State-wise Economic Viability of Long-Duration Energy Storage Systems in the United States

Alexandre Moreira; Andre Marcato; Miguel Heleno; Patricia Silva

REVIEW 3 major objections 5 minor 18 references

State-wise Economic Viability of Long-Duration Energy Storage Systems in the United States

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Under the 2030 cost target of $1,100 per kilowatt, only four contiguous U.S. states would find 100-hour storage economical for replacing gas and coal in 2050.

desk verdict A solid, policy-relevant state-by-state LDES viability map whose headline 'only 4 states' claim hangs on an untested no-trade assumption and a single storage configuration. read the letter →

arxiv 2505.07624 v1 pith:FIJ73VYC submitted 2025-05-12 math.OC

Alexandre Moreira , Patricia Silva , Miguel Heleno , Andre Marcato This is my paper · ORCID

classification math.OC

keywords long-durationenergystorageviabilitycost100-hourcapacityexpansionthermalgenerationretirementstate-wiseanalysiseconomicsround-tripefficiency

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that the current U.S. cost target for multi-day storage is set too high to justify replacing gas and coal plants in most states. It computes a viability cost for each of the 48 contiguous states: the maximum price per kilowatt that a 100-hour, 42.5%-efficient storage system can carry and still keep total system cost no higher than a system that keeps its thermal plants. The results range from -$11.81/kW to $5,993.94/kW, and only Kansas, North Dakota, Nebraska, and Virginia clear the $1,100/kW target. The finding matters because it separates the cost question from the value question: even if storage technology reaches its target price, most states would not choose it for full thermal replacement under these assumptions.

What carries the argument

The central object is the LDES viability cost, defined as the maximum avoided total system cost divided by the installed storage power capacity. It is computed by solving two optimization problems in sequence: a baseline model that minimizes total system cost with existing generators only, and an opportunity-value model that allows new intermittent capacity and batteries, retires gas and coal, fixes a candidate LDES capacity, and requires total cost to stay at or below the baseline. The largest per-kilowatt value found across the tested 0-150 GW capacity range is the state's maximum viability cost.

What would settle it

Re-run the two models with cross-state transmission allowed at existing or planned transfer capacities; if the number of states clearing the $1,100/kW threshold exceeds four, the headline result is an artifact of the isolation assumption.

Watch

Extended reading notes

Core claim

For each state's 2050 system, the authors replace gas and coal generation with a mix of new wind, solar, 4-hour batteries, and a fixed amount of 100-hour storage, and then ask how much that storage could cost per kilowatt before the replacement stops being cost-neutral. They call this the viability cost. Across states it spans from -$11.81/kW to $5,993.94/kW, and the headline result is that only Kansas, North Dakota, Nebraska, and Virginia exceed the $1,100/kW multi-day storage target. Replacing thermal plants in the 43 states with positive viability costs would require at least 646.09 GW (64.61 TWh) of 100-hour storage, and 1,009.30 GW (100.93 TWh) at the capacities that maximize viability cost.

Load-bearing premise

The paper models each state as an isolated grid with no electricity imports and no new transmission, so a state with poor local wind and solar cannot draw on clean power from outside; the four-state result depends on that isolation.

Editorial extensions

If this is right

If the target falls to $500/kW, the viable list grows to nine states: California, Colorado, Idaho, Kansas, Montana, North Dakota, Nebraska, New Mexico, and Virginia.
At $300/kW, 17 states clear the bar, adding Maine, Minnesota, North Carolina, Nevada, Oklahoma, Texas, Washington, and West Virginia.
The minimum 100-hour storage needed for thermal replacement in the 43 states with positive viability costs is about 646.09 GW (64.61 TWh), and 1,009.30 GW (100.93 TWh) at the capacity that maximizes each state's viability cost.
States with high viability costs generally have low thermal participation, low thermal utilization, wind-dominated intermittent generation, and higher thermal fixed O&M costs; low-viability states generally rely on thermal, have solar-dominated intermittent mixes, and lower intermittent capacity factors.
Five states (Alabama, Connecticut, Delaware, New Jersey, and Ohio) have negative viability costs, meaning replacing gas and coal with intermittent generation plus storage raises total system cost at any storage price.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If states were allowed to trade electricity across their borders, the list of viable states would probably grow, because renewable-rich states could sell surplus power to resource-poor neighbours; the paper's isolation assumption is a plausible driver of the four-state result.
The gap between the paper's minimum 646 GW of 100-hour storage and the 225-460 GW national estimate suggests that wider demand projections assume storage plays roles other than full thermal replacement, or that full thermal replacement is too strict a benchmark for many states.
Negative-viability states could still find LDES worthwhile for services such as capacity adequacy or transmission deferral; the paper's value measure counts only avoided thermal-plant costs.
A direct test of robustness would vary round-trip efficiency and duration around the iron-air baseline (42.5% RTE, 100 hours) to see whether the set of four states changes when the storage technology's parameters move.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

First thing to know: this is a competent, clearly written extension of the authors' own California method to all 48 contiguous states, and the new output—per-state $/kW viability costs for 100h, 42.5% RTE LDES in 2050—is genuinely new and useful for policy conversation. The driver analysis (low thermal share, high wind share, higher thermal FO&M => higher viability cost) is plausible and supported by the table.

The modeling is internally coherent. The viability cost is the max avoided cost divided by LDES capacity, an optimization output compared to an external DOE target, so there's no circularity. The negative-cost states (AL, CT, DE, NJ, OH) are explained by weak renewables, not by a fitting artifact. The supplementary dashboard and per-state figures are a plus.

The soft spot that matters: Section 4.6 treats each state as an island with no inter-state trade or new transmission. The paper states this clearly, but never tests it. This is not a minor detail. States with poor local wind/solar must build expensive local intermittent capacity to retire gas and coal, which lowers the computed avoided cost and hence the viability cost. In reality those states would import clean power, which would raise their true viability cost. The five negative-cost states are small or import-dependent systems; that pattern strongly suggests the assumption is doing real work. So the headline 'only 4 states clear $1,100/kW' is conditional on an assumption that could plausibly change the ranking. The 4x/10x build caps and the single 100h/42.5% storage configuration are secondary but also need sensitivity analysis.

The reader's stress-test holds up on reading the paper; I don't see a way to dismiss it. That said, the central argument is not flawed—it's a conditional result that is honestly labeled. The paper would improve with a scenario allowing regional trade or at least a discussion of how imports would shift the threshold. No code or data files are provided, which limits reproducibility, though the public data sources are named.

Verdict: this deserves peer review. I'd send it out. It's the kind of applied result a referee can push to a much stronger conditionally-accepted paper. For my own work I wouldn't cite the headline number without checking the sensitivity, but I'd bring it to reading group as a case study in how strong assumptions drive state-level storage results.

Referee Report

3 major / 5 minor

Summary. The paper estimates, for each of the 48 contiguous U.S. states in a 2050 scenario, the maximum capital cost (in $/kW) of a 100-hour, 42.5% round-trip-efficiency long-duration energy storage (LDES) system such that the state could retire all gas- and coal-fired generation and replace it with a mix of intermittent renewables, short-duration storage, and LDES without increasing total system cost. This 'viability cost' is computed as the ratio of the maximum opportunity value from a constrained optimization (Eqs. 2-3) to the installed LDES capacity. The headline result is that only Kansas, North Dakota, Nebraska, and Virginia have viability costs above the DOE's $1,100/kW target; five states have negative viability costs. The paper also reports aggregated LDES capacity requirements and correlates high viability costs with low thermal utilization, high wind shares, and higher thermal fixed O&M costs.

Significance. The paper provides a consistent, transparent method for state-level LDES valuation using public datasets, and its negative result about the DOE target, if robust, would be policy-relevant. The framework is reproducible in principle and the supplementary material is extensive. However, the headline ranking is conditioned on several strong modeling choices that are stated but not stress-tested, most notably the autarkic treatment of each state and the exclusion of all firm low-carbon alternatives other than gas/coal retirement. These choices could materially change the set of states above the target, so the central claim should be interpreted with caution until sensitivity analyses are provided.

major comments (3)

[Section 4.6] The state isolation assumption is load-bearing for the headline claim. The model requires each state's balancing areas to meet the state's demand with no inter-state trade and no new transmission, which forces resource-poor states to build local intermittent capacity at high cost and lowers their computed avoided cost and hence their LDES viability cost. The five states with negative viability costs (AL, CT, DE, NJ, OH) are small or import-dependent systems, consistent with this mechanism. Since the paper's central result is that only four states clear the DOE target, the authors should either relax this assumption (e.g., group states into regions with trade) or provide a rigorous sensitivity analysis showing that the ranking is unchanged. Without that, the 'only 4 states' conclusion is not robust to a reasonable and realistic modification of the model.
[Sections 2.1 and 4.4] The opportunity-value scenario retires gas and coal and admits only intermittent generation, SDES, and LDES as replacements; nuclear, geothermal, and fossil with carbon capture are not candidate resources. The resulting viability cost is thus conditional on the exclusion of these alternatives, which in many states could provide firm capacity at lower cost than large amounts of local renewables plus storage. The paper should report at least a bounding analysis (e.g., adding a generic firm low-carbon resource at ATB cost) and discuss how the DOE-target comparison would change. As it stands, the policy conclusion that the target is sufficient in only four states is framed more strongly than the model supports.
[Section 4.7] The candidate renewable build limits of 4x (and 10x for CT, DE, PA) the existing installed capacity are ad hoc and could bind for states with limited land or already high renewable penetration. If these caps bind, the optimization cannot choose the least-cost replacement mix, and the implied viability costs are depressed. The authors do not report which states or technologies hit these caps, nor do they test the sensitivity of the state ranking to the multiplier. This should be documented and, ideally, relaxed in a sensitivity case.

minor comments (5)

[Section 4.2] The phrase 'baseline bodel' appears to be a typo for 'baseline model'.
[Equations (2)-(3)] The variables C_over, q_over, c_VC, and x_power are introduced without definitions or units; the paper should include a notation table or refer explicitly to [7] with a short description so the metric is self-contained.
[Section 2.2] The definition of 'minimum required LDES capacity' (646.09 GW) should be stated precisely; it is not clear whether this is the smallest capacity at which viability cost is non-negative or the capacity that maximizes viability cost.
[Supplementary information] The supplementary figures appear to contain garbled text (sequences of 'uni0000' characters) in the provided version; the authors should verify the PDF rendering.
[Throughout] The term 'IES' (intermittent energy sources) is used throughout but not defined at first use; please define all abbreviations on first appearance.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: viability costs are optimization outputs benchmarked against an external DOE target.

full rationale

The paper's derivation chain is self-contained. It takes external 2050 state-level data (Cambium 2022, ReEDS, ATB, AEO), solves a baseline cost-minimization model and then an opportunity-value maximization model (Eqs. 1-3), and defines the LDES viability cost as the break-even per-kW cost that keeps total system cost at the baseline level (c_VC in Eq. 2 under constraint Eq. 3). The headline claim that only four states clear US$1,100/kW is a direct comparison of these computed break-even costs to an externally set DOE target, not a fitted parameter renamed as a prediction. The only self-reference is the use of the authors' earlier optimization framework [7]; that framework is a general mathematical model applied here to new state-level inputs, and the state ranking is not an input to [7], so the citation is methodological rather than load-bearing circularity. The state-isolation assumption (Sec. 4.6) may affect robustness, but it is a stated modeling assumption, not a circular reduction of the result to its inputs. The correlational driver analysis (Sec. 2.3) is descriptive and does not claim a first-principles derivation. No step in the paper exhibits Eq.-to-Eq. equivalence by construction or a fitted-input-called-prediction pattern.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The analysis relies on several hand-set modeling parameters and domain assumptions. The free parameters (capacity caps, reserve margin, K-means clusters, storage configuration) are arbitrary choices that affect results; the axioms include the island-state assumption, full thermal retirement, the Cambium 2050 scenario, and the exclusion of other firm low-carbon resources. No new physical entities are introduced.

free parameters (5)

Candidate intermittent capacity multiplier = 4x (10x for CT, DE, PA)
Maximum investment in candidate solar and wind generators is set to 4 times the installed capacity of the corresponding fixed generator (10 times for CT, DE, PA). This arbitrary cap limits the ability to replace thermal generation and directly affects viability costs.
Candidate SDES capacity multiplier = 10x existing SDES capacity
Short-duration storage investment is capped at 10 times the state's existing SDES capacity, an ad hoc constraint that could force more LDES deployment and raise its apparent value.
Reserve margin = 4% of demand
A fixed 4% reserve requirement is imposed across all time periods and states; no sensitivity is reported, and the value is chosen by the authors as standard practice.
K-means generator clusters = 3 per balancing area
Generators are aggregated into three representative cost categories per balancing area to reduce computational complexity, with no validation of the clustering error or its impact on results.
LDES duration and round-trip efficiency = 100 hours, 42.5% RTE
All results assume a single technology configuration based on iron-air storage references. Other durations and efficiencies are not tested, so the findings do not generalize to all LDES technologies as the title and abstract suggest.

assumptions (5)

domain assumption Each state is an isolated system with no inter-state electricity trade and no new transmission investment
Section 4.6 states that the combined generating capacity of all balancing areas within a state must meet the state's demand. This is the most load-bearing assumption because it prevents resource-sharing across state lines and likely lowers viability costs in resource-poor states.
ad hoc to paper Gas and coal generators are fully retired in the opportunity value model
Section 4.4 says retirement of gas and coal generators is considered, and the analysis replaces them completely. Partial retirement or retention of some thermal capacity is not explored, so the viability costs apply only to the full-replacement scenario.
domain assumption The 2050 Cambium 2022 energy matrix (no tax credit phaseout scenario) is an accurate projection of each state's future system
Section 4.5 takes fixed generators, load, and installed capacities from NREL's Cambium 2022 dataset. All results inherit any errors or biases in that projection.
ad hoc to paper No other firm low-carbon technologies (nuclear, geothermal, CCS) are candidate replacements for gas and coal
The opportunity value model allows investment only in intermittent sources, SDES, and LDES. Other firm clean resources that could substitute for thermal generation are excluded, potentially overstating the value of storage.
standard math The optimization problems are solved to global optimality
The paper does not provide solver details, tolerances, or optimality certificates, but the methodology implicitly assumes the reported optima are correct.

how reviews work

0 comments

Cite this review

Pith. "Pith review of State-wise Economic Viability of Long-Duration Energy Storage Systems in the United States." pith.science (2026). https://pith.science/paper/FIJ73VYC

@misc{pith2026250507624,
  author       = {Pith},
  title        = {Pith review of: State-wise Economic Viability of Long-Duration Energy Storage Systems in the United States},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FIJ73VYC}},
  note         = {Machine review of arXiv:2505.07624}
}

abstract

Long-duration energy storage (LDES) assets can be fundamental resources for the next-generation power systems. However, LDES technologies are still immature and their future technology costs remain highly uncertain. In this context, we perform in this paper an extensive study to estimate the maximum LDES technology costs (which we define as viability costs) under which LDES systems would be economically viable in each state of the contiguous U.S. according to their characteristics. Our results indicate that only 4 states (out of 48) would be able to remove firm conventional generation supported by LDES systems without increasing their total system costs under the current US-DOE cost target of 1,100 US$/kW for multi-day LDES. In addition, we find that states with the highest LDES viability costs have in general low participation of thermal generation, a high share of wind generation, and higher thermal-related fixed operation and maintenance (FO&M) costs.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

18 extracted references · 18 canonical work pages

[1]

Frontiers in Energy Research8(2020)

Zhang, J., Guerra-Fernandez, O.J., Eichman, J., Pellow, M.A.: Benefit analysis of long-duration energy storage in power systems with high renewable energy shares. Frontiers in Energy Research8(2020)

work page 2020
[2]

In: 2023 North American Power Sympo- sium (NAPS), pp

Colombo, M., Kurtz, S.: Value of long-duration energy storage and oxy- combustion in renewables driven grids. In: 2023 North American Power Sympo- sium (NAPS), pp. 1–5 (2023). IEEE

work page 2023
[3]

Nature Communications15(1), 9501 (2024)

Staadecker, M., Szinai, J., S´ anchez-P´ erez, P.A., Kurtz, S., Hidalgo-Gonzalez, P.: The value of long-duration energy storage under various grid conditions in a zero-emissions future. Nature Communications15(1), 9501 (2024)

work page 2024
[4]

Energy Environ

Guerra, O.J., Zhang, J., Eichman, J., Denholm, P., Kurtz, J., Hodge, B.-M.: The value of seasonal energy storage technologies for the integration ofwind and solar power. Energy Environ. Sci.13, 1909–1922 (2020)

work page 2020
[5]

Joule 5(8), 2077–2101 (2021)

Hunter, C.A., Penev, M.M., Reznicek, E.P., Eichman, J., Rustagi, N., Baldwin, S.F.: Techno-economic analysis of long-duration energy storage and flexible power generation technologies to support high-variable renewable energy grids. Joule 5(8), 2077–2101 (2021)

work page 2021
[6]

Nature Energy6(5), 506–516 (2021)

Sepulveda, N.A., Jenkins, J.D., Edington, A., Mallapragada, D.S., Lester, R.K.: The design space for long-duration energy storage in decarbonized power systems. Nature Energy6(5), 506–516 (2021)

work page 2021
[7]

IEEE Transactions on Energy Markets, Policy and Regulation (2024) 15

Silva, P., Moreira, A., Heleno, M., Marcato, A.L.M.: Boundary technology costs for economic viability of long-duration energy storage systems in california. IEEE Transactions on Energy Markets, Policy and Regulation (2024) 15

work page 2024
[8]

Accessed on May 17, 2023

Gagnon, P., Cowiestoll, B., Schwarz, M.: Cambium 2022 Data. Accessed on May 17, 2023. https://scenarioviewer.nrel.gov

work page 2022

Show all 18 references

[9]

Accessed on March 22, 2023

Cole, W., Brown, M., Brown, P., et al.: Regional Energy Deployment System Model 2.0 (ReEDS 2.0). Accessed on March 22, 2023. https://www.nrel.gov/ analysis/reeds/index.html

2023
[10]

Accessed on March 29, 2023

NREL: ATB Electricity Data 2022. Accessed on March 29, 2023. https://atb.nrel. gov/electricity/2022/data

2022
[11]

Energy Information Administration: Annual Energy Outlook 2023

U.S. Energy Information Administration: Annual Energy Outlook 2023. Accessed on July 1, 2023. https://www.eia.gov/outlooks/aeo/

2023
[12]

Technical report, U.S

Scott, K., Hendrickson, S., Ryan, N., Dawson, A., Kort, K., Shrager, B., Siberry, V., Spitsen, P., Babinec, S., Balducci, P., Zhou, Z.: Pathways to commercial liftoff: Long duration energy storage. Technical report, U.S. DOE (March 2023). https://liftoff.energy.gov/

2023
[13]

Accessed on March 7, 2025

DOE: Long-Duration Energy Storage Demonstrations Program – Multiday Iron Air Demonstration. Accessed on March 7, 2025. https://www.energy.gov/sites/ default/files/2024-06/FactSheet LDESAward Xcel 6.5.24 v3.pdf

2025
[14]

Walsh, F.: A review of the iron–air secondary battery for energy storage

McKerracher, R.D., Ponce-de-Le´ on, C., Wills, R.G.A., Shah, A.A., C. Walsh, F.: A review of the iron–air secondary battery for energy storage. ChemPlusChem 80(2015)

2015
[15]

Accessed on March 19, 2025 (2025)

Silva, P., Moreira, A., Heleno, M., Luis Marques Marcato, A.: Economic via- bility of 100-h long-duration energy storage (LDES) systems in 2050 - Tableau Online. Accessed on March 19, 2025 (2025). https://public.tableau.com/ views/LDES national study online v1/Painel 1?:langua...

2025
[16]

International Journal of Climatology19, 471– 488 (1999)

Klink, K.: Climatological mean and interannual variance of united states surface wind speed, direction and velocity. International Journal of Climatology19, 471– 488 (1999)

1999
[17]

Technical report, NREL (January 2023)

Gagnon, P., Cowiestoll, B., Schwarz, M.: Cambium 2022 scenario descriptions and documentation. Technical report, NREL (January 2023). NREL/TP-6A40-84916. https://nrel.gov/publications.html

2022
[18]

Technical report, NREL (June 2021)

Ho, J., Becker, J., Brown, M., Brown, P., Chernyakhovskiy, I., Cohen, S., Cole, W., Corcoran, S., Eurek, K., Frazier, W., et al.: Regional energy deployment system (ReEDS) model documentation: Version 2020. Technical report, NREL (June 2021). NREL/TP-6A20-78195. https://nrel.g...

2020

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

Frontiers in Energy Research8(2020)

Zhang, J., Guerra-Fernandez, O.J., Eichman, J., Pellow, M.A.: Benefit analysis of long-duration energy storage in power systems with high renewable energy shares. Frontiers in Energy Research8(2020)

work page 2020

[2] [2]

In: 2023 North American Power Sympo- sium (NAPS), pp

Colombo, M., Kurtz, S.: Value of long-duration energy storage and oxy- combustion in renewables driven grids. In: 2023 North American Power Sympo- sium (NAPS), pp. 1–5 (2023). IEEE

work page 2023

[3] [3]

Nature Communications15(1), 9501 (2024)

Staadecker, M., Szinai, J., S´ anchez-P´ erez, P.A., Kurtz, S., Hidalgo-Gonzalez, P.: The value of long-duration energy storage under various grid conditions in a zero-emissions future. Nature Communications15(1), 9501 (2024)

work page 2024

[4] [4]

Energy Environ

Guerra, O.J., Zhang, J., Eichman, J., Denholm, P., Kurtz, J., Hodge, B.-M.: The value of seasonal energy storage technologies for the integration ofwind and solar power. Energy Environ. Sci.13, 1909–1922 (2020)

work page 2020

[5] [5]

Joule 5(8), 2077–2101 (2021)

Hunter, C.A., Penev, M.M., Reznicek, E.P., Eichman, J., Rustagi, N., Baldwin, S.F.: Techno-economic analysis of long-duration energy storage and flexible power generation technologies to support high-variable renewable energy grids. Joule 5(8), 2077–2101 (2021)

work page 2021

[6] [6]

Nature Energy6(5), 506–516 (2021)

Sepulveda, N.A., Jenkins, J.D., Edington, A., Mallapragada, D.S., Lester, R.K.: The design space for long-duration energy storage in decarbonized power systems. Nature Energy6(5), 506–516 (2021)

work page 2021

[7] [7]

IEEE Transactions on Energy Markets, Policy and Regulation (2024) 15

Silva, P., Moreira, A., Heleno, M., Marcato, A.L.M.: Boundary technology costs for economic viability of long-duration energy storage systems in california. IEEE Transactions on Energy Markets, Policy and Regulation (2024) 15

work page 2024

[8] [8]

Accessed on May 17, 2023

Gagnon, P., Cowiestoll, B., Schwarz, M.: Cambium 2022 Data. Accessed on May 17, 2023. https://scenarioviewer.nrel.gov

work page 2022