All five weekday entries (Monday August 10 through Friday August 14) are present in the research log. Three items required consolidation: the Bank of America 230-GW adequacy analysis appeared in both the August 10 and August 13 entries from the same source and is deduplicated as standing context; Xcel Minnesota’s Capacity*Connect utility-owned VPP appeared in both the August 10 and August 13 entries from two different outlets and is consolidated into a single item; and the “Why DERMS Is Earning a Place Alongside SCADA” editorial appeared in both the August 10 and August 13 entries from the same URL and is deduplicated. Entergy Louisiana’s demand-response portfolio is consolidated from two angles — the SEPA/NCCETC budget detail (August 11 entry) and the Louisiana PSC approval itself (August 14 entry). The August 10 and August 13 entries carry the daily task’s own note that no qualifying articles published in the trailing 48 hours, so several items reach back to April–July source dates; they are reported on the day they entered the log, with original publication dates shown. The week’s defining development is new and unambiguous: PJM disclosed that roughly 3,800 MW of Northern Virginia data-center load tripped offline in a single normally-cleared fault on July 22 — the largest such event in PJM history — and is now weighing ride-through requirements as an interconnection condition, while CAISO posted large-load technical standards on August 12 that specify exactly that.
For six weeks this series has argued that flexibility is becoming a condition of large-load service. This week the argument inverted into something stronger and considerably less comfortable: unmanaged large load is now being characterized by its own grid operator as a reliability liability. PJM told its Operating Committee on August 11 that a normally-cleared fault on a 230-kV line in Dominion’s Northern Virginia zone caused approximately 3,800 MW of data-center load to disconnect — system load fell 3.8% in two waves, from 99,984 MW to 96,205 MW — and that the machines were, in the words of PJM’s own reliability engineering staff, “too sensitive.” Nothing broke. The protection logic worked exactly as designed, which is precisely the problem, because it is the same design replicated across every clustered campus on the system. Two prior events of roughly 1,500 MW each occurred in the same zone in 2024 and 2025; this one was more than double both combined.
The regulatory machinery moved to meet it within the same week. CAISO posted a large-load technical-requirements straw proposal on August 12 specifying ride-through capability, post-fault power recovery, ramp rates, monitoring and telemetry, modeling data, and commissioning — the first RTO to put those obligations in writing as baseline interconnection requirements. FERC’s six show-cause responses come due Sunday, August 17, across dockets EL26-67 through EL26-72. And NERC is under a FERC directive to finalize computational-load registry criteria and initial reliability standards by December 31, 2026 — though the voltage and frequency ride-through specifications themselves slip into 2027, a gap PJM warns must close “before the load comes on the system.”
The synthesis for demand-side planners is a reframing rather than a new number. Every prior week’s case for dispatchable DER, storage, and demand response rested on scarcity: firm capacity cannot be built fast enough, so flexibility fills the gap. This week adds a second, independent argument that does not depend on the supply stack at all. Load that can vanish in milliseconds and return uncontrolled is a stability problem, and orchestrated, telemetered, dispatchable distributed resources are the buffer against exactly that volatility. Grid-edge visibility stopped being a procurement preference this week and started becoming a reliability prerequisite.
🔋 Energy Storage
CAISO’s revised demand-response framework, expected August 19, could pull upwards of 2 GW of behind-the-meter batteries into the wholesale market through what one advocate called “a small accounting change [that] could significantly change the battery market in California.” Today BTM batteries are effectively locked out of CAISO’s resource-adequacy market — exports past the meter earn only retail net-metering or net-billing-tariff credit — but the draft under the Demand and Distributed Energy Market Integration initiative would assign wholesale value to BTM aggregations that reduce load within any of the grid’s 20-plus sub-load-aggregation points. This is the valuation-and-access layer that determines whether a DERMS or VPP business case can monetize the roughly $66/kW-year capacity value this series tracks, rather than settling for retail-rate offsets. Two constraints keep it honest: the framework retains DR’s definition as load curtailment, and aggregators seeking genuine net export must still enter the generation interconnection queue — so the “accounting change” unlocks the load-reduction envelope, not unrestricted export. A parallel CPUC demand-response rulemaking, opened by a February scoping order covering valuation methodologies, RA valuation, cost-effectiveness, and CAISO market integration, plus bridge-year funding to sustain existing programs through 2028–2029, could lift participation further — but decisions may not land until Q4 2026 at the earliest and as late as February 2028, so the timeline for capturing this value is measured in years, not quarters. The sub-LAP construct is the procurement signal: granular locational telemetry and aggregation capability at sub-LAP resolution are becoming table stakes for wholesale DER value in California. (Source: Utility Dive, August 5 — August 14 entry)
Note the calendar collision worth planning around: August 19 carries both the CAISO revised DR framework and the CAISO large-loads technical-standards workshop. These are separate initiatives moving on the same date — one opening wholesale access for distributed storage, the other imposing ride-through and telemetry obligations on large loads. Read together they describe CAISO’s whole theory of the grid edge in a single day.
⚡ Virtual Power Plants (VPP) & Demand Flexibility
Xcel Energy’s Minnesota Capacity*Connect is the first utility-owned-and-operated VPP in the United States, approved by the Minnesota PUC to spend up to $430 million deploying roughly 200 MW of customer-sited batteries in 1–3 MW increments through 2028, with a $50 million contribution from Google tied to its in-state data center (deduplicated — appeared in the August 10 and August 13 entries from two outlets). The financing design is what makes it IRP-relevant rather than merely novel: MISO energy and capacity revenues repay nearly the entire deployment cost, leaving a residential bill impact of only $0.67–$1.50 per year through 2030. And the commission attached the mandate this series has repeatedly flagged as the missing piece of every DER business case — Xcel must establish concrete estimates of avoided and deferred distribution value by its November 2027 integrated grid plan, with quarterly progress reports. Utilities have historically struggled to put a defensible number on what a DER is worth for deferring a distribution upgrade; Minnesota has now ordered one produced on a schedule. The contested part is the ownership model itself, and it is the live docket fight worth tracking. MnSEIA, SEIA, and CCSA objected to the absence of a parallel third-party program, arguing the same capacity could be delivered more cheaply by orchestrating DER that ratepayers already own, and that a rate-based, utility-owned battery fleet forecloses the competitive market FERC Order 2222 was meant to open. That is the make-versus-buy question at the heart of every VPP business case, now being litigated with real money: whether the ~$66/kW-year capacity value is best captured through utility-owned assets (rate base, guaranteed return, direct dispatch control) or third-party aggregation (lower capital exposure, competitive pricing, customer participation). Because the program emerged from Minnesota’s integrated distribution planning rather than a standalone VPP mandate, it also blurs the bulk-system and distribution-value boundary in a way commissions elsewhere will cite. The Google co-funding structure is separately worth noting as a template: a hyperscaler paying for distributed capacity that offsets the very load its data center adds. (Sources: Canary Media, April 8; Utility Dive — August 10 and August 13 entries; deduplicated)
The Louisiana PSC approved five Entergy Louisiana demand-response programs targeting 155 MW by 2030 — roughly the capacity of a small gas plant — under an $81 million 2026–2030 budget, while denying the proposed performance-incentive mechanism (consolidated from two entries). The portfolio is a practical template for a diversified DSM value stack: three residential offerings (smart thermostat, EV charging, battery storage), an agricultural irrigation load-control program, and a commercial-and-industrial curtailment program, plus a non-residential aggregated MISO-emergency capacity component. For DERMS procurement, orchestrating thermostats, EV chargers, batteries, irrigation pumps, and C&I loads inside one program is exactly the cross-DER coordination problem a modern platform must solve — five device classes, three customer segments, two dispatch logics. The advocacy detail is the one to carry into IRP work: the Alliance for Affordable Energy pressed for enrollment to be prioritized in load pockets — localized areas where supply-demand imbalance threatens reliability — so the same megawatts of curtailment deliver maximum deferral value and outage protection. That is the demand-side analogue of non-wires-alternative geographic targeting, and it is the difference between a DR program that earns system-average capacity value and one that earns locational T&D deferral value on top. The 155 MW target is a useful mid-size IOU benchmark against this series’ illustrative 70 MW portfolio deferring roughly $150M in CapEx. Note also what the commission declined: the performance-incentive mechanism was denied, which is a caution for any utility building a DR business case that assumes shareholder incentive treatment. (Sources: Alliance for Affordable Energy, March 18 — August 14 entry; SEPA/NCCETC, July 23 — August 11 entry; consolidated)
Puget Sound Energy issued its 2026 Demand Response RFP holding about 129 MW of DR today and needing roughly 500 MW of year-round DR capacity by 2030 to satisfy a 10% peak-reduction planning requirement. This is the cleanest biddable procurement in the log this week and it puts three useful numbers behind a DR value stack at once. The near-fourfold gap between 129 MW and ~500 MW is the acquisition runway. The year-round availability condition is the design signal that matters most for platform requirements — a DR product that must perform in winter and summer, not on a handful of summer peak days, drives materially different dispatch, telemetry, and M&V architecture than a classic summer-peak program, and bidders should price that accordingly. And the explicit tie to a 10% peak-reduction standard is the precedent worth citing: it converts DR from an optional IRP sensitivity into a planning obligation with a defined MW target, the same statutory-hook dynamic seen in Virginia’s H.B. 429 VPP-modeling requirement. An open, competitive solicitation for hundreds of megawatts of DR as firm capacity is also a live avoided-cost comparable — a utility buying DR as a capacity resource rather than emergency load relief validates the deferral-value framing directly. (Source: Puget Sound Energy 2026 Demand Response RFP — August 12 entry)
New York enters summer 2026 with more than 1,566 MW enrolled in demand-response programs against a forecast peak of 31,578 MW — roughly 5% of system peak standing in for firm capacity. The framing matters as much as the ratio: the New York DPS counted enrolled DR as a resource in its summer supply assessment, not as an emergency afterthought, which is what a mature statewide portfolio looks like from a planning perspective. The forecast peak also came in below last year’s actual 31,857 MW, evidence that demand-side measures and efficiency are blunting load growth in a large ISO territory even amid data-center pressure elsewhere. For planners benchmarking their own targets, ~5% of peak is a directional reference point. And the operational implication runs straight to the DERMS layer: enrolling and reliably calling on 1,566 MW across many thousands of participants requires precisely the telemetry, forecasting, and settlement infrastructure a DERMS RFP must specify — the scale figure and the platform requirement are the same fact stated two ways. (Source: New York Department of Public Service, Summer 2026 electricity supply assessment — August 12 entry)
New Jersey’s VPP straw proposal reaches its comment deadline Monday, August 17 under Docket QO26030099 (previously covered — re-entered the log August 11). The program spans all four state electric utilities (PSE&G, JCP&L, Atlantic City Electric, Rockland Electric), enrolls home batteries, smart thermostats, and EV chargers, and compensates through bill credits, discounts, or direct payments under Governor Sherrill’s January 20 Executive Order No. 2 statewide energy emergency. BPU President Ben Hertz-Shargel’s framing is the avoided-cost thesis in plain regulatory language — tapping “technology that’s already sitting in driveways and living rooms … and turn[ing] them into grid assets that reduce utilities’ cost to serve customers.” The multi-utility design is the part with export value: a common program across four IOUs forces consistent enrollment, telemetry, and settlement rules, which is the standardization problem most single-utility programs never have to solve. Note again the August 17 collision with the FERC show-cause responses. (Source: NJ BPU, July 27 — August 11 entry; previously covered, deadline now imminent)
🔌 DERMS & Grid Integration Technology
DERMS has crossed the SCADA threshold: in 2026 utilities are moving DERMS out of the innovation budget and into 3-, 5-, and 10-year capital plans alongside ADMS expansions, substation automation, and AMI (deduplicated — appeared in the August 10 and August 13 entries from the same source). That reclassification is the procurement fact of the year on the technology lane, because it changes what an RFP must contain. Innovation-budget pilots buy capability; capital-plan infrastructure buys SCADA-grade reliability, defined response times, and failover and redundancy terms — and it must integrate with legacy SCADA, DMS, GIS, and OMS rather than sit in a silo. Itron’s Nick Tumilowicz frames the integration principle sharply and it belongs in any DERMS/ADMS specification: a DERMS that dispatches DER without connecting to the broader distribution operating picture “simply replicates the coordination problem it was meant to solve.” DERMS extends ADMS; it does not replace it. The operational drivers are the ones this series tracks — roughly 78 GW of coal retirements over two decades removing the upstream flexibility distribution operators relied on, ~200 GW of projected U.S. load growth by 2030, and updated interconnection standards that convert rooftop inverters, home batteries, and EV chargers from passive loads into dispatchable grid assets. The underweighted element is AMI as a sensing-and-control layer: smart meters across tens of millions of endpoints give DERMS voltage and current visibility at grid-edge resolution that substation-level SCADA structurally cannot match, which is what turns demand flexibility into a feeder-management capability rather than a system-level blunt instrument. The piece is candid about friction — many utilities run legacy SCADA/DMS stacks needing extensive upgrade before an advanced DERMS deploys, and multi-vendor interoperability and communications modernization inflate both cost and timeline, which is exactly why interoperability standards (IEEE 2030.5, CIM) and phased-deployment roadmaps have to be specified up front rather than negotiated later. The closing caution is organizational rather than technical: DERMS delivers full value only when grid-operations and customer-program teams share objectives. Proof points cited: SMUD, Xcel Colorado’s Renewable Battery Connect (15 MW-plus in under six months), and Hawaiian Electric. (Source: T&D World / Itron, June 9 — August 10 and August 13 entries; deduplicated)
🏗️ Data Centers & Large Load Growth
PJM is weighing new interconnection reliability requirements — including ride-through standards — for “computational loads” after approximately 3,800 MW of data-center load tripped offline in Dominion’s Northern Virginia zone on July 22, the largest such event in PJM history. A normally-cleared fault on a 230-kV line produced a large generation-load imbalance and swings in voltage and frequency; PJM’s total load fell about 3.8%, from 99,984 MW to 96,205 MW, in two waves, and the operator restored its Area Control Error within nine minutes against NERC’s 30-minute standard. The system held. What did not hold was the load, and PJM Operating Committee Chair Emanuel Bernabeu’s assessment leaves no ambiguity: “This was a normally cleared fault. They should not disconnect from the grid.” Two prior sudden load-transfer events of roughly 1,500 MW each occurred in the same zone in 2024 and 2025, so the trend line is steep — this event exceeded both combined. For this series’ thesis the significance is a reversal of polarity. Every prior week framed large load as a demand problem that flexible resources help serve; this frames unmanaged large load as a stability problem that flexible, curtailable, orchestrated resources help absorb. Those are different arguments with different audiences, and the second one lands with reliability engineers rather than rate analysts. The regulatory follow-through is already scheduled: under FERC directive, NERC must finalize registry criteria and initial reliability standards for computational loads by December 31, 2026, with a follow-on standards plan due in March — but voltage and frequency ride-through specifications slip into 2027, leaving a window PJM says must close “before the load comes on the system.” For DERMS and ADMS procurement and for IRP work, the operative conclusion is that grid-edge visibility, telemetry, and enforceable performance requirements — for loads and for DER — are becoming reliability prerequisites rather than optional features. (Source: Utility Dive, August 12 — August 14 entry)
PG&E’s data-center pipeline doubled to 12.7 GW in a single quarter, but only 490 MW — under 4% — carries a signed interconnection agreement, and the utility’s own forecast expects to serve just 1.8 GW by 2030. The quarter’s jump came almost entirely from 26 early-stage projects totaling 8.2 GW that had merely paid an initial application fee, a category holding only 1.7 GW in March; another 3.9 GW sits in final engineering. The analysis reads the disclosure exactly right — “a marketing funnel, not a load forecast” — and warns that planners sizing equipment against 12.7 GW will over-buy. Set this beside Exelon’s 40% single-quarter correction from last week’s digest and the pattern is no longer anecdotal: applying any financial-commitment filter to a large-load pipeline collapses it by an order of magnitude. The historical rhyme is instructive and worth citing in a proceeding — one project filing at multiple interconnection points simultaneously echoes the early-2010s renewable interconnection boom that CAISO ultimately saw largely canceled. CEO Patti Poppe’s stated preference for facilities under 1 GW that fit existing substation headroom, because they “move faster” through the queue, is itself a locational-flexibility signal favoring right-sized loads and the DER hosting capacity a utility can actually credit. The disciplinary takeaway for a DSM filing: anchor avoided-cost deferral math to executed-agreement load and the utility’s own served-load forecast, never to headline pipeline figures a commission can readily discount. (Source: Mgrid.org, August 10 — August 11 entry)
CenterPoint Energy raised its 10-year capital plan by $1.2 billion to $66.7 billion and disclosed roughly 14 GW of ERCOT “Batch Zero” eligible large-load submissions — a volume that, if realized, would lift Houston Electric’s peak demand by more than 65% — while pegging demand-charge cash flow at approximately $6 million per GW per month. That per-GW figure is the most useful number in the entry, and it is a double-edged one. It quantifies precisely how lucrative serving firm large load is to the wires utility, which is exactly why non-wires alternatives and DER hosting must be argued head-to-head against the wires build rather than assumed into the plan — the incentive structure does not favor them on its own. Where last week’s FirstEnergy disclosure showed the fossil-generation-build answer to data-center load, CenterPoint is the transmission-and-distribution counterpart: the same demand surge that pulls gas plants into rate base pulls billions in poles-and-wires CapEx behind it. The company claims its Texas projects could yield more than $5 billion in ten-year customer savings, a figure any commission will test against who actually bears the interconnection and capacity costs. And the 14 GW figure carries the same caveat as PG&E’s 12.7 GW — these are Batch Zero submissions, screened through a process that Governor Abbott’s audit has now paused entirely. (Source: Investing.com / CenterPoint Q2 2026 earnings materials, July 28 — August 13 entry)
FirstEnergy’s contracted data-center demand rose roughly 50% in a single quarter — up 2.1 GW to 6.4 GW — and Monongahela Power and Potomac Edison are seeking West Virginia PSC approval for the $2.7 billion Maidsville Energy Center: a 1,200-MW gas plant (~$2.5B, online end-2031) plus 70 MW of solar ($182M). The financing detail is the flashpoint: Mon Power proposes to recover part of the build through a surcharge on its existing customers, even though the plant is oriented primarily around a single data-center customer — the exact cost-allocation exposure large-load tariff dockets in Pennsylvania, Oregon, and Texas are trying to fence off. Note the distinction from the pipeline items above: this is contracted demand, not queue volume, which makes it the more credible number and the harder one to challenge. That cuts both ways for a demand-side filing. It strengthens the “need” case, and it sharpens the counterfactual: a 1,200-MW gas plant with a five-plus-year lead time funded partly by captive ratepayers is the slow, expensive, emissions-heavy path that dispatchable DER, storage, and DR can defer or right-size on a materially shorter clock. The IRP discipline is unchanged — anchor firm-generation need to contracted, cost-caused load, insist that large loads bear their own interconnection and capacity costs, and force DER and non-wires alternatives into the same avoided-cost comparison as the gas build. (Source: Utility Dive, August 2026 — August 12 entry)
Adequacy context (previously covered, deduplicated): Bank of America’s analysis re-entered the log twice this week, on August 10 and August 13, from the same source — more than 230 GW of new capacity needed over five years against roughly 93 GW of accredited utility additions, a gap exceeding 100 GW; data centers alone adding ~125 GW of load at a 4.1% CAGR through 2030; 7.5 GW of data-center on-site generation under construction and 60-plus GW in pre-construction; large gas turbines committed through 2030; and coal retirements delayed or canceled across Maryland, Wisconsin, Indiana, Utah, Kansas, Nebraska, and Mississippi. It was covered in the July 24 digest and remains the standing scarcity citation. The line that has aged best is the deliverability framing: “the market is no longer constrained by demand … it is constrained by where power can actually be delivered.” (Source: Utility Dive, July 17 — August 10 and August 13 entries; deduplicated, previously covered)
📋 Regulatory & Policy
FERC’s six large-load show-cause orders reach their response deadline this Sunday, August 17 — dockets EL26-67 (PJM), -68 (SPP), -69 (NYISO), -70 (MISO), -71 (CAISO), and -72 (ISO-NE). Each RTO and its transmission owners must either show cause why its tariff remains just and reasonable without large-load-specific provisions, or state what tariff revisions would remedy the Commission’s concerns, across five reform areas: efficient transmission study processes; cost transparency so new large loads do not shift costs onto existing customers; clear co-location and behind-the-meter-generation rules; new transmission services for flexible large loads; and processes for studying generation dedicated to co-located loads. The fourth area is the one that matters most to demand-side practice, and its language is doing real work — by formally recognizing that large loads with flexible operating profiles “can provide value to the grid,” FERC is edging toward treating demand flexibility as a compensable grid service under the same avoided-cost logic that underpins DR and non-wires alternatives. The cost-transparency reform embeds the ratepayer-protection framing DER business cases already rely on. FERC’s proposed large-load definition — peak load in excess of 50 MW interconnecting at above 69 kV, single site, non-co-located — sets the threshold governing which loads face these obligations, and it is worth comparing against PJM’s own 50 MW at a single point of interconnection (or multiple points within one mile) and NERC’s 20 MW at 60 kV computational-load criterion. Three different thresholds, three different purposes, one customer population. Interveners have until September 16 to comment on the responses; that record will shape large-load cost allocation and DER hosting treatment for years. The August 3 abeyance deadline has passed — requests were limited to 90 days, with FERC on record that it will scrutinize them heavily and disfavor extensions — but which operators actually filed remains unconfirmed in public reporting and stays an open verification item. (Sources: Husch Blackwell, June 25 — August 14 entry; Day Pitney, McGuireWoods — external verification)
CAISO posted its large-load technical-requirements straw proposal on August 12, and its content is the direct regulatory answer to PJM’s 3,800-MW trip. Under the expedited Large Loads Initiative — prompted by the June 2026 FERC show-cause order in EL26-71 — the proposal sets baseline technical requirements for reliable interconnection and operation of large loads, including data centers, on the ISO-controlled grid: ride-through capability, post-fault power recovery, ramp rates, monitoring and telemetry, modeling data, and commissioning. That list is worth reading twice, because it is functionally the same requirement set utilities have been imposing on generators for decades, now applied to load. A stakeholder workshop follows August 19 with comments due September 2, and CAISO’s compliance filing is due at FERC November 16. For DERMS and ADMS planning the monitoring-and-telemetry and modeling-data requirements are the operative ones: they establish that large-load visibility is an interconnection deliverable, which is the same data plumbing a DERMS needs and a strong argument for building it once. This is meaningful movement on a watch item this series has carried for three weeks. (Source: CAISO Large Loads Initiative — external verification)
DOE’s Section 202(c) emergency authority is now a documented systemic pattern rather than a series of isolated events (consolidated from two entries). POWER Magazine’s continuously updated ledger counts 43 Section 202(c) orders and extensions signed since May 2025 — an authority historically invoked a handful of times per decade now firing on a near-weekly cadence — and tallies that retirement-deferral orders alone have stalled the closure of at least 4.4 GW of coal capacity as of April 2026. The ledger sorts the orders into two families: retirement-deferral orders forcing generators to run past planned shutdown, and operating-limit orders letting plants exceed environmental or permit ceilings during scarcity. The individual case logged this week is Order 202-26-37, signed July 26 and extended as 202-26-37A through August 10, directing SPP to dispatch named generators and call on behind-the-meter backup across a 17-state footprint from northern Texas to Montana during a late-July heat stretch — with environmental runtime limits stepping aside for the order period, and DOE pointing to a roughly 35 GW pool of idle standby generation at industrial sites, hospitals, and data centers. A counting caution belongs with these figures: POWER’s “43” counts orders and extensions since May 2025, while the SPP analysis describes Order 202-26-37 as the thirty-seventh of 2026 under the 202-26-XX numbering. These are different denominators, not a contradiction, and prior entries in this series have used both. Cite the pattern and the cadence; do not cite a single running tally as authoritative. For a DSM business case the argument compresses to one line: when the federal fallback for insufficient firm capacity is holding aging, high-emission fossil units open at ratepayer expense — with cost recovery for fuel, wear, and emissions compliance repeatedly deferred to after-the-fact FERC proceedings — the avoided-cost and ratepayer-protection value of clean, dispatchable demand-side flexibility rises by exactly that much. Owners of the 35 GW standby fleet should note the second-order point: federal orders now reach into private backup generation routinely, which blurs the line between a private asset and a de-facto grid resource and foreshadows formalized, biddable large-load flexibility products. (Sources: POWER Magazine, August 2026 — August 12 entry; Mgrid.org, August 3 — August 11 entry; consolidated)
The SEPA/NCCETC Q2 2026 policy roundup remains the densest source of program comparables, and its detail re-entered the log August 11 (previously covered July 31). On compensation: DTE Electric’s proposed two-year Michigan Residential Battery VPP Pilot at $105/kW-year (capped at $504 per participant); Illinois’ ICC approval of ComEd’s and Ameren’s Scheduled-Dispatch VPP tariffs plus ComEd’s Bring-Your-Own-Device Load Reduction program, with a cost-effectiveness evaluation due end-2026. On the DERMS and data-access layer: Maryland’s PSC ordered a third-party-accessible DER registry, a Data Exchange Work Group, interconnection-tool alignment with forthcoming VPP applications, and progress reporting on utility DERMS timelines — concrete regulatory scaffolding for the orchestration requirements a DERMS RFP must specify. On IRP obligations, Virginia’s statutory swarm remains the forward signal: H.B. 429 requires utilities to model IRP scenarios exceeding maximum energy-savings targets via VPPs using aggregated DR or storage, H.B. 434 directs evaluation of non-wires alternatives, and H.B. 1467 mandates an Appalachian Power VPP pilot of up to 150 MW of DER aggregations by July 2027. Hawaii’s PUC edge-DERMS grid-services docket (2026-0084) rounds out the set. (Source: DSIRE Insight / NCCETC & SEPA, July 23 — August 11 entry; previously covered)
Two verification items resolve with corrections this week. First, NERC’s computational-load timeline: the August milestone is a revised draft of the Computational Load Entity registration criteria incorporating comments from the window that closed May 15 — not, as framed in last week’s digest, the opening of a new comment window. Board approval remains targeted for December 5 and the FERC filing for December 31. The criteria themselves are unchanged and still broader than most operators assume: aggregate connected load capability of 20 MW or greater, at a single point of interconnection to the Bulk Power System at 60 kV or above, hosting 1 MW or more of computational load. Second, MISO’s Zero Injection GIA filing: no August filing is confirmed in public reporting. MISO continues to target FERC acceptance in time for end-2026 implementation, requiring revisions to Tariff Attachment FF and applicable Business Practice Manuals, and utilities have filed objections on reliability and legal grounds. This item has now been open for three consecutive weeks. (Sources: NERC Rules of Procedure posting; Steptoe; RTO Insider — external verification)
🔬 EPRI Research Spotlight
EPRI-anchored research re-entered the log on August 10 after four dry weeks — though the vehicle is the June 26 FlexMosaic and “speed-to-power” coverage previously carried in the August 7 digest rather than genuinely new EPRI output, which makes this the fifth consecutive week without a fresh EPRI publication in the log. A direct check confirms no major EPRI announcement since the FlexMosaic framework and the DCFlex expansion to nine demonstration sites.
The item most worth foregrounding is the quantitative finding that had not been pulled out on its own: a 2026 Duke University Nicholas Institute study finds that shaving just 1–2% off data-center peak demand lowers electricity rates 0.5–2.8% while protecting reliability. That is an unusually clean rate-impact number for a demand-side intervention, expressed in the units a commission actually rules on, and it belongs in any large-load flexibility filing this cycle. The mechanism EPRI supplies is the “headroom” argument — flexible large loads let operators interconnect capacity faster while minimizing rate-raising infrastructure, converting an interconnection bottleneck into a procurable demand-side resource.
The FlexMosaic five-class taxonomy remains the valuation grammar this series expects to see referenced in binding flexibility agreements: Class A (infrequent extreme stress) through Class E (frequency stabilization), keyed to notification time, duration, frequency, depth, and speed of response, with Class D (sudden supply/demand swings) and Class E identified as unlocking the most system value by mitigating local thermal overloads and voltage drops. Read against this week’s PJM disclosure, Class D and E are no longer abstractions — a 3,800-MW uncontrolled disconnection is the Class D scenario stated in the negative, and the FlexMosaic classes are the vocabulary for specifying what a controlled response should look like instead. EPRI has also described its framework as a voluntary standard creating “a shared, credible way to define flexibility from large loads … based on the magnitude, timing, duration, and frequency of their response,” explicitly aimed at shortening interconnection timelines.
Supporting figures carried forward: U.S. data-center demand projected at 66 GW in 2027, up from 31 GW in 2025, with summer peak share rising to 8.5% and EPRI’s estimate of up to 17% of U.S. electricity by 2030. The regulatory tailwind remains FERC’s June 18 order directing system operators to provide transmission for flexible large loads, alongside NERC’s 2026 large-load risk-mitigation guidelines crediting data-center flexibility as ratepayer protection.
Standing watch: the NERC Computational Load Entity revised draft is due this month, and EPRI’s DCFlex demonstration data remains the most credible technical record available to that proceeding. Check EPRI.com directly again next run — six weeks without new research is long enough that the drought is itself worth noting in the next digest.
🚩 Utility-Sector Relevance Flags
⚑ 3,800 MW Disconnected From a Normally-Cleared Fault — Load Is Now a Stability Problem
Topic: Bulk-System Reliability / Large-Load Ride-Through / DER as Buffer
Relevance: PJM’s July 22 Northern Virginia event is the largest data-center load loss in its history, more than double the two ~1,500 MW events of 2024 and 2025 combined, and PJM’s own reliability engineering characterizes the machines as “too sensitive” — they should not have disconnected. System load fell 3.8% in two waves; ACE was restored in nine minutes against a 30-minute standard. Nothing failed, which is what makes clustered campuses running identical protection logic a correlated multi-gigawatt risk. This gives the demand-side case a second, supply-independent argument: orchestrated, telemetered DER buffers exactly this volatility.
Action Signal: Implement — Audit protection-logic and ride-through settings across clustered large loads on your system now, and begin writing ride-through performance and EMT model delivery into large-load interconnection agreements rather than waiting for the 2027 NERC specifications. Utilities with Northern Virginia-style load concentration should treat correlated trip exposure as a named planning scenario this cycle.
⚑ CAISO Wrote the Requirement List — Ride-Through, Telemetry, Modeling, Commissioning
Topic: Large-Load Interconnection Standards / Telemetry / DERMS Data Layer
Relevance: The August 12 straw proposal is the first RTO document to specify baseline large-load technical requirements in writing: ride-through capability, post-fault power recovery, ramp rates, monitoring and telemetry, modeling data, and commissioning. It is functionally the generator interconnection requirement set applied to load. Workshop August 19, comments September 2, FERC compliance filing November 16. Other RTOs facing the same show-cause pressure will draw from it.
Action Signal: Engage — Read the CAISO list as the likely national template and map it against your own large-load interconnection requirements to find the gaps. The monitoring/telemetry and modeling-data provisions overlap substantially with DERMS data requirements — scope them as one integration project, not two, and file comments by September 2 if you operate in CAISO.
⚑ August 17 Is a Triple Deadline
Topic: Regulatory Calendar / Show-Cause Responses / VPP Docket
Relevance: Sunday, August 17 carries show-cause responses from all six RTOs and their transmission owners (EL26-67 through EL26-72) and the NJ BPU VPP straw-proposal comment deadline under Docket QO26030099. Comments on the show-cause responses run to September 16. The flexible-large-load service reform area is where FERC comes closest to treating demand flexibility as a compensable grid service, which makes the September 16 comment record the more important of the two windows for demand-side advocacy.
Action Signal: Engage — Confirm your filing position on both dockets this week. For the show-cause record, the September 16 comment window is where demand-side and DER interests should concentrate resources; the RTO responses themselves are the target, not the original orders.
⚑ Three Large-Load Thresholds, One Customer Population
Topic: Registration / Compliance Scope / Tariff Design
Relevance: FERC’s show-cause definition (>50 MW at >69 kV, single site, non-co-located), PJM’s Interim Resource Adequacy Service definition (≥50 MW at a single POI or multiple POIs within one mile), and NERC’s Computational Load Entity criteria (≥20 MW aggregate at a single POI at ≥60 kV, hosting ≥1 MW of computational load) capture overlapping but materially different populations. NERC’s is the broadest by a wide margin and reaches colocation and enterprise sites, not just hyperscale campuses. The revised NERC draft is due this month; Board approval December 5, FERC filing December 31.
Action Signal: Implement — Run all three screens across your service territory and produce one reconciled customer list. Customers will receive obligations from more than one of these frameworks and will look to their utility to explain the difference; being able to do so is a customer-relationship advantage and a prerequisite for offering flexibility as the alternative to compelled curtailment.
⚑ PG&E: 12.7 GW Pipeline, 490 MW Signed, 1.8 GW Expected — The Filter Is the Finding
Topic: Load Forecasting / IRP Discipline / Capital Allocation
Relevance: Under 4% of PG&E’s pipeline carries an executed interconnection agreement, and the utility’s own 2030 served-load forecast is 14% of the headline. The quarter’s doubling came almost entirely from 26 projects that paid an application fee. Following Exelon’s 40% single-quarter correction, two large utilities in two weeks have shown that any financial-commitment filter collapses a large-load pipeline by roughly an order of magnitude. Meanwhile CenterPoint discloses ~14 GW of Batch Zero submissions into a process Texas has paused entirely.
Action Signal: Implement — Use executed-agreement load and the utility’s own served-load forecast as the denominator in every avoided-cost and deferral calculation this cycle, and state the filter explicitly in filings. Where a proceeding relies on headline pipeline figures, the PG&E and Exelon disclosures are the two most citable rebuttals available.
⚑ The First Utility-Owned VPP Comes With a Distribution-Value Mandate
Topic: VPP Ownership Model / DER Valuation / Cost Recovery Precedent
Relevance: Xcel’s $430M / ~200 MW Capacity*Connect is the first U.S. utility-owned-and-operated VPP, funded almost entirely by MISO market revenues ($0.67–$1.50 per residential customer per year), with $50M from Google. The commission ordered Xcel to quantify avoided and deferred distribution value by its November 2027 integrated grid plan with quarterly reports — a direct attack on the valuation gap that has weakened DER business cases for a decade. Third-party aggregators are contesting the ownership model as foreclosing the Order 2222 competitive market.
Action Signal: Watch — Track the Minnesota docket and the quarterly avoided-distribution-value reports specifically; if Xcel produces a workable methodology, it is exportable to every other jurisdiction and is worth more to the sector than the 200 MW. Utilities weighing a VPP build should treat the make-versus-buy question as genuinely open and document the comparison, because commissions are now being asked to rule on it directly.
⚑ Year-Round DR Is a Different Product — PSE Needs 500 MW of It by 2030
Topic: DR Procurement / Program Design / Platform Requirements
Relevance: PSE holds ~129 MW and must reach ~500 MW of DR available year-round by 2030 under a 10% peak-reduction planning requirement. The year-round condition changes the dispatch, telemetry, and M&V architecture substantially versus a summer-peak program, and the statutory peak-reduction hook converts DR from an IRP sensitivity into a planning obligation with a defined target — the same mechanism as Virginia’s H.B. 429.
Action Signal: Engage — Bidders should price year-round availability explicitly rather than extrapolating summer-peak economics. Utilities without a peak-reduction planning standard should note that PSE’s, Virginia’s, and Illinois’s statutory hooks are the mechanism actually driving DR procurement to scale, and consider advocating for one.
⚑ Load-Pocket Targeting Is the Free Multiplier on a DR Portfolio
Topic: DR Program Design / Locational Value / Non-Wires Alternatives
Relevance: The Louisiana advocacy ask on Entergy’s 155 MW / $81M portfolio — prioritize enrollment inside load pockets where supply-demand imbalance threatens reliability — is the demand-side analogue of non-wires-alternative geographic targeting, and it is the difference between earning system-average capacity value and earning locational T&D deferral value on top of it. It costs nothing but enrollment-targeting discipline. Note also that the commission denied the proposed performance-incentive mechanism.
Action Signal: Implement — Overlay your DR enrollment map on your constrained-feeder and deferred-project map before the next program cycle, and weight acquisition spend accordingly. Do not assume shareholder performance-incentive treatment in a DR business case; Louisiana just declined it.
⚑ 43 Emergency Orders and Counting — But Cite the Pattern, Not the Tally
Topic: 202(c) Cadence / Avoided-Cost Framing / Backup-Generation Fleet
Relevance: POWER’s ledger counts 43 Section 202(c) orders and extensions since May 2025, with retirement-deferral orders stalling at least 4.4 GW of coal closures; the SPP order (202-26-37, July 26, extended through August 10 across 17 states) reaches into a ~35 GW pool of idle behind-the-meter standby generation. Note the counting caution: “43 since May 2025” and “37th of 2026” use different denominators and both appear in this series. Cost recovery for fuel, wear, and emissions is repeatedly deferred to after-the-fact FERC proceedings.
Action Signal: Watch — Use the cadence and the 4.4 GW coal-deferral figure as the ratepayer-protection framing in avoided-cost filings, and avoid quoting a single running total as authoritative. The 35 GW standby fleet is the quantified conversion opportunity: federal orders already reach into it, which argues for programmatic, compensated enrollment before it is commandeered uncompensated.
📌 Sources
August 10, 2026 Entry
– Utility Dive — Bank of America Warns AI Data-Center Growth Will Leave the U.S. More Than 100 GW Short of Firm Capacity Through 2030 (July 17, 2026)
– Utility Dive — Data Centers, Flexibility, Utilities and “Speed to Power”: EPRI FlexMosaic and the Duke Rate-Impact Study (June 26, 2026)
– Canary Media — Xcel Minnesota Wins Approval for the Nation’s First Utility-Owned Virtual Power Plant (April 8, 2026)
– T&D World (Nick Tumilowicz, Itron) — Why DERMS Is Earning a Place Alongside SCADA (June 9, 2026)
August 11, 2026 Entry
– Mgrid.org — DOE Orders SPP to Dispatch Backup Generation Across 17 States, Then Extends It Into August (August 3, 2026)
– Mgrid.org — PG&E’s Data Center Pipeline Doubles to 12.7 GW, but Only 490 MW Has a Signed Agreement (August 10, 2026)
– New Jersey Board of Public Utilities — VPP Straw Proposal and Stakeholder Docket QO26030099 (July 27, 2026)
– DSIRE Insight (NCCETC & SEPA) — VPP and Supporting DER Policy Developments, Q2 2026 (July 23, 2026)
August 12, 2026 Entry
– POWER Magazine — DOE Has Issued More Than 40 Section 202(c) Emergency Orders Since May 2025: An Updated Log (August 2026)
– Utility Dive — FirstEnergy Data-Center Demand, West Virginia Maidsville Gas Plant, Q2 Earnings (August 2026)
– Puget Sound Energy — 2026 Demand Response Request for Proposals
– New York Department of Public Service — Summer 2026 Electricity Supply Assessment
August 13, 2026 Entry
– Investing.com — CenterPoint Q2 2026 Slides: Demand Surge Drives Capital Plan Boost (July 28, 2026)
– Utility Dive — AI Data-Center Growth and Utility Generation Plans (Bank of America capacity gap)
– T&D World — Why DERMS Is Earning a Place Alongside SCADA
– Utility Dive — VPP vs. VPP: Customer-Owned DER Aggregators Challenge Xcel-Owned Batteries in Minnesota
August 14, 2026 Entry
– Utility Dive — PJM Eyes Data Center, Crypto Reliability Requirements After 3.8 GW of Load Trips Offline (August 12, 2026)
– Husch Blackwell — FERC Issues Show Cause Orders to All Six RTOs/ISOs on Large Load Interconnection (June 25, 2026)
– Utility Dive — California DER Wholesale Market Participation Could Grow 2 GW From CAISO Framework (August 5, 2026)
– Alliance for Affordable Energy — PSC Approves Demand Response Programs (March 18, 2026)
External Verification (targeted searches, not from the daily log)
– California ISO — Large Loads Initiative: Technical Requirements Straw Proposal Posted (August 12, 2026)
– California ISO — Large Loads Stakeholder Initiative Page
– California ISO — July 20, 2026 Informational Report on Large Loads and Co-Located Loads, EL26-71
– RTO Insider — Va. Data Center Ride-Through Performance Faulted for July Voltage Disturbance
– PJM Inside Lines — PJM, Dominion Review Large Load Transfer Event
– American Public Power Association — PJM, Dominion Review Large Load Transfer Event
– Day Pitney — FERC Issues Show Cause Orders to Six RTOs/ISOs on Large Load Integration
– McGuireWoods — FERC Issues Section 206 Show Cause Orders Directing All Six RTOs/ISOs to Justify or Reform Large Load Integration Rules (June 2026)
– PJM — Interim Resource Adequacy Service Executive Summary (July 27, 2026)
– PJM — Reliability Backstop Procurement FERC Filing, Docket ER26-3380-000 (July 31, 2026)
– NERC — Computational Load Entity, Summary of Changes (Rules of Procedure posting)
– Steptoe — NERC Releases Proposed Registration Requirements for “Computational Load” Customers
– RTO Insider — Utilities Object to MISO Zero-Injection Gen Plan for Large Loads
– EPRI — DCFlex Initiative Expands to Nine Demonstration Sites Across U.S., Europe
Next digest: week ending August 21, 2026. Carried verification items — the six RTO show-cause responses filed August 17 and what each proposed for flexible-large-load service; which operators filed August 3 abeyance requests (still unconfirmed); CAISO’s August 19 revised DR framework and the same-day large-loads workshop; PJM’s Interim Resource Adequacy Service docket number once posted; MISO Zero Injection GIA filing status (open three weeks); the NERC Computational Load Entity revised draft expected this month; PJM’s ride-through rulemaking path and whether other RTOs follow CAISO’s technical-standards template; ERCOT Batch Zero audit scope and any restart; FERC response to the five-state “but-for” rehearing; and a direct EPRI check after six weeks without new research.