Executive answer
The evidence supports a layered answer rather than a simple import-versus-self-sufficiency ranking. A state can produce more electricity than it sells over a year and still face local deliverability problems, while a state that routinely relies on outside power can operate reliably if it has firm external resources and enough transmission. The useful question is therefore not just how much electricity crosses a state boundary, but whether power can reach the relevant load during the hours when the system is stressed. [S24] [S25] [S260] [S261]
The same caution applies to prices. EIA’s state average retail price is a revenue-per-kWh measure, not a direct estimate of the cost of imported electricity. Fuel, generation mix, wholesale prices, transmission, distribution, customer mix, taxes, policy programs and regulatory recovery can all move the final rate. Berkeley Lab’s recent work likewise treats retail-price growth as the combined result of multiple cost drivers rather than a single energy-balance variable. [S106] [S355] [S356] [S394]
This report therefore separates four analytical layers: annual state energy balance, deliverable interregional transfer capability, local flexibility such as generation, storage and demand response, and the outage-and-restoration performance experienced by customers. Keeping those layers distinct makes it possible to explain why the same level of external reliance can be benign in one system and risky in another. [S95] [S118] [S357] [S361]
January-July 2026 utility-scale generation relative to retail sales ranged from 36.5% in Massachusetts to 169.9% in Pennsylvania, confirming large structural differences in state supply. [S105] [S107]
California’s apparent generation gap changes dramatically when roughly 23.0 TWh of estimated small-scale solar is added to utility-scale output, showing how statistical boundaries affect dependency measures. [S105] [S388]
2026 year-to-date state retail prices ranged widely from Texas to California, so abundant local generation or low import dependence does not mechanically produce low customer prices. [S106] [S107] [S355]
During the 2021 cold emergency, ERCOT’s external options were far smaller than the imports available to MISO and SPP, highlighting the operational importance of actual interconnection capability. [S42]
In September 2026, scheduled interchange converted NYISO’s -951 MW in-state day-ahead margin into a positive 931 MW overall margin, demonstrating imports as a planned adequacy resource. [S214]
Key findings
- 36.5% · 66.4% · 169.9% — The January-July 2026 generation-to-sales screen ranged from 36.5% in Massachusetts and 66.4% in Maryland to 169.9% in Pennsylvania, confirming that structural reliance on external supply differs sharply across states. [S105] [S107]
- 82.5% → ~99.5% — California appears to cover only about 82.5% of retail sales with utility-scale generation, but adding roughly 23.0 TWh of estimated small-scale solar brings the screen close to 100%, so the measurement boundary can change the dependency label. [S105] [S388]
- 26.32 · 27.79 · 15.78¢/kWh — 2026 year-to-date retail prices of about 26.32 cents/kWh in Massachusetts, 27.79 in California and 15.78 in Pennsylvania contradict a simple claim that greater local generation surplus necessarily produces lower customer prices. [S106] [S107] [S355]
- 1,220MW vs ~13,000MW — Winter Storm Uri exposed the difference between annual balance and deliverability: ERCOT had roughly 1,220 MW of import capability while MISO and SPP imported nearly 13,000 MW from less-affected parts of the Eastern Interconnection. [S42]
- -951MW + 1,882MW = +931MW — NYISO’s September 2026 report showed 1,882 MW of scheduled interchange turning a -951 MW in-state day-ahead margin into a +931 MW overall margin, demonstrating that external power can be an intentional adequacy resource. [S214] [S260] [S261]
- ~52GW U.S. utility batteries by Jun 2026 — Customer resilience depends on deliverable transmission, generator availability, storage and demand response, extreme-weather preparation and distribution restoration, so annual state import shares cannot substitute for direct reliability and outage measures. [S95] [S118] [S357] [S361] [S371]
A January-July 2026 screen shows how different state supply structures remain. Utility-scale generation was only about 36.5% of retail sales in Massachusetts and 66.4% in Maryland, while Pennsylvania generated about 169.9% of its retail sales. These are screening ratios rather than measured interstate imports, but they clearly identify very different starting positions for local supply and external procurement. [S105] [S107] [S387]
California demonstrates why the accounting boundary matters. Utility-scale generation was 111,466 GWh against 135,102 GWh of retail sales, an apparent coverage ratio of about 82.5%. EIA-861M separately estimates about 23,021 GWh of small-scale solar generation. Once distributed solar is acknowledged, the apparent gap nearly disappears, so a utility-scale-only measure can misclassify states with large behind-the-meter resources. [S105] [S388]
Price comparisons reject a one-to-one link between local surplus and customer cost. Year-to-date 2026 average retail prices were about 27.79 cents/kWh in California, 26.32 in Massachusetts and 15.78 in Pennsylvania, while Texas was around 10.41. These values come from states with very different generation and market structures, which is precisely why an import-share variable cannot substitute for a full retail-cost analysis. [S106] [S107] [S355]
Research scope and definitions
EIA’s net interstate flow is not a matrix of directly measured bilateral state-to-state transactions. It is an annual state accounting construct that reconciles generation with disposition, sales and losses; positive values indicate net inflow and negative values net outflow. The measure is valuable for structural comparison, but it does not identify the supplying state, the physical transmission path or the firmness of the underlying transaction. [S24] [S25]
Political state boundaries also do not match the boundaries used to operate the grid. The Lower 48 is organized mainly around the Eastern, Western and ERCOT interconnections, while balancing authorities and RTOs or ISOs schedule generation, demand and interchange across multi-state footprints. Hourly balancing-authority interchange is therefore an operating measure with different geography and timing from annual state net flow. [S70] [S220] [S394]
For that reason, this report uses ‘self-sufficiency’ only as shorthand for the annual relationship among in-state generation, sales and net energy balance. ‘Resilience’ refers to whether generation, transmission, reserves and restoration capabilities continue to serve load during stress. A low annual import share cannot prove resilience, and a high share cannot prove fragility without evidence on deliverability and system performance. [S95] [S260] [S371]
State supply structure in 2026: what generation and sales can tell us
Current EIA monthly data provide a useful check on whether older structural classifications still point in the same direction. From January through July 2026, Massachusetts recorded 11,223 GWh of utility-scale generation against 30,744 GWh of retail sales. Maryland recorded 23,604 GWh against 35,562 GWh. Pennsylvania moved in the opposite direction, with 144,043 GWh of generation against 84,758 GWh of sales. These comparisons describe a supply structure; they are not direct measurements of interstate trade. [S105] [S107]
California is the most important sensitivity test in the current data. Utility-scale output of 111,466 GWh compared with 135,102 GWh of retail sales suggests a sizable gap if the analysis stops at large plants. Yet EIA’s monthly framework separately estimates about 23,021 GWh of small-scale photovoltaic generation. The difference between those two accounting choices is almost as large as the apparent utility-scale shortfall itself. [S105] [S107] [S388]
Monthly figures should also be treated as a current-condition layer rather than as final annual history. EIA explains that the Electricity Monthly Update relies on samples, model-based prediction and later revisions, and that annual datasets may differ after additional collection and validation. The 2024 State Electricity Profiles remain a stronger baseline for finalized structure, while 2026 data are best used to detect whether the direction of change has shifted. [S26] [S387] [S388]
| Market | Utility-scale generation | Retail sales | Generation/sales screen | 2026 YTD avg. price |
|---|---|---|---|---|
| Massachusetts | 11,223 GWh | 30,744 GWh | 36.5% | 26.32¢/kWh |
| Maryland | 23,604 GWh | 35,562 GWh | 66.4% | — |
| California | 111,466 GWh | 135,102 GWh | 82.5% | 27.79¢/kWh |
| California + small-scale solar | +23,021 GWh | 135,102 GWh | ~99.5% | 27.79¢/kWh |
| Pennsylvania | 144,043 GWh | 84,758 GWh | 169.9% | 15.78¢/kWh |
The generation/sales ratio is a structural screening measure, not a measured interstate-import share. California’s second line is a sensitivity that adds estimated small-scale solar. [S105] [S107] [S388]
Why retail prices do not line up with self-sufficiency
Retail electricity prices combine far more than the cost of producing energy. Average revenue per kilowatt-hour reflects wholesale energy and capacity costs, transmission, distribution, customer composition, taxes, policy programs and regulatory timing. Distribution investment has itself become a major driver in many utilities as aging equipment, wildfire mitigation, resilience upgrades and load growth require new spending. [S106] [S355] [S356]
The 2026 cross-state figures impose a simple but important discipline on causal claims. Pennsylvania’s large generation surplus does not make its retail price a universal floor, and Massachusetts and California both have high prices despite different resource structures. California’s distributed solar also means that a high retail price can coexist with a much smaller apparent energy gap once behind-the-meter generation is counted. A price ranking is therefore not an import-dependence ranking. [S105] [S106] [S107] [S388]
Wholesale markets add another layer of variation. FERC’s market primer and EIA’s regional wholesale updates show that fuel costs, demand, generation availability and transmission congestion affect locational prices hour by hour. Retail rates then add network and regulatory costs on top of those wholesale outcomes. A model that assigns state retail-price differences to annual net imports alone omits the largest parts of the actual cost stack. [S217] [S394]
Transfer capability and resource adequacy: why MW matters more than annual MWh in emergencies
Annual imports answer an energy-volume question, but emergency operations are constrained by instantaneous capability. NERC’s MOD-001 framework calculates available transfer capability from facility ratings, power-flow models, contingencies, existing reservations, reliability margins and other network conditions. MOD-004 separately allows transmission capacity to be preserved as Capacity Benefit Margin for generation-deficiency conditions. These standards formalize the distinction between realized annual imports and deliverable emergency support. [S260] [S261]
MISO’s loss-of-load analysis makes the same distinction in planning practice. Capacity import and export limits are modeled separately from annual energy flows and can become constrained by source generation, transmission contingencies, redispatch limitations or voltage. A region can import many megawatt-hours over a year yet still have insufficient firm transfer during a specific peak or outage condition. [S241]
PJM likewise assigns value to external support through resource-adequacy methods and accepts external capacity subject to locational import and deliverability limits. The relevant planning question is not whether the resource sits outside a state, but whether it can be counted under defined risk conditions. That is a fundamentally different concept from an annual state import percentage. [S140] [S245]
Extreme-weather evidence: Uri, Elliott and the more recent cold events
Winter Storm Uri remains the clearest illustration of how interconnection structure can dominate annual energy balance. The joint FERC-NERC review found that ERCOT’s direct-current ties provided only about 1,220 MW of total import capability. MISO and SPP, by contrast, imported nearly 13,000 MW from less-affected parts of the Eastern Interconnection. ERCOT used the imports it could schedule, but the physical scale of the ties limited the available option set. [S42]
Winter Storm Elliott in 2022 demonstrates the opposite boundary condition. Imports often helped systems maintain reserves and avoid shedding load, but many neighboring balancing authorities entered energy emergencies at the same time and had little surplus to share. A transmission path does not create energy; the reliability value of the connection collapses when correlated weather removes the neighboring surplus on which the import depends. [S43]
Reviews of the 2024, 2025 and 2026 cold periods show that preparedness and rules also change outcomes. Generator winterization, gas-electric coordination, situational awareness and operating procedures improved in several areas, while residual vulnerabilities remained. The evolution of EOP-012 and the new TPL-008-1 extreme-temperature planning standard confirm that resilience is a managed performance outcome, not a permanent property inferred from a state’s annual energy balance. [S62] [S63] [S117] [S118] [S226]
| Case | Observed condition | Role of external power | Key limitation |
|---|---|---|---|
| ERCOT · Uri 2021 | ~1,220 MW import capability | Used available imports but had a small option set | Thin interconnection itself [S42] |
| MISO + SPP · Uri 2021 | Nearly 13,000 MW imported | Large support from less-affected Eastern areas | Requires regional surplus [S42] |
| Elliott 2022 | Many neighbors in emergency together | Imports supported reserves in some hours | Correlated stress reduced energy to share [S43] |
| NYISO · 2026-09-28 | -951 MW in-state + 1,882 MW interchange | Produced +931 MW overall margin | Schedule must remain deliverable [S214] |
| ISO-NE · 2025 | Imports ~7% of regional net energy | Transactions rose in high-demand/high-price hours | Hydrology and transmission outages matter [S180][S305] |
| CAISO · summer 2026 | 2,547 MW planning surplus plus external resources | Out-of-state resources broaden portfolio | Regional heat, wildfire and drought can correlate risk [S185] |
When external power works as a reliability resource: New York, New England and California
NYISO provides a direct operating example. On September 28, 2026, its day-ahead report showed an in-state capacity margin of -951 MW and expected interchange schedules of 1,882 MW, producing a positive overall margin of 931 MW. In this case external power was not an accidental dependency discovered after the fact; it was part of the operating plan that made the day-ahead position adequate. [S214]
New England is another useful case because it is structurally accustomed to external supply. ISO-NE reports that imports supplied about 7% of regional net energy in 2025 and that the region has 14 interconnections with New York and eastern Canada. Its 2025 market-monitoring report also shows that imports became much larger during high-demand and high-price hours, indicating that external transactions respond to scarcity and relative prices rather than behaving as a fixed annual dependency. [S180] [S181] [S305]
California combines external resources with a rapidly changing local portfolio. CAISO’s 2026 summer assessment reported a 2,547 MW planning surplus and included new out-of-state resources such as SunZia wind from New Mexico. Yet CAISO also warns that region-wide heat, wildfire and drought can reduce the reliability value of imports because neighboring systems may be stressed at the same time. Interconnection adds options, but it does not eliminate correlated regional risk. [S185] [S235] [S257]
Why a generation-surplus state can still face reliability constraints: Pennsylvania and PJM
Pennsylvania illustrates the difference between annual energy abundance and real-time deliverability. Its January-July 2026 utility-scale generation was far above retail sales, and nuclear generation alone contributed 43,921 GWh. That makes the state an obvious supply-surplus case in an annual screen. It does not, however, prove that every load pocket can be served under all contingencies. [S105] [S194]
PJM approved an immediate-need reliability project in Pennsylvania in 2026 to address generator-deliverability violations and strengthen a 500-kV corridor. The problem was not an absence of generation in the broad region; it was the ability of the network to move generation to the locations that needed it under planning criteria. This is exactly the kind of constraint that an annual state balance cannot reveal. [S196]
Rapid load growth further weakens the assumption that today’s surplus will persist. PJM’s transmission plan is responding to generator retirements, data-center-driven demand growth and changing regional flows, and 2026 approvals included roughly $11.8 billion of transmission projects. The relevant forward-looking variables are therefore new load, resource retirements and project completion dates, not just the historical sign of net interstate flow. [S194] [S195] [S250]
Long-run change from 2011 to 2026: from annual energy to flexibility and restoration
Since 2011, U.S. reliability policy and market design have increasingly valued services that annual energy statistics cannot measure. FERC Order 755 and 755-A required organized markets to compensate frequency-regulation resources not only for reserved capacity but also for performance. The change recognized that fast and accurate balancing has a distinct reliability value even when the annual megawatt-hour balance appears adequate. [S293]
Storage has made that distinction more important. EIA reports that U.S. utility-scale battery capacity grew at an average annual rate of about 70% over the prior three years, reaching 43.6 GW at the end of 2025 and nearly 52 GW by June 2026. Batteries may not change a state’s annual generation total dramatically, but they can change when the system needs imports, how much reserve it carries and how quickly it responds to short-duration scarcity. [S46]
Demand is evolving at the same time. Data centers, manufacturing and electrification are creating large concentrated loads in regions that may have looked comfortable under older annual balances. FERC’s 2026 actions directed six jurisdictional RTOs and ISOs to justify or revise rules for integrating large loads. Self-sufficiency should therefore be treated as a moving system condition shaped by generation, storage, transmission and demand together. [S250] [S256]
Distribution networks and customer resilience: the layer state energy balances miss
Bulk-system adequacy is not the same thing as customer outage performance. Storms, wildfire, flooding, vegetation and aging distribution equipment can interrupt service even when regional generation and transmission are adequate. Berkeley Lab’s distribution-resilience work therefore evaluates hazard exposure, vulnerability and restoration capability separately from bulk energy supply. [S357]
DOE’s resilience metrics make the distinction explicit. The recommended measures include customer interruptions, SAIFI, SAIDI, CAIDI, equipment repair time and the time required to restore defined shares of customers. NERC’s Transmission Availability Data System separately tracks the frequency, duration and causes of bulk transmission outages. Annual state import shares are absent because they do not directly measure either interruption or restoration performance. [S361] [S371]
This means that a highly import-dependent state is not automatically a state with poor customer resilience, just as a generation-surplus state is not automatically quick to restore service after a major storm. A complete evaluation must add distribution hardening, undergrounding, vegetation management, microgrids, monitoring and restoration capability to the bulk-power analysis. [S120] [S357] [S361]
Counterevidence and alternative interpretations
The first important counterexample is that external power repeatedly reduces rather than increases risk. Scheduled interchange in New York, scarcity-period imports in New England and the large interregional transfers used during the January 2024 Arctic storms all show that interconnection can provide insurance, lower production costs and reduce the need for emergency actions. Defining imports themselves as vulnerability would invert the lesson from these cases. [S23] [S214] [S285] [S305]
The second counterexample is that local abundance does not guarantee deliverability. PJM’s generator-deliverability project, ERCOT’s restrictions on planned transmission outages for summer 2026 and NERC’s generator-outage evidence show that nominal generation capacity can be unavailable or trapped behind network constraints. A local-generation surplus is therefore not a reliability score any more than an import share is a fragility score. [S172] [S196] [S236]
The third issue is measurement uncertainty. Monthly state data contain estimates and revisions; utility-scale generation excludes some distributed generation; industrial self-generation may not appear in retail-sales relationships; and state lines do not match system-operator boundaries. Fine-grained rankings can therefore reflect data definitions as much as real structural differences. The analysis is strongest when it focuses on large contrasts and on operating evidence that directly tests deliverability. [S151] [S387] [S388]
Derived analysis: a four-layer diagnostic for external-power dependence
This report proposes a four-layer diagnostic rather than a single self-sufficiency score. Layer one is the structural annual balance of generation, sales and net flow. Layer two is deliverable transfer capability and the firmness of external resources. Layer three is local flexibility, including available generation, storage, demand response and operating reserves. Layer four is customer-facing outage and restoration performance. A strong result in one layer should never be assumed to imply strength in the others. [S24] [S260] [S261] [S361]
Applied qualitatively, the framework separates several U.S. cases without turning them into a league table. ERCOT during Uri represents a system with substantial local resources but thin external transfer options. NYISO in September 2026 illustrates planned reliance on scheduled interchange. ISO-NE combines structural imports with multiple interfaces and market rules that adjust flows with scarcity. Pennsylvania/PJM shows that a generation surplus can coexist with internal deliverability needs. These are different risk pathways, not ordinal grades. [S42] [S181] [S196] [S214]
The framework also changes the choice of remedy. More generation is not the answer to every dependency problem. The binding constraint may instead be transmission, a firm external contract, storage, demand response, generator winterization or distribution restoration. The practical value of the framework is that it directs attention to the layer where the constraint actually appears before policy or investment is chosen. [S51] [S118] [S120] [S257]
Scenarios and practical implications after 2026
One scenario is faster-than-expected load growth combined with slower transmission and generation delivery. Regions experiencing concentrated data-center or manufacturing demand could move from apparent surplus to tighter import and transfer needs even if their historical energy balance looked comfortable. Under that scenario, interconnection queues, retirements, transmission project milestones and large-load studies matter more than a backward-looking five-year import average. [S194] [S195] [S250] [S256]
A second scenario is continued rapid growth in batteries and distributed resources. Storage can reduce peak net load and change the hours when external power is valuable without adding many annual megawatt-hours of generation. California already shows how distributed solar changes the state-level accounting boundary. As behind-the-meter resources expand, utility-scale generation alone will become an even weaker proxy for dependence. [S46] [S185] [S388]
For operational monitoring, four data blocks should be tracked together: finalized state generation, sales and net-flow balances; RTO or balancing-authority interchange and import capability; local storage, resource availability and demand response; and outage and restoration metrics. Price analysis should add fuel, wholesale-market, transmission, distribution and regulatory-cost data as a separate layer rather than attributing customer rates to the import balance. [S217] [S260] [S355] [S361]
Decision summary
What is well established
The evidence establishes that annual state import or generation balance is not a stand-alone indicator of retail price or resilience. Firm transfer and scheduled interchange have directly covered shortfalls, while generation-surplus areas have still required network reinforcement for deliverability.
What remains conditional
External power is beneficial to the extent that neighboring systems retain simultaneous surplus, transmission paths and contracts remain usable during the relevant stress hours, and local generation, storage and demand response can absorb short-duration disruptions or congestion.
What cannot yet be answered
The available evidence does not support one universal causal coefficient stating how much external dependence changes retail prices or outage duration in every state. That would require harmonised state, utility and hourly data on transactions, power flows, costs, interruptions and restoration.
Practical actions
- Monitor state self-sufficiency with annual generation, sales and net flow alongside the relevant RTO or balancing-authority import capability and actual interchange so structural dependence is separated from deliverability.
- When comparing retail prices, track fuel, wholesale energy and capacity, transmission, distribution, customer mix and policy costs separately from import dependence rather than assigning high rates to external procurement alone.
- Evaluate resilience with generator availability under extreme weather, storage and demand response, transmission outages and customer restoration times so the assessment reflects actual service interruption and recovery rather than annual energy abundance.
- Forecast future dependence using large-load interconnection, generator retirements, storage additions, transmission-project completion and distributed solar growth instead of mechanically extending historical net-import averages.
Limitations and final conclusion
The main limitation is geographic mismatch. State political boundaries, utility territories, balancing authorities and RTO or ISO footprints overlap in different ways across the country. The same numerical state import share can therefore describe very different operating arrangements in New England, PJM, MISO, California or ERCOT. Current 2026 monthly values also include estimates that may change when final annual datasets are published. [S70] [S194] [S210] [S387]
Even with those limitations, the evidence supports a strong conclusion. External dependence can be measured as a structural energy-balance condition, but it cannot be used by itself as a price or resilience indicator. The actual risk is determined by the combination of firm transfer capability, simultaneous neighboring surplus, local generation and storage, internal network constraints, operating flexibility and restoration performance. Where these conditions are strong, imports can act as both an economic and reliability resource; where they are weak or highly correlated with regional stress, dependence becomes more fragile. [S42] [S43] [S214] [S260]
The final answer to the research question is therefore conditional but not vague: states can be identified as structurally more or less reliant on external electricity, yet the meaning of that reliance depends more on when, where and how reliably outside power can be delivered than on the annual quantity imported. U.S. electricity self-sufficiency should be evaluated by separating annual state balance from interregional deliverability, local flexibility and customer restoration outcomes. [S24] [S25] [S260] [S261] [S361]
Sources and evidence notes
S24 · U.S. Energy Information Administration (EIA) — EIA’s glossary defines net interstate flow as a state accounting balance rather than a bilateral transaction matrix. That distinction is fundamental because the metric cannot identify the supplying state or the physical transmission path. EIA Glossary — Net interstate flow of electricity
S25 · U.S. Energy Information Administration (EIA) — The SEDS technical notes explain the calculation of state electricity flows and losses, their separation from international trade, and the use of estimation. They support treating annual net flow as a structural indicator rather than a direct operating flow measure. State Energy Data System 2024: Consumption — Section 6 Electricity
S42 · Federal Energy Regulatory Commission (FERC), NERC, and Regional Entities — The Uri inquiry records roughly 1,220 MW of ERCOT import capability compared with nearly 13,000 MW imported by MISO and SPP from less-affected parts of the Eastern Interconnection. It is the clearest case showing why interconnection scale changes emergency options. The February 2021 Cold Weather Outages in Texas and the South Central United States
S43 · Federal Energy Regulatory Commission (FERC), NERC, and Regional Entities — The Elliott inquiry shows both sides of external supply: purchased power often helped preserve reserves, but simultaneous emergencies across neighboring areas left little energy to share. Import value is therefore conditional on regional correlation. Winter Storm Elliott Report: Inquiry into Bulk-Power System Operations During December 2022
S46 · U.S. Energy Information Administration (EIA) — EIA reports extraordinary growth in utility-scale batteries, reaching 43.6 GW at the end of 2025 and nearly 52 GW by June 2026 after average annual growth of about 70% over the prior three years. Storage changes the timing of scarcity and import needs. Battery storage capacity averaged 70% growth over the last three years
S95 · North American Electric Reliability Corporation (NERC) — NERC’s State of Reliability evaluates severe weather, bulk-system performance, frequency response and storage rather than using state energy balances as a proxy. It provides an official framework for separating annual structure from reliability outcomes. 2025 State of Reliability — Assessment Overview of 2024 Bulk Power System Performance
S105 · U.S. Energy Information Administration (EIA) — EIA’s Electric Power Monthly state generation table provides the 2026 utility-scale generation values used in the Massachusetts, Maryland, Pennsylvania and California supply screen. Electric Power Monthly Table 1.3.A — Utility Scale Facility Net Generation by State
S106 · U.S. Energy Information Administration (EIA) — EIA’s end-use update provides state average revenue per kWh and retail-sales trends. The definition matters because the measure is average revenue divided by sales, not an individual tariff or a pure energy-cost metric. Electricity Monthly Update — End Use
S118 · Federal Energy Regulatory Commission (FERC) — FERC’s approval of the revised EOP-012 cold-weather standard changed preparedness requirements beginning in October 2025. Reliability conditions across the study period therefore cannot be treated as institutionally constant. FERC Takes Action to Enhance Reliability of the U.S. Electric Grid
S140 · PJM Interconnection — PJM’s resource-adequacy study explicitly includes a Capacity Benefit of Ties, showing that external interconnection is modeled as a bounded reliability benefit rather than automatically classified as dependency risk. 2025 PJM Effective Load Carrying Capability and Reserve Requirement Study (ELCC/RRS)
S180 · ISO New England — ISO-NE reports that imports supplied about 7% of regional net energy in 2025, down from 16% in 2021. The figures show that structural reliance changes over time with resource conditions and infrastructure. 2026 New England Regional Profile
S185 · California Independent System Operator (CAISO) — CAISO’s 2026 summer assessment combines a 2,547 MW planning surplus and new external resources with warnings about region-wide heat, wildfire and drought. It captures both the benefit of interconnection and the limitation created by correlated stress. 2026 Summer Loads and Resources Assessment Overview
S194 · PJM Interconnection — PJM’s RTEP documents generator retirements, rapid data-center-driven demand growth and changing regional flows. It demonstrates why a current state generation surplus is not a fixed forecast of future adequacy. Regional Transmission Expansion Plan (RTEP)
S196 · PJM Interconnection — PJM’s Pennsylvania reliability project addresses generator-deliverability violations on a 500-kV corridor. It directly shows that abundant generation and network deliverability are different concepts. PJM Board Approves Transmission Project Addition, Changes To Support Grid Reliability
S214 · New York Independent System Operator (NYISO) — NYISO’s September 28, 2026 report shows 1,882 MW of scheduled interchange turning a -951 MW in-state margin into a +931 MW overall margin. It is a concrete example of imports functioning as planned adequacy supply. NYISO Capacity Report — September 28, 2026
S241 · Midcontinent Independent System Operator (MISO) — MISO’s LOLE study calculates capacity import and export limits separately from annual energy flows and tests generation, transmission and voltage constraints. It supports using MW transfer capability rather than annual MWh for adequacy analysis. Planning Year 2026-2027 Loss of Load Expectation Study Final Report
S260 · North American Electric Reliability Corporation (NERC) — NERC MOD-001 defines ATC and AFC using ratings, contingencies, reservations, power-flow models and reliability margins. It formally separates realized energy imports from transfer capability available under stressed conditions. MOD-001-1a — Available Transmission System Capability
S261 · North American Electric Reliability Corporation (NERC) — NERC MOD-004 establishes Capacity Benefit Margin procedures for preserving transmission capacity during generation-deficiency events. It shows that external support can be an intentionally reserved reliability tool. MOD-004-1 — Capacity Benefit Margin
S355 · Lawrence Berkeley National Laboratory (Berkeley Lab) — Berkeley Lab’s 2026 retail-price work attributes rate trends to multiple drivers including gas, capacity, transmission, distribution, policy costs, generation mix and load growth. It constrains any explanation based on imports alone. Retail Electricity Price Trends and Drivers: Data Update—2026 Edition, July Update
S357 · Lawrence Berkeley National Laboratory (Berkeley Lab) — Berkeley Lab’s distribution-resilience work evaluates hazards, vulnerability and restoration capability separately from bulk supply. It supports treating customer resilience as a distinct layer of the analysis. Integrating Resilience Planning in Distribution System Planning
S361 · U.S. Department of Energy, Grid Deployment Office — DOE’s resilience metrics focus on interruptions, SAIFI, SAIDI, CAIDI, repair time and customer restoration milestones. State import shares are absent because they do not directly measure outage or recovery performance. Grid Resilience Formula Grant Metrics — Tracking Guidance
S387 · U.S. Energy Information Administration (EIA) — EIA’s monthly methodology explains the use of samples and model-based estimation and notes that later data collection can revise results. This is why 2026 figures are treated as current-condition evidence rather than final history. Electricity Monthly Update — Methodology and Documentation
S388 · U.S. Energy Information Administration (EIA) — The EIA-861M methodology documents state adjustments and separate estimates for small-scale solar. It is the methodological basis for the California boundary sensitivity test. Form EIA-861M Detailed Data — Sampling, State Adjustments, and Small-Scale PV Estimates
S394 · Federal Energy Regulatory Commission (FERC) Staff — FERC’s Energy Primer explains that wholesale prices are location- and time-specific and reflect dispatch and transmission constraints. It supports keeping wholesale price, retail price and annual state balance as separate analytical variables. Energy Primer: A Handbook of Energy Market Basics
Green Map creates custom-edited map images using open geographic data sources such as geoBoundaries, Natural Earth, OpenStreetMap, and government open data. These maps are edited visual materials, not raw data files, and are provided for education, documents, presentations, and graphic reference.