Arkansas Solar Project for Meta (November 7, 2026) (first appearance; carried in one paragraph for the annual-match-versus-hourly-firm gap, not as a DER item)
– Environment+Energy Leader — Toronto Hydro Secures 12 MW to Manage Peak Grid Demand (published February 22, 2026; page updated November 9, 2026) (entered the log November 9 under the “most recent relevant piece” rule from a page refreshed that day; the procurement is from October 2025 and the publication from February 2026, so it is carried as a dated benchmark in Benchmark Hygiene and is NOT reported as a new development)

November 11, 2026 Entry
– Foley Hoag — BAAJA Blast Part 4: Plugging In — Distributed Energy Resources and Virtual Power Plants Under the Bipartisan American Affordability and Jobs Act (November 6, 2026) (first appearance for the bill’s DER/VPP wholesale-market-access provisions; consolidated with the MGRID permitting-letter item above as one item — deduplicated)
– PA Environment Daily — PUC Updates Comment Schedule On New Tools To Manage Data Centers During Emergency Periods; Nov. 17 Technical Conference On Data Center Costs (November 5, 2026) (movement — the October 1 tentative order in Docket M-2026-3064961 was reported in the late-September/early-October digest; this is the distinct November 5 clarification order narrowing the proceeding after FERC’s rejection of parts of PJM’s IRAS)
– MarketScale — Octopus Energy US Gets 50% to 60% of Smart-Device Owners Into Demand Response (November 6, 2026; updated November 8, 2026) (first appearance; distinct from the Uplight summer-portfolio release below — enrollment rates versus portfolio output, same combined company, two datasets)
– Environment+Energy Leader — Distributed Energy Resources and the Load Growth Gap (updated November 11, 2026) (first appearance)

November 12, 2026 Entry
– Bloomberg — Texas Grid Boss Wants Data Centers to Cut Power Before Blackouts (November 11, 2026) (first appearance; the SB 6 rules and Docket 59220’s uncompensated-and-uncapped finding are carried benchmarks cited in one line)
– Interesting Engineering — Oracle Helps US Utilities Cut Peak Power Demand by Record 1 GW in 2026 (November 11, 2026) (first appearance)
– Energy-Storage.news — Arizona Utilities Approve New Technologies to Accelerate Residential Storage and VPP Deployment (early November 2026) (first appearance; distinct from last week’s FranklinWH item, which covered Entergy Louisiana and Alliant Energy — different utilities and programs, same vendor — and from the October 26 KJZZ Arizona reserve-margin report and the November 4 APS IRP piece carried in prior digests)
– Troutman Energy Report — FERC Again Rejects Without Prejudice TransAlta’s Cost Recovery Proposal for Compliance with DOE 202(c) Order (November 11, 2026) (first appearance; the 2026 Section 202(c) order count and the J.H. Campbell vacatur are carried benchmarks cited in one line)

November 13, 2026 Entry
– New Jersey Monitor — NJ Regulators Approve Virtual Power Plant, Battery Storage Programs (November 12, 2026) (movement in its strongest form — the BPU’s VPP proposal has appeared five times in this series since July 27 as the “undecided” entry in the enrollment-channel comparison; this is the decision, and it resolves that entry as dual-channel)
– Uplight via GlobeNewswire — Uplight Continues to Scale Demand Flexibility Portfolio, Delivering Reliable Capacity and Empowering Utility Partners to Meet Record Summer Demand and Keep Bills Affordable (November 11, 2026) (first appearance; reported as one cluster with the Oracle and Octopus items because the analytic point is the cluster, but three distinct vendors and datasets)
– Hitachi Energy via PR Newswire — Hitachi Energy Recognized as a Leader in the 2026 IDC MarketScape for Worldwide Advanced Distribution Management Systems (November 11, 2026) (first appearance, citing IDC MarketScape doc #US54133826, October 2026; the GE Vernova GridOS and EnergyHub items carried in prior digests are different vendors and products)
– Stoel Rives — Energy Regulatory Update, November 11, 2026 (first appearance; summarizing CPUC November 19 voting-meeting Item 7 in R.22-11-013 and Resolution E-5455. The CPUC’s final 2026 Avoided Cost Calculator in Resolution E-5476 and the ~$608M DR bridge funding in R.25-09-004 are carried benchmarks from prior digests, cited in one line each)

In-Window Targeted Gap-Fills
– NERC — Project 2024-02 Planning Energy Assurance, “Long-Term Energy Reliability Assessments,” Standard Initial Draft, Formal Posting 1 (October 2025) (first appearance; resolves the highest-value tracking item in this series, open for three runs — the energy assurance standard’s project number, title, applicability, requirements R1–R8, and the fact that Demand Response appears only at R2.2 as a methodology input while DERs, VPPs and load flexibility do not appear at all. Initial draft stakeholder language, not an adopted standard)
– NERC — Standards Committee Agenda Package, October 21, 2026 (first appearance; supplies Project 2024-02’s status in the 2027–2029 Reliability Standards Development Plan — among planned completions, goal “completed by October 2026,” ~12 SMEs and ~2,496 work hours — and Project 2026-02’s December 2026 drafting goal and December 31, 2026 FERC filing deadline)
– Troutman Energy Report — NERC Seeks FERC Approval of New Registration Categories Related to Computational Loads (November 2026) (first appearance; closes the tracked “NERC early-November FERC filing” item — the petition went in October 21, 2026 in Docket RR27-1-000, not the November 4 date NERC’s own Legal Update anticipated. Rules of Procedure revisions only; standards and implementation plans to follow)


Method: All five weekday entries — Monday November 9 through Friday November 13 — are present in the research log; no days are missing. The window is strict Monday–Friday and no item from a prior week is reported as new. Items are reported on the day they entered the log with their original publication dates shown; several carry pre-week publication dates from inside the daily entries’ trailing-window reach (the Octopus/MarketScale piece of November 6 updated November 8, the PA Environment Daily clarification-order report of November 5, the Foley Hoag BAAJA analysis of November 6, the MGRID PJM-winter and EDP/Meta pieces of November 7, the MGRID PG&E and New York storage pieces of November 8, and the Energy-Storage.news Arizona piece of early November). Cross-week deduplication, which is where this week’s work was: the New Jersey BPU’s VPP program has appeared FIVE times in this series since July 27 as a proposal and as the “undecided” entry in the four-way enrollment-channel comparison, so the November 12 unanimous approval is reported as the DECISION and as movement, not as a new program — and it resolves that comparison entry as dual-channel (utility-direct or via a DER manager), which is the analytic payoff rather than the 547.1 MW. The Arizona item is distinct from last week’s FranklinWH/Entergy Louisiana/Alliant piece (different utilities, different programs, same vendor and same battery model) and from both the October 26 KJZZ Arizona reserve-margin report and the November 4 APS IRP filing carried in prior digests; EnergyHub, FranklinWH and Palmetto are all carried vendors and are cited as such. The PA PUC item is the distinct November 5 clarification order, not the October 1 tentative order reported in the late-September/early-October digest. PJM’s winter outlook is distinct from MISO’s Winter PRA carried last week — different operator, different proceeding, different source — and the two sets of winter figures are reconciled in Benchmark Hygiene rather than mixed. Sunrun’s Puerto Rico fleet is a first appearance in the tracked record and is explicitly separated from the September–October California dispatch figures (580 MW, 601 MW, 745 MW), which are a different fleet and not comparable; Base Power’s Central Georgia EMC co-op deal from the October 23 digest is cited as carried context rather than re-reported. The quiet-window reach-back shape flagged on the November 6 run caught one item: Toronto Hydro’s Local Demand Response procurement entered the November 10 daily entry under the “most recent relevant piece” rule from a page refreshed November 9, but the auction is from October 2025 and the publication from February 2026, so it is carried as a dated locational benchmark in Benchmark Hygiene and is NOT reported as news — checked by name against prior digests rather than by URL, which is what the daily log’s window-local dedupe cannot do. Within-log deduplication: the BAAJA coverage is consolidated from two sources across two daily entries into one item (deduplicated) — the November 10 entry’s permitting-letter piece and the November 11 entry’s Foley Hoag analysis of the bill’s separate DER/VPP wholesale-market-access provisions, two halves of one bill; and the Oracle, Uplight and Octopus releases are reported as one cluster, three distinct vendors and datasets, because the analytic point is that three depth disclosures landed in one week and none is independently verified. Zero source URLs overlap last week’s Sources section. Carried benchmarks cited in one line each rather than re-reported: the $325/MW-day PJM clear and the $555/MW-day backstop cap, the 6,831-MW shortfall, PJM’s recovered qualifying-offset criteria, the October 23 MISO LMR penalty order effective June 1, the CPUC’s ~$608M DR bridge funding and Societal-Cost-Test exemption, R.21-06-017’s orchestration framework and Load Flexibility Scenario, SB 913’s June 2028 deadline and device-level telemetry precondition, NERC’s October 26 Board adoption of the computational-load standards, the ~8% Missouri winter accreditation, Google–Voltus >1 GW to 2040, SRP’s 60% pre-payment, FirstEnergy’s ~$250M/GW multiplier, Brattle’s 4% benchmark, Ember’s ~3% shiftable share, CAISO’s four FILI instrumentation requirements, OpenAI’s unaccreditable 1,000 MW, and the ~$66/kW-year and ~$20.70/MWh avoided-cost anchors. The DERMS and ADMS searches in all five daily entries again returned mostly evergreen explainers (Uplight “What Is DERMS,” Oracle, GE Vernova “ADMS vs. DERMS”), DistribuTECH 2026 session listings and market-sizing reports — including the November 5 GlobeNewswire “DERMS Market to 2036” release, excluded as marketing — making this the SEVENTH consecutive week with no standalone DERMS vendor or deployment item; the Hitachi Energy/IDC MarketScape recognition is the first standalone ADMS item in that run and is used for one capability claim rather than as a deployment signal, with the DERMS section otherwise built on the Arizona grid-edge hardware layer and the Carnegie Mellon visibility diagnosis. Targeted search was used only to verify or fill clear gaps, and it resolved three items, one of them the most consequential gap-fill in several months. First and most important: THE ENERGY ASSURANCE STANDARD, the highest-value tracking item in this series and unresolved after three runs, is NERC Project 2024-02 “Planning Energy Assurance,” producing a standard titled “Long-Term Energy Reliability Assessments” (number still TBD), applicable to Planning Coordinators, Resource Planners and Transmission Planners — and in its October 2025 initial draft Demand Response appears exactly once, at R2.2, as one element of a methodology a Planning Coordinator must document, alongside R2.3.5 on energy-storage resource constraints, while DERs, VPPs and load flexibility appear nowhere. R3 defines unacceptable risk in terms of frequency and duration of unserved-load events, maximum unserved load and total unserved energy; R6–R8 require Corrective Action Plans that may consist of a planner’s own mitigating actions or recommendations to third parties including measures needing regulatory approval, which is the provision that could make a CAP a demand-side procurement. The NERC Standards Committee’s October 21, 2026 agenda package places the project among planned completions with a goal of “completed by October 2026,” about 12 subject-matter experts and ~2,496 work hours, with a further comment period and the final ballot still TBD — so the open channel is that comment period, the final ballot, and each Planning Coordinator’s R2 methodology. The caveat is material and is stated in the body and the flag: this is initial-draft stakeholder language, not an adopted standard, and the posted draft still carries “[Standard Name and Number TBD].” Second: NERC’s FERC filing of the computational-load registry criteria, tracked since September, WAS filed — on October 21, 2026 in Docket RR27-1-000, not the November 4 date NERC’s own Legal Update anticipated — seeking expedited approval of Rules of Procedure revisions creating the Computational Load Owner and Computational Load Operator entity types at a threshold of at least 50 MW supplied at 100 kV or greater, estimated to capture about 68,288 MW or roughly 86% of computational load projected in service by end-2027, with the proposed Reliability Standards and implementation plans to be filed later. This confirms the sequencing this series predicted: the registry binds first and the standards follow. Third: the EPRI direct check found no publication dated inside the window — the EIGHTEENTH consecutive week — so the EPRI Spotlight again applies the carried FlexMosaic five-class framework, this time against the Project 2024-02 finding, because FlexMosaic’s four class-determining parameters (notification time, response duration and frequency, and depth and speed of response) are precisely the “capabilities and operational limitations” R2.2 requires a methodology to address and declines to define. What targeted search did not resolve this run: Dominion’s November 15 VPP tariff filing was not yet locatable (the deadline falls two days after the window closes, and whether the 2026 IRP preferred plan carries the VPP as a resource remains open); CAISO’s November 16 EL26-71 compliance filing and the transmission-owner tariff language are likewise due after the window; the ER26-3380 amended tariff sheets were not read directly this run, so whether the prior-delivery-year DR exclusion and the three-year registration-history minimum survive into the amended tariff, and what came of the November 4 offset-evidence decisions, remain open — the PJM docket filings are retrievable and should be read directly next run rather than through coverage; PJM’s deployed demand-response megawatts for the 2026 shoulder-season events remain unpublished; no coverage of the November 4 CEC interoperability workshop output was found; and the CPUC Societal-Cost-Test implementation vehicle and the R.21-06-017 joint Tier 3 advice letter were not located. The Pennsylvania Bulletin publication date that starts the comment clock in Docket M-2026-3064961 is now unverified after five runs and the November 5 clarification order has reset that clock — pacodeandbulletin.gov refuses programmatic retrieval and no secondary source states it, so this needs a manual check or a call to the PUC Secretary’s Bureau and should not absorb further search budget. The November 17 PA PUC technical conference on large-computational-load cost allocation is four days past this window’s close and is the most actionable near-term date on the board.