All five weekday entries (Monday July 13 through Friday July 17) are present in the research log. This digest covers them in full. Two long-running watch items resolved this week: New Jersey’s Governor Sherrill signed the data center tariff package July 7 (confirmed in the July 13 entry), and — via direct external verification of the PUCT docket, as flagged in the last two digests — ERCOT’s Batch Zero framework was confirmed approved by the PUCT on June 18, 2026, on schedule; the vote occurred but was never captured in the daily log across five consecutive weeks.

The defining item of the week arrived Thursday: FERC’s July 16 order in Docket RD26-7-000 directing NERC to file mandatory reliability standards for “computational loads” — data centers and crypto facilities — by December 31, 2026, with registry criteria that would, for the first time, make the loads themselves registered entities accountable under Section 215. The evidentiary record behind the order (a July 2024 event that shed ~1,500 MW of data-center load in milliseconds; ERCOT crypto sites losing 17–95% of consumption during normally cleared faults) converges with this week’s IEEE Spectrum analysis of AI workload volatility to reframe large flexible load as a bulk-system reliability obligation rather than a customer-service question. Meanwhile the July heat pattern continued: DOE issued its second summer 202(c) order for PJM on July 14 as the RTO forecast its third-highest peak ever (164,553 MW), just 12 days after PJM’s post-event analysis confirmed the all-time record of 168,158 MW on July 2 — with demand response delivering a verified 6,113 MW of load reduction at peak, roughly $403 million in annual avoided capacity value at the Brattle ~$66/kW-year benchmark. Distributed solar earned equally hard numbers: the Acadia Center calculated $130 million in avoided ISO-NE wholesale costs during the heat dome, with rooftop solar out-contributing the region’s nuclear fleet during peak afternoon hours on July 2. On the policy side, New Jersey’s signed three-bill package added a DER capacity-offset mechanism not visible in pre-signing coverage, Louisiana’s PSC approved a five-program, 155 MW Entergy DR portfolio, and a Utility Dive investigation put a price on the supply-side alternative: DOE’s 202(c) campaign to keep seven fossil plants online costs ratepayers ~$550 million per year — while four of the eleven ordered units aren’t even operating.

🔋 Energy Storage

SEIA reported the strongest first quarter for U.S. energy storage on record — 9.7 GWh installed in Q1 2026, up 32% year-over-year — and identified AI data center demand as the primary growth driver for the first time, displacing renewable integration as the market’s main accelerant. Utility-scale batteries accounted for 7.8 GWh (80% of the total), with C&I at 648 MWh and residential at 515 MWh; Google and Meta have announced procurements of tens of thousands of MWh to support data center operations. The revised forecast of 610 GWh cumulative installations by 2030 implies roughly an order-of-magnitude growth in the installed U.S. battery fleet within four years — an enormous dispatchable pool that, if reachable through DERMS platforms and VPP programs, materially alters the capacity-planning calculus that currently defaults to gas peakers. Notably, the residential segment held at 515 MWh despite the OBBBA’s elimination of the Section 25D residential solar credit effective January 1, suggesting state incentives (e.g., NYSERDA’s VPP-linked storage incentive) and VPP performance payments are sufficient to sustain baseline deployment. Texas, Arizona, and California led utility-scale installs, with Georgia, Iowa, and Mississippi emerging as new growth markets — a geography that mirrors EPRI’s data-center load projections. (Source: Utility Dive / SEIA — July 13 entry)

Tesla deployed 13.5 GWh of energy storage in Q2 2026 — its second-largest quarter ever, up 53% sequentially and 40% year-over-year — meaning a single vendor’s quarterly deployment exceeded the entire U.S. market’s record Q1 total. Cumulative Tesla deployments now exceed 132 GWh since 2016, on pace to beat the company’s 46.7 GWh full-year 2025 record. The strategic overlap with the VPP market is the operative point for this log: the same Powerwall units driving deployment totals feed the VPP-aggregable fleet behind the Tesla/Sunrun/Renew Home 16.8 GW framework (covered in the July 3 digest), compounding the addressable capacity pool with every residential install. Wood Mackenzie’s broader market context — developers planning a record 24 GW of utility-scale storage additions in 2026, a 30% jump over 2025 — positions storage as the fastest-growing resource category and the enabling technology for the shift from emergency-only DR toward continuous DERMS-optimized balancing. Tesla’s July 22 earnings call is expected to detail positioning for state-mandated VPP programs (Illinois 2027, New Jersey). (Source: pv magazine USA — July 15 entry)

☀️ Distributed Solar

The Acadia Center calculated that behind-the-meter solar saved New England customers at least $130 million in wholesale energy costs during the June 28–July 4 heat event — avoiding 28–43% of daily wholesale energy expenditures and, between 2 p.m. and 7 p.m. on July 2, contributing more to the ISO-NE fuel mix than the region’s entire nuclear fleet. Rooftop and other BTM solar supplied up to 25% of total regional demand at peak production hours. Two planning implications stand out for IRP work. First, the avoided-cost value scales with stress: when wholesale prices spike, distributed solar’s per-MWh value rises proportionally, making it a natural hedge against exactly the price volatility driving ratepayer complaints. Second, BTM solar is actively shifting the system peak from mid-afternoon to early evening — creating the precise post-sunset gap that DERMS-dispatched batteries (e.g., Eversource’s ConnectedSolutions+ substation pilots, covered in the July 10 digest) are designed to fill. This is the most granular quantification of distributed solar’s operational value during extreme stress yet to reach this series. (Source: Vermont Public / Maine Public, New England News Collaborative — July 13 entry)

Canary Media’s multi-region post-heat-wave analysis confirms clean energy resources were load-bearing infrastructure — not supplementary contributors — during the record heat dome: Texas solar covered more than 30% of ERCOT’s record 83 GW July peak, New England offshore wind cut oil-fired generation below prior-year heat-wave levels, and SPP leaned on wind for central-U.S. reliability. PJM’s data point is the most consequential for capacity planning: the RTO set its all-time demand record without ultimately needing the emergency resources authorized under the DOE 202(c) orders — demand response, distributed solar, and utility-scale clean generation carried the event without last-resort federal intervention being exercised. The portfolio effect is the citable takeaway: geographic and temporal diversity across wind, solar, and demand-side resources delivered aggregate reliability exceeding what any single technology or region provides in isolation — the value proposition DERMS multi-resource dispatch exists to capture. (Source: Canary Media — July 14 entry)

⚡ Virtual Power Plants (VPP) & Demand Flexibility

PJM’s official post-event analysis confirmed demand response delivered 6,113 MW of load reduction during the all-time record peak of 168,158 MW on July 2, and 5,037 MW on July 3 — substantially above the preliminary “4,000+ MW” estimates in earlier reporting. At the Brattle ~$66/kW-year avoided-cost value, 6,113 MW of demonstrated DR capacity represents roughly $403 million in annual avoided capacity value, delivered under exactly the conditions DR programs are designed for. The operational context sharpens the case: forced generation outages of 18,100–19,400 MW ran 41–51% above the three-year top-peak-day average, meaning the grid absorbed record demand and above-average supply-side failure simultaneously — PJM’s dispatch lead said the RTO used “every single generator available” and called the capacity position an ongoing “area of concern.” The 3.25 GW operating beyond environmental permit limits under 202(c) authority underscores that current margins depend on emergency authorities with real environmental costs, strengthening the argument for demand-side reliability insurance. These are the definitive numbers to cite from the July heat dome; official DR evaluation follows within 60 days. (Source: PJM Inside Lines — July 13 entry)

Foley Hoag’s July survey of the state VPP landscape documents enabling action across at least ten states in 2025–2026, crystallizing the shift from voluntary utility pilots to legislatively mandated capacity resources. New York’s GREAT Act establishes a statewide VPP program requiring IOUs to create dedicated home-battery and EV riders; Illinois’ Clean and Reliable Grid Affordability Act (signed January 2026) pairs a ~3 GW grid-scale storage procurement with mandatory household VPP programs; Massachusetts targets 3.5 GW from new load-management strategies; and Minnesota’s $430M distributed-battery program rounds out a first quarter in which four states quantified demand-side aggregation targets at GW scale. For IRP practitioners, statutory VPP mandates lower the evidentiary burden for booking avoided-cost value — capacity that regulators are required to procure is easier to defend as firm and deferrable than capacity a utility merely proposes. (Source: Foley Hoag LLP — July 17 entry)

Latitude Media’s deep analysis of Xcel’s Capacity*Connect — the first utility-owned VPP in the U.S., and a standing watch item in this series — surfaces the structural tensions that will shape the national utility-owned-vs-third-party VPP debate. This is a movement update on the $430M/200 MW Minnesota program (PUC approval and SEIA/MnSEIA opposition covered in prior digests); new this week are the specifics of the criticism: the sole-source Sparkfund partnership with no competitive procurement, cost estimates exceeding comparable third-party VPP programs, and an interconnection-queue bypass for utility-owned batteries that third-party DERs don’t get — what critics call a regulatory double standard, and what CCSA characterizes as “the utility effectively extending its monopoly.” The PUC-required independent evaluation (due November 2027) sets up a natural experiment: by 2027–2028 the industry will have operational data from utility-owned (Xcel), tariff-based third-party (ComEd SDVPP), and hyperscaler-funded (Google/Voltus, Tesla/Sunrun/Renew Home) models — an evidence base that will likely determine which ownership architecture regulators favor nationally. (Source: Latitude Media — July 15 entry; movement update on prior digest coverage)

The Tesla/Sunrun/Renew Home 16.8 GW VPP framework — first covered in the July 3 digest — reappeared in this week’s log with deployment details worth recording: 300 MW is ready for immediate deployment in Virginia’s Data Center Alley (growing to 500+ MW by 2030), California accounts for 4.66 GW (27.7%) of the pool, and participating households were paid $67 million in 2025 for device flexibility. This is deduplicated context, not a new item — but the Virginia-ready capacity and the explicit positioning of VPP aggregations as large-load interconnection offsets (“deployed within months… requiring no additional hardware, software, interconnection, water, or land”) tie the framework directly to the FERC show-cause and NJ capacity-offset developments elsewhere in this digest. (Source: pv magazine USA — July 15 entry; deduplicated against July 3 digest coverage)

Peninsula Clean Energy and Silicon Valley Clean Energy jointly launched a two-part demand flexibility RFP — each CCA targeting 5 MW (10 MW combined) of existing flexible-load capacity by 2026 and 25 MW each (up to 50 MW combined) of new demand flexibility by 2030 via DR, load-modifying programs, and dynamic pricing. The signal is that CCAs — not just vertically integrated IOUs — are now active procurers of demand flexibility as a planning resource with defined MW targets, expanding the addressable market for aggregation platforms and third-party DR providers. The joint-procurement structure (two CCAs’ load pooled into one RFP for a larger, more bankable enrollment base) is a replicable template for municipal aggregators as California leans on demand-side resources for peak and hosting-capacity management. (Source: CalCCA — July 17 entry)

🔌 DERMS & Grid Integration Technology

NERC’s inverter-based resource ride-through standards PRC-029-1 and PRC-030-1 take effect October 1, 2026 for BES facilities (January 1, 2027 for non-BES) — prohibiting momentary cessation during grid disturbances and requiring inverters to remain connected and inject reactive current through voltage and frequency excursions, with penalties up to $1.54 million per day per violation. Category 1 covers BES IBR facilities above 75 MVA aggregate; Category 2 covers non-BES IBRs above 20 MVA at 60 kV or higher — a scope capturing most utility-scale solar and a growing share of the battery fleet that DERMS platforms coordinate. The companion PRC-030-1 mandates structured investigation and correction of unexpected IBR performance events, creating a compliance feedback loop that doubles as dispatch-optimization data. For DERMS procurement teams, the deadline (now 75 days out for BES facilities) accelerates the need for real-time telemetry that can verify ride-through performance, document compliance for NERC auditors, and coordinate fleet responses — capabilities core to advanced DERMS but absent from basic SCADA, and a regulatory-risk-mitigation line item that strengthens the DERMS business case on the same investment. The standards also read as the supply-side mirror of FERC’s July 16 computational-load order (below): both responses to the same class of millisecond-scale, fleet-wide disturbance behavior. (Source: pv magazine USA — July 14 entry)

CAISO’s Track 1 “DDEMI” Draft Final Proposal, discussed at a July 9 stakeholder meeting, advances metering reforms that recognize customer-level exports within a resource-level load-curtailment network — an unglamorous but economically decisive M&V fix for behind-the-meter demand response. The reform prevents a DR resource’s dispatched reduction from being understated or double-counted when some enrolled meters are net exporters at dispatch — directly material to the settled capacity value regulated utilities book in IRP filings, since DR’s avoided-cost claim is only as credible as its baseline methodology. It also clears a persistent FERC Order 2222 friction point by clarifying the metering and telemetry rules DER aggregators must satisfy in CAISO, and signals the customer-level sub-metering granularity next-generation DERMS platforms must support. CAISO methodology choices routinely become de facto templates for other ISOs, making Track 1 a leading indicator for national DR settlement rules. (Source: CAISO via Stoel Rives — July 16 entry)

Itron’s TD World analysis “Why DERMS Is Earning a Place Alongside SCADA” — cited briefly in last week’s digest — received full treatment in this week’s log, adding the operational proof points: Xcel Colorado aggregated 15+ MW of customer batteries in under six months via Renewable Battery Connect (the state’s first battery-based VPP), against a backdrop of ~78 GW of coal retirements over two decades and 200 GW of U.S. load growth projected by 2030. Deduplicated against the July 10 digest’s coverage; the incremental takeaways are that utilities are now writing DERMS specifications to SCADA-grade standards (24/7 reliability, defined response times, failover/redundancy) and that full DERMS value requires organizational alignment between grid-operations and customer-program teams — an operating-model change IRP and DSM planners should reflect in implementation cost and timeline assumptions, not just a software purchase. (Source: TD World — July 17 entry; deduplicated against July 10 digest coverage)

🏗️ Data Centers & Large Load Growth

FERC ordered NERC to establish mandatory reliability standards for computational loads — the most consequential federal reliability action of the year for the data-center problem this series tracks. The July 16 order in Docket RD26-7-000, approved 5–0, converts NERC’s voluntary schedule into an enforceable requirement: new or modified standards governing data centers, crypto mines, and other IT facilities must be filed by December 31, 2026 (Phase II work plan by March 1, 2027), and NERC must develop Compliance Registry criteria (Appendix 5B) that would for the first time make large loads themselves registered entities directly accountable under FPA Section 215 — the framework that has governed generators and transmission owners since 2005. The evidentiary record is what utilities should internalize: NERC documented a July 10, 2024 Eastern Interconnection event in which a single 230 kV fault cascaded into six successive faults and the near-simultaneous loss of ~1,500 MW of data-center load, and a review of 26 ERCOT ride-through events found crypto facilities shedding 17–95% of pre-disturbance consumption within milliseconds of normally cleared faults. Chairman Laura Swett called the schedule “not optional.” The practical implication: data-center flexibility and grid-supportive controls behavior are migrating from voluntary “good citizen” commitments toward auditable, standard-backed obligations — which strengthens the case for interconnection-conditioned flexibility and for the telemetry-and-controls stack (i.e., DERMS) that makes compliance verifiable. Drafting is already underway, with standards and glossary revisions out for a 45-day comment period in August and NERC Board approval targeted December 5. (Source: POWER Magazine — July 17 entry)

IEEE Spectrum’s analysis supplies the engineering substrate beneath FERC’s order: data center demand volatility — not just scale — is rewriting grid operating rules, with synchronized AI training clusters producing power fluctuations of 70% or more within milliseconds. The categorical distinction matters for planning: unlike renewable intermittency (supply-side, weather-driven), compute variability is demand-side and driven by workload scheduling — creating a “structural mismatch” where compute infrastructure scales in quarters while electrical infrastructure expands over years. Compute-thermal coupling compounds the effect: cooling demand rises nonlinearly with processing intensity, so a single training job starting or finishing propagates fluctuations through multiple facility systems at once. Geographic concentration (Northern Virginia’s Data Center Alley) means localized substation and corridor stress that system-wide demand metrics never surface. For DERMS and VPP designers, the validated requirement is sub-second, locational response — frequency deviations and voltage fluctuations, not just hourly peak-shaving — which expands the DERMS dispatch mandate toward real-time grid stabilization. (Source: IEEE Spectrum — July 14 entry)

The NCCETC/SEPA “DELTa” database now tracks 77 large-load tariffs across 60 utilities in 36 states — with 29 approved by state regulators in 2025 alone, more than double the 14 approved across all of 2018–2024. The structural trends matter for DER strategy: roughly two-thirds of 2025-approved tariffs set 50–150 MW minimum thresholds (up from the ≤25 MW norms of earlier years), reflecting deliberate regulatory focus on hyperscale customers and cost-causation — upfront network-upgrade costs, minimum-load guarantees, and exit fees. The design features increasingly embed demand-flexibility hooks: Pennsylvania’s proposed model tariff offers lower charges for customers with onsite generation, unused interconnection capacity, or interruptible service, and allows up to 20% load reduction after the initial term. With EPRI projecting data centers could reach 17% of U.S. generation by 2030, and state actions proliferating (NY Energize NY, NC task force, UT SB 132, TX SB 6, CA SB 57, MO executive order), large-load cost allocation is now a first-order affordability battleground in which DSM, load flexibility, and bring-your-own-capacity constructs are being written into the regulatory toolkit as explicit alternatives to unmitigated CapEx. (Source: DSIRE Insight / NCCETC / SEPA — July 16 entry)

📋 Regulatory & Policy

New Jersey Governor Sherrill signed the three-bill energy affordability package on July 7 — resolving the signing-decision watch item from the last two digests — and the signed law reveals a DER capacity-offset mechanism not visible in pre-signing coverage: data centers can offset their capacity obligations by paying third parties to reduce demand elsewhere on the grid through efficiency, DR enrollment, BTM storage, and managed electrification. That provision effectively makes hyperscale facilities a funding source for the DER and VPP programs utilities and DERMS platforms manage — a direct financial linkage between large-load growth and distributed energy investment. The core tariff terms held from the legislative version: a first-of-its-kind ratepayer class for facilities ≥50 MW, 85% take-or-pay for 10 years, and mandatory DR participation. The companion bills are independently consequential: S1673 mandates RTO membership and strips the 50-basis-point ROE adder (~$60M/year ratepayer savings), and the Advanced Grid Technologies Act (S4411) subjects supplemental transmission projects — $14.7B and 79% of NJ’s ratepayer transmission costs over 17 years — to BPU certificate review, with an expedited 120-day track for projects using advanced transmission technologies. That review gate creates a formal venue for non-wires alternatives, including DERMS-coordinated DR and distributed storage, to compete against traditional transmission spend. (Source: Utility Dive / NJ Governor’s Office — July 13 entry; resolves watch item)

ERCOT Batch Zero: resolved — the PUCT approved ERCOT’s batch interconnection framework on June 18, 2026, exactly as scheduled. After five consecutive weeks without a research-log entry, this was confirmed via direct external verification of PUCT/ERCOT announcements, as the prior two digests recommended. The approved protocol revisions establish a batch study process for large loads of 75 MW or greater, grouping qualifying loads into a single system-wide study to allocate available transmission capacity — against a queue ERCOT now sizes at more than 438 GW of large-load requests, nearly 89% from data centers. Next milestones: ERCOT expects to notify Batch Zero applicants of their project classification in August 2026 (when the batch’s full scope becomes known), with a final statewide transmission plan expected in Fall 2027. The lesson flagged in the July 10 digest is now confirmed twice over: the daily log can miss regulatory action entirely, not just lag it — external docket checks remain necessary for deadline-bearing watch items. (Source: external verification — PUCT/ERCOT announcement, June 18, 2026; Utility Dive)

DOE issued its second summer 202(c) emergency order for PJM on July 14 as another Mid-Atlantic heat wave drove a forecasted July 15 peak of 164,553 MW — which would rank third-highest in PJM’s 97-year history, just 12 days after the all-time record. The order (effective through July 21) again authorizes dispatch beyond environmental permit limits and backup-generation direction at large loads as a last resort before firm load shed, and PJM issued both a Maximum Generation Alert and a Load Management Alert signaling possible Pre-Emergency and Emergency DR activation. The economics point for DSM business cases: when DR is activated or on standby multiple times in a single month, the annualized avoided-cost value per MW rises substantially, because the same capacity asset delivers reliability value across repeated discrete events rather than one annual peak. DOE’s estimate that 35+ GW of unused backup generation exists nationwide also quantifies the untapped demand-side resource base that currently lacks the coordination infrastructure — DERMS — to be dispatched outside emergency declarations. (Source: PJM Inside Lines / DOE — July 15 entry)

A Utility Dive investigation put hard numbers on the cost of the supply-side alternative: DOE’s 202(c) campaign keeping seven fossil plants (11 units) online costs ratepayers approximately $550 million per year — while four of the 11 units are not operating, and NERC classifies most affected regions as facing “normal” reliability risk. The plant-level detail is stark: the Campbell coal plant in Michigan cost $401M through Q1, recovered $221M in MISO market sales, and will socialize $180M across MISO ratepayers in states with no connection to the retirement decision; TransAlta Centralia produced zero electricity in 2026 while seeking $19.9M in recovery. NERC’s own reliability lead called the orders “a blunt instrument,” conceding “some are helpful, some maybe not.” For DR advocates this is the sharpest economic counterfactual yet to reach the series: $550M/year for capacity that largely isn’t generating, versus the ~$20M cost of a 70 MW DR portfolio — and versus the 6,113 MW of verified DR delivered during PJM’s record peak. The 90-day rolling order structure also creates a planning paradox: long-term adequacy rationale, short-term authority, undermining the multi-year horizons grid operators actually plan against. (Source: Utility Dive — July 14 entry)

🏭 Utility Programs & Deployments

The Louisiana PSC approved five Entergy Louisiana demand response programs projected to shave 155 MW by 2030 — roughly the capacity of a small gas plant — spanning residential smart thermostats, EV charging, and battery storage (stackable for qualifying customers), agricultural irrigation control, and C&I curtailment. The approval is a clean regulated-utility template for the avoided-cost business case this series is built on: a commission explicitly accepting a multi-sector demand-side portfolio with a defined MW target as a capacity resource, framed by advocates as “investing in people, not expensive new infrastructure.” The Alliance for Affordable Energy’s recommendation to prioritize enrollment in load pockets — supply-demand-imbalanced areas where targeted DR delivers the highest locational value — foreshadows the next maturity step from system-wide enrollment toward locationally optimized dispatch, which is precisely the DERMS and hosting-capacity-analytics value proposition. (Source: Alliance for Affordable Energy / Louisiana PSC — July 16 entry)

🔬 EPRI Research Spotlight

EPRI’s DCFlex initiative and “FlexMosaic” framework anchor Utility Dive’s deep dive on data-center flexibility — finding AI training and inference workloads can offer 18–55% flexibility relative to average consumption while meeting quality-of-service requirements, magnitudes that dwarf traditional residential DR. FlexMosaic defines five flexibility classes (A, infrequent extreme-stress response, through E, frequency stabilization), with EPRI noting that Classes D and E — mitigating local thermal overloads and voltage drops — unlock the most system value: a direct signal that the highest-value data-center dispatch is sub-second and locational, not hourly peak-shaving, consistent with the volatility findings and FERC’s RD26-7-000 order elsewhere in this digest. The economic hook comes from a Duke University Nicholas Institute study: a 1–2% reduction in data-center peak demand can lower electricity rates 0.5–2.8% while protecting reliability. The operating model is equally instructive for DERMS architecture: Emerald AI positions itself as an orchestration layer between utilities and data centers in which utilities retain full dispatch authority while the platform translates grid requirements into feasible dispatch targets and returns telemetry, verification, and event-compliance reporting — the M&V-grade closed loop a resource-level DERMS must provide. Near-term proof points to watch: Emerald AI flexing an NVIDIA data center served by Silicon Valley Power, and the 96 MW Aurora AI Factory in Manassas, Virginia, online late 2026. With U.S. data-center demand projected at 66 GW by 2027 (from 31 GW in 2025) and summer-peak share doubling to 8.5%, these pilots are the evidence regulators will use in deciding whether to mandate flexibility as an interconnection condition. (Source: Utility Dive / EPRI DCFlex — July 16 entry)

EPRI’s 17%-of-U.S.-generation-by-2030 projection also anchored this week’s DELTa large-load tariff analysis, and DCFlex’s flexibility-class taxonomy is a likely reference framework as NERC drafts the computational-load standards FERC ordered July 16 — watch for EPRI participation in the 45-day comment period opening in August.

🚩 Utility-Sector Relevance Flags

⚑ FERC’s Mandatory Computational-Load Standards: Large Loads Become Regulated Reliability Entities
Topic: Regulatory / Large Load / Grid Reliability Standards
Relevance: The July 16 order (RD26-7-000) directing NERC standards by December 31, 2026 and Section 215 registry criteria converts data-center flexibility from a voluntary commitment into a coming auditable obligation — and the standards NERC drafts will define the ride-through, modeling, and controllability envelope that utility interconnection agreements and DERMS integrations must accommodate.
Action Signal: Engage — The 45-day comment window opens in August; utilities with large-load queues should participate directly and align interconnection-study assumptions with the draft standards rather than retrofitting after December approval.

⚑ PJM’s Verified 6,113 MW of DR at Record Peak: The Definitive Performance Citation
Topic: Demand Response Validation / Capacity Value / IRP
Relevance: PJM’s official post-event analysis puts verified DR performance at 6,113 MW during the 168,158 MW all-time peak — worth ~$403M/year in avoided capacity at the Brattle ~$66/kW-year benchmark, delivered while forced outages ran 41–51% above average. This replaces the preliminary “4,000+ MW” figure used in earlier coverage.
Action Signal: Implement — Cite the 6,113 MW / $403M figures in IRP filings and DR cost-effectiveness analyses now; note the official 60-day evaluation may adjust final numbers.

⚑ ERCOT Batch Zero: Resolved After Five Weeks — Approved June 18, Classification Notices in August
Topic: Regulatory / Large Load / Interconnection
Relevance: External docket verification confirmed the PUCT approved the batch framework on schedule against a 438 GW queue (89% data centers). The five-week log gap on a deadline-bearing item confirms the series’ standing lesson: external checks are mandatory for watch items with dates attached.
Action Signal: Watch — August applicant-classification notices will reveal Batch Zero’s actual scope; ERCOT-territory utilities and developers should track classification outcomes as the first hard signal of which loads clear the new process.

⚑ New Jersey’s Signed Capacity-Offset Mechanism: Data Centers as a DER Funding Source
Topic: Regulatory / Large Load / DER Market Design
Relevance: The signed law lets ≥50 MW data centers offset capacity obligations by funding third-party demand reduction — efficiency, DR, BTM storage, managed electrification — creating a compliance-driven revenue stream for DER programs that did not exist in any state framework before this signing.
Action Signal: Engage — NJ-territory utilities and aggregators should scope offset-eligible program structures now, ahead of BPU implementation rules; utilities elsewhere should flag the offset mechanism as a template in pending large-load tariff proceedings.

⚑ NERC PRC-029/030 Ride-Through Compliance: 75 Days Out, $1.54M/Day Exposure
Topic: IBR Compliance / DERMS Requirements / Fleet Operations
Relevance: October 1 BES effectiveness prohibits momentary cessation and mandates reactive-current injection through disturbances for IBR facilities >75 MVA (Category 2 >20 MVA at ≥60 kV from January 1, 2027), with penalties to $1.54M/day/violation. Compliance verification requires exactly the real-time telemetry and fleet coordination advanced DERMS provides.
Action Signal: Implement — Utilities and IBR operators should complete ride-through settings audits and telemetry gap assessments before October 1; DERMS procurement teams should add PRC-029/030 compliance verification to platform requirements.

⚑ 202(c) Economics: $550M/Year for Plants That Largely Aren’t Running
Topic: Capacity Planning / DR Cost-Effectiveness / Emergency Authority
Relevance: The Utility Dive investigation’s numbers — $550M/year across seven plants, four of 11 units idle, $180M socialized across MISO ratepayers, NERC calling benefits “mixed” — form the sharpest supply-side counterfactual yet for demand-side investment, landing the same week DOE issued its second summer PJM 202(c) order.
Action Signal: Implement — Pair the $550M/year figure with PJM’s verified 6,113 MW DR performance in IRP and PUC filings as the documented cost comparison between emergency supply-side retention and demonstrated demand-side capacity.

⚑ EPRI DCFlex 18–55% Flexibility Finding: The Technical Basis for Mandated Data-Center DR
Topic: Data Centers / Demand Flexibility / EPRI Research
Relevance: EPRI’s finding that AI workloads can flex 18–55% at acceptable QoS — with the Duke study showing a 1–2% peak cut lowers rates 0.5–2.8% — supplies the technical and economic evidence base regulators need to condition interconnection on flexibility, exactly as FERC’s RD26-7-000 order moves the obligation federal.
Action Signal: Engage — Utilities negotiating large-load interconnection agreements should reference FlexMosaic classes (especially D/E locational stabilization) in flexibility requirements, and track the Emerald AI/Silicon Valley Power and Aurora AI Factory pilots as citable operational precedents.

⚑ FERC July 20 Generation-Adequacy Reports: Three Days Out
Topic: Regulatory / Large Load / IRP
Relevance: All six RTOs/ISOs must file informational generation-adequacy reports Monday July 20 under the June 18 show-cause orders — each disclosing, in its own words, where adequacy is tightest. The 60-day tariff revisions remain due ~August 17.
Action Signal: Implement — Pull each RTO’s filing directly upon release next week; the disclosed adequacy gaps are directly citable inputs for pending IRP avoided-cost filings and DR program justifications.

📌 Sources

July 13, 2026 Entry
– PJM Inside Lines — PJM Serves Load Through Record-Breaking July Heat (July 10, 2026)
– Vermont Public / Maine Public (New England News Collaborative) — Rooftop Solar Eases New England’s Electric Demand in Heat Wave (July 12, 2026)
– Utility Dive — New Jersey Transmission Oversight, Data Center Affordability Bills Signed (July 7–8, 2026)
– Utility Dive / SEIA — US Energy Storage Installations Hit Q1 Record, Up 32% Year-Over-Year (May 20, 2026)

July 14, 2026 Entry
– IEEE Spectrum — Data Centers and Grid Instability (July 3, 2026)
– Utility Dive — DOE Emergency Power Plants: Reliability Benefits and Costs (June 25, 2026)
– pv magazine USA — Inverter-Based Resource Performance History Leads to Regulatory Change (April 7, 2026)
– Canary Media — Clean Energy Helped the Grid Survive the Heat Wave (July 10, 2026)

July 15, 2026 Entry
– PJM Inside Lines — PJM Hot Weather Operations Update (July 14, 2026)
– pv magazine USA — Sunrun, Tesla, Renew Home Announce Plans for 16.8 GW Virtual Power Plant Program (June 25, 2026)
– pv magazine USA — Tesla Announces 13.5 GWh Energy Storage Deployments in Q2 (July 2, 2026)
– Latitude Media — Xcel’s Tricky Work of Building the First Utility-Owned VPP (April 9, 2026)

July 16, 2026 Entry
– Utility Dive — Data Centers Are Ready to Negotiate Flexibility for Speed (June 26, 2026)
– Alliance for Affordable Energy — PSC Approves Demand Response Programs (March 18, 2026)
– Stoel Rives — Energy Regulatory Update (July 8, 2026)
– DSIRE Insight / NCCETC / SEPA — US Data Center Gold Rush Drives Surge in New Utility Tariffs (April 20, 2026)

July 17, 2026 Entry
– POWER Magazine — FERC Orders Mandatory NERC Reliability Standards for Data Center and Other Computational Loads (July 16, 2026)
– Foley Hoag LLP — Virtual Power Plants: The Distributed Energy Revolution Has Arrived (July 2026)
– TD World — Why DERMS Is Earning a Place Alongside SCADA (June 9, 2026)
– CalCCA — Peninsula Clean Energy & Silicon Valley Clean Energy Jointly Launch Demand Flexibility Initiatives (2026)

External Verification (ERCOT Batch Zero)
– ERCOT / PUCT — PUCT Approves ERCOT’s Batch Zero Process for Connecting Large Electricity Users (June 18, 2026)
– Utility Dive — Texas, Facing 438 GW Queue, Approves Initial Large-Load Interconnection Process (June 2026)
– ERCOT — Market Notice M-B062326-01: Implementation of the Batch Zero Process (June 23, 2026)