According to recent analysis published by MV2 of Munim, Inc. utilizing data from the U.S. Energy Information Administration's (EIA) monthly generator inventory, known as EIA-860M, grid connection delays are becoming increasingly prevalent. The research investigates whether planned power plants actually come online during their projected operational month, highlighting a notable downward trend across multiple generation types.
Tracking Grid Connection Slips
The core metric tracked in the research measures the share of generators listed as planned for the next month that successfully report operating within that timeframe. According to source s1, this operating share dropped significantly from 0.469 during the 2016–2019 period down to 0.349 for 2022–2025. Source s2 further contextualizes this dataset by treating monthly Form EIA-860M filings as rolling forecasts, scoring planned operating and retirement months against subsequent agency reporting. Every calendar year from 2021 through 2026 sits below every year from 2016 to 2020 in this long-term tracking series.
Technology Breakdowns and Structural Shifts
While rapid expansion in renewable energy often receives blame for grid integration bottlenecks, the analysis indicates that schedule adherence drops are largely intrinsic to individual technologies rather than merely a consequence of shifting energy mixes. Source s1 reports that solar reliability dropped from 0.506 to 0.354, while battery storage reliability fell from 0.442 to 0.299. Although battery installations surged from 3% to 25% of the generator mix—a structural shift accounting for about 21% of the overall decline—the remaining 79% of the drop occurs internally within specific technology categories. Gas generation initially appeared stable or rising, but closer inspection revealed that a single plant's 12 combustion turbines, repeatedly listed across 102 entries between June 2016 and July 2018, artificially skewed the figures; without that anomaly, gas performance remained flat.
Understanding Broader Implications for Grid Planning
This investigation matters because grid operators, policymakers, and energy developers rely heavily on the EIA-860M inventory to forecast capacity additions and manage regional grid reliability. When nearly two-thirds of generators planned for an upcoming month miss their target completion date, electricity planning models face compounding forecasting errors. Furthermore, these findings challenge the conventional assumption that battery and solar integration issues alone drive interconnection delays. By demonstrating that scheduling misses are worsening internally across legacy and emerging technologies alike, the research highlights systemic friction in project completion pipelines. Source s2 adds critical context regarding data hygiene, exposing how third-party aggregators like Catalyst Cooperative's PUDL changelog can inadvertently drop recurring status updates due to deduplication rules, and how EIA's own historical workbooks occasionally omit retired units or shift reporting scopes. For energy analysts and software developers building reproducible energy workflows, these discrepancies mean that raw data ingestion requires rigorous validation, pre-registered testing, and careful handling of multi-year generator histories to avoid drawing false conclusions about grid development velocity.
Methodological Limitations and Code Availability
Both sources emphasize strict methodological boundaries and known data limitations. The monthly share metrics calculated in s1 count generator-months rather than distinct physical plants, meaning a repeatedly delayed project multiplies its weight in the statistics. Source s2 outlines additional structural limitations, noting that missing records in historical EIA workbooks—such as the November 2024 file omitting 279 generators—and shifting definitions on retired generator sheets require transparent, version-controlled code to audit. To address these hurdles, the associated open-source repository munimtechnologies/eia860m-slip provides pre-registered predictions, plain Python scripts, and standard tests under an Apache-2.0 license, allowing independent verification of every number cited in the reports.
Editorial graphic and headline text by Signal Titans.
Source: Why fewer U.S. power plants come online in their planned month: mostly not the solar and battery boom · GitHub – munimtechnologies/eia860m-slip: Do U.S. generators planned in EIA-860M come online (or retire) when planned? Scores every planned month as a forecast. MV2 of Munim, Inc. · GitHub
SIGNAL TITANS
Leave a Reply