Energy storage: the right technology depends on the system in which it will compete

Storage investors must evaluate how technological configurations compete, what benefit they create for the electricity system, and how much they can earn. Their profitability also depends on how those conditions will change during the asset’s life. This Perspective is for storage investors: developers, generators evaluating hybridization, funds, and financiers. Before investing, they should establish with evidence whether capital is going to the configuration that will produce the expected returns.

0 · Structure of this note

Section 1 sets out the decisions before investing. Sections 2 and 3 examine system need and competitiveness by duration. Sections 4 and 5 compare alternatives and localized costs. Section 6 distinguishes system benefit, capturable revenue, and cost. Section 7 examines changes during the asset’s life. Section 8 presents the Capital Projects Decision System. Verified facts, calculations, estimates with assumptions, and hypotheses are kept separate.

1 · Investment begins by deciding whether storage, and in which configuration, can produce the expected return

Before choosing a supplier, contract, or financing structure, storage investors make 2 decisions that condition all the others.  The first is whether storage is an attractive investment compared with the other solutions available to the system. The second is the configuration: technology, power, energy, node, in-service date, and operating approach. An asset can have proven technology, an attractive contract, and flawless construction. Even if it meets all its specifications, it can fall short of the expected return if those 2 decisions were wrong.

Chile’s public data show why the question matters.  The example of a hypothetical 4-hour battery at Crucero, a 220 kV bus in the north, examines 2 remuneration components: arbitrage and capacity. Using prices from October 2025 to September 2026, a strategy that charges during the day and discharges at night is calculated. Its arbitrage margin reaches USD 105/kW per year. Capacity remuneration provides a reference of 69. Annual capital and operating cost is estimated at USD 92–109/kW (section 6) [D; E]. A year earlier, the same margin was 137. It fell 23% in 12 months [D: 105 / 137 = 0.77], while battery capacity in the system grew (section 7) [D]. These are references calculated using past prices and perfect foresight.  Profitability remains to be demonstrated. Investors face a rapidly changing opportunity and must answer 3 questions beyond the tender documents. How much revenue the asset retains when already announced capacity comes online. What remuneration sustains it throughout its life. Whether a 4-hour battery is the appropriate configuration at that node.

Answering those questions requires understanding the system in which the asset will compete.  The next 6 sections cover what needs to be understood, and the last describes how it becomes a decision-making capability.

2 · Competitive solutions depend on system need; published hours are the starting point for an independent evaluation

A storage asset’s revenue depends on the service the system pays for. Services differ in 3 dimensions: response speed, delivery duration, and electrical attributes.

Need	Response time	Sustained delivery	Source
Fast frequency response	250 milliseconds	Up to 15 minutes	ERCOT, Texas operator [V]
Frequency regulation	5 seconds	—	ERCOT [V]
Contingency reserves	10 minutes; or 30 minutes	2 hours; or 4 hours	ERCOT [V]
Peaking capacity	—	2–6 hours	NREL, United States national laboratory [V]
Capacity and intraday energy shifting	—	4–12 hours	NREL [V]
Interday and multiday backup	—	10–36 hours; 36–160 hours	United States Department of Energy [V]
Seasonal	—	More than 160 hours	Department of Energy [V]
System strength	Not measured in time: grid-forming capability, voltage and frequency control, fault current, black start	Not measured in hours	Chile’s Coordinador Eléctrico Nacional [V]

The horizons range from minutes to more than 160 hours and correspond to different services.  The thresholds are United States classifications, with approximate boundaries. What transfers is the breadth of the interval. System strength is specified and verified separately. Capacity in MWh, by itself, leaves those electrical attributes unproven. An asset designed to provide a service listed in the table earns revenue from that service, and its business case depends on the system continuing to need it.

The 3 recent instruments in the region publish duration references, minimum requirements, or delivery windows, with different scopes.  Mexico presents a reference for receiving projects, Colombia defines contractual products, and the Dominican Republic sets requirements for a renewable tender. The published hours are identified in the documents examined, but the technical derivation behind them is not.

Country	Instrument	How duration appears
Mexico	Convocatoria de la Secretaría de Energía para proyectos estratégicos de generación y almacenamiento (Secretaría de Energía call for strategic generation and storage projects), May 2026: a mechanism for receiving applications and granting permits	A 3-hour reference for systems "not associated" with a power plant; paragraph 18.2 clarifies that the table is "a reference" and that "different project proposals may be received" [V]. The binding expansion plan, PLADESE 2025–2039, considers 4 hours for battery systems included in its power plant installation program [V]
Colombia	Resolución 40208 del Ministerio de Minas y Energía (Ministry of Mines and Energy Resolution 40208), April 2026	Product 3 combines blocks of 08:00–18:00 and 18:00–22:00; product 4 corresponds to 18:00–22:00. Both allow plants under the battery storage conditions of article 9, and those contractual windows do not set a minimum battery duration [V]. The explanatory memorandum runs its expansion model "with the 4 hourly demand blocks considered for the auction": the window enters as an input [V]
Dominican Republic	Resolución SIE-092-2025-LCE de la Superintendencia de Electricidad (Superintendencia de Electricidad Resolution SIE-092-2025-LCE)	"Minimum storage capacity of four (04) hours throughout the term of the Contract" for batteries at renewable plants in a tender of up to 600 MW, in a list including synthetic inertia, black start, and grid-forming capability [V]

Investors must verify compliance with the published reference or obligation as part of the entry evaluation. They must test profitability with independent evidence on need, competition, and revenue.  In the North region, the Mexican call publishes a reference of 245 MW and 3 hours and allows different proposals [V]. The plan’s table for that region publishes 141 MW for 2025–2030 [V]. The 2 figures may measure different universes, and comparing them requires reconciling dates, categories, and regional coverage. Neither the call nor chapter 3 of the plan explains which study produced the 3 hours [V].

In Mexico, operating solar portfolios are already consolidating, and each faces the decision to add storage. In 2026, Zora Power, a Riverstone company, acquired 4 operating solar plants totaling 162.4 MWp, bringing its Mexican portfolio to 461.2 MWp [V, according to Riverstone; 3 of the 4 plants confirmed by independent sources]. The sources report no storage at those plants. For that owner, the question in this section is the next decision: which system need the battery would serve at each node, and with how many hours [H].

Duration can be derived from the system.  The Coordinador Eléctrico Nacional, Chile’s system operator, published a study in August 2023 that left duration as a variable. On a simplified 22-bus network, it first identified interregional transmission. It then jointly optimized generation and storage. The resulting optimal duration was 6–8 hours, with 1,000–4,000 MW in the north [V]. That result applies to Chile, to that simplified network, and to its expansion assumptions.  The duration of each specific asset is determined separately. Investors can derive the need from a system representation to establish what will sustain their asset.

3 · Duration changes the economic competitiveness of storage technologies

Technologies allocate their costs differently between power and energy.  In lithium iron phosphate (LFP) batteries, cost is primarily associated with energy. At 1,000 MW, they cost USD 394/kWh installed at 2 hours and 299 at 100 hours. In pumped hydro storage, cost is primarily associated with power. At 1,000 MW, it falls from USD 503/kWh at 4 hours to 32 at 100 hours [V, PNNL, United States national laboratory].

The levelized cost of storage per MWh discharged (LCOS) reflects that difference [V, PNNL v2024, 2023 values in the United States, USD/MWh]:

Technology, 1,000 MW	4 h	10 h	24 h	100 h
LFP lithium	174	168	296	954
Pumped hydro storage	186	97	97	122
Compressed air energy storage	130	91	93	106
Thermal storage	192	145	190	359
Hydrogen (2021 values)	—	346	350	373

At the same scale for all technologies, pumped hydro storage and compressed air cost 42–46% less than lithium at 10 hours [D: 97 / 168 = 0.58; 91 / 168 = 0.54]. At 24 hours, lithium costs 1.76 times as much as at 10 [D: 296 / 168]. At 4 hours, lithium and pumped hydro storage cost almost the same, USD 174/MWh and USD 186/MWh. Compressed air has the lowest cost in the table, 130, provided a site and operation compatible with that reference are available [V]. At 24 hours, lithium costs more than pumped hydro storage, compressed air, and thermal storage [D: 296 versus 97, 93, and 190].

The table combines technology, operating assumptions, and site, so it does not say which technology wins in a specific project. It does say that duration changes the conditions under which each technology competes.  Competitiveness also depends on power, site, utilization, and cost structure, which sections 4 and 5 incorporate. For investors, the question is whether their proposed technology is competitive at the duration the system will pay for throughout the asset’s life. If that duration changes, the cost ranking may change with it, and another technology may offer the same service at a lower price in the next tender.

In some markets, regulatory eligibility defines the competitive field before costs are compared.  For products 3 and 4 of Resolución 40208, Colombia makes participation conditional on plants incorporating battery storage [V]. That condition defines the competitive field for those contracts.

4 · Storage competes to meet system needs and capture revenue in specific markets and contracts

A storage asset faces 2 distinct forms of competition, and its business case must measure both.

Competition	Who competes	What it determines for investors
For system need	Storage, gas generation, transmission, demand management, redispatch, accepting curtailment	Whether the system will need storage and how much. A line or a peaking turbine can reduce the need, or the price, on which the asset depends
For revenue	Resources eligible for the same product, market, or contract: other storage, generation, hybrid configurations, and demand, where the rules allow	How much the asset can earn in each market or contract

Options that are not even storage compete for the same need:

Need	Option A	B	C	D
Evening peak	Solar with storage	Peaking gas turbine	Demand management	Pumped hydro storage, if a site is available
Congestion at a node	Storage at the node	Transmission line	Redispatch	Demand management
Solar curtailment	Short-duration storage	Long-duration storage	Transmission	Accepting curtailment
System strength	Storage with grid-forming capability	Synchronous condenser	Synchronous generation	Combination

Accepting curtailment is also a configuration, and it limits arbitrage revenue.  In the Chilean study, storage reduces renewable energy curtailment by up to 40% compared with the case without storage. The efficient curtailment level remains close to 15% [V]. In that model, eliminating all curtailment costs more than it is worth.

The components change together.  A new transmission line changes the appropriate node, and the node changes duration, technology, and revenue. A battery earning revenue from congestion may lose part of its business case if a line that resolves it enters service. The effect depends on the node and the direction of flows [H: illustrative example]. The investment unit is therefore the complete configuration, and the question "how many battery hours?" comes late.

5 · Country, site, and utilization change the cost and economics of each alternative

Applying a "country factor" to an international reference would combine what transfers with the technology and what does not.  A configuration’s cost is separated into 4 layers:

Layer	What it contains	How it is treated
Technology	Efficiency, degradation, cycle life, and much of the equipment price	Transfers: the international reference is retained
Country	Construction, freight, tariffs and taxes, permits, cost of capital	Replaced with publicly available country-specific data
Site	Interconnection, civil works, geology, caverns or reservoirs, network reinforcements	Replaced with project data
System	Hourly charging price, cycles, displaced resource, congestion, services	Determines utilization and revenue; produced by system simulation

Country affects the return.  Chile’s Comisión Nacional de Energía (CNE) publishes USD 722/kW for 4-hour batteries [V, May 2026]. The country risk premium is 1.10% in Chile and 2.85% in Colombia, compared with 0.23% in the United States [V, Damodaran, January 2026]. In an illustrative sensitivity, adding that difference to a real rate of 7% raises the annual capital charge of a 15-year asset by 6–17% [E: the cost of capital of each project is estimated separately, with its debt and equity structure]. The published plans of Mexico and Colombia contain no reference investment cost for storage [V].

Utilization also matters, and the technical data sheet does not determine it.  International references assume high utilization. Lazard assumes 1 daily cycle for 350 days per year [V]. PNNL assumes the maximum permitted by duration and warranty [V]. With fewer cycles, fixed cost is spread over fewer MWh. Return also depends on the margin in those hours and payments for availability and other services.

6 · The benefit a configuration creates for the system and the revenue investors can earn are 2 distinct evaluations

A configuration is evaluated using 3 quantities.

Quantity	What it measures
System benefit	The net reduction in system operating and unserved-energy costs against an explicit counterfactual, including charging, losses, and the configuration’s incremental operation, before its investment and replacements
Capturable revenue	What investors can earn from energy, capacity, services, and contracts, with mutually compatible rules and constraints
Total cost	Investment, operation, and replacements for each alternative over a common horizon; taxes and financing enter only the financial evaluation, and each item is recorded once

These support 2 connected evaluations, and each cost is counted once.  The economic evaluation compares each alternative against the same counterfactual, with equivalent service, reliability, and horizon. If benefit is already net of operation, only investment and replacements are deducted. It determines whether the configuration benefits the system. The financial evaluation distinguishes 2 cash flows. Project cash flow is calculated before debt service. Equity cash flow is calculated after financing disbursements and payments. Each is discounted at a consistent rate to determine whether it meets the required return. Debt repayment capacity is tested separately. A configuration may benefit the system and be unfinanceable under current remuneration, or be financeable and provide little benefit. The financial evaluation must represent remuneration rules, their effective period, and the project’s contractual protection, and test those subject to revision through sensitivity analysis.

Benefit depends on what the MWh displaces.  United States heat rates and levelized fuel prices from PLADESE 2025–2039 illustrate gross fuel cost. For a gas-fired combined cycle, the result is USD 38/MWh. Applying the heat rate of petroleum-fueled internal combustion engines to diesel yields USD 364/MWh. The ratio is 9.57 times [D: 7.548 MMBtu/MWh × USD 5.04/MMBtu versus 10.512 × 34.62; United States Energy Information Administration, table 8.2; PLADESE, table 3.7; the technology mapping is an assumption]. These are gross fuel cost references at full load; a specific plant requires its own efficiency and fuel. This is the gross cost of displaced fuel.  Estimating the net benefit of storage requires accounting for charging energy and its losses. Other incremental costs and applicable additional benefits must also be included. Each item is counted once. The difference shows that the same MWh carries a different weight depending on the system in which it is discharged.

With public data, the 2 evaluations of a 4-hour battery connected at Crucero reach this point:

Component	Public data	Result
Investment cost	CNE, May 2026 cost report: USD 722/kW, with an interval of 616–879	USD 722/kW [V]
Annual capital and operating cost	Fixed investment of USD 722/kW, 15-year life, and real cost of capital of 8–10%; operation and maintenance at 1–2% of investment per year, according to CNE	USD 92–109/kW per year [E]. Project cash flow reconciles the items included in the USD 722/kW and adds only those excluded, such as connection or replacements; degradation and taxes are represented in their cash flows
Capacity revenue	Capacity node price at Crucero 220 kV: USD 9.18/kW-month (CNE, July 2026). The transitional rule in Decreto 70 (Decree 70) recognizes 98% of a 4-hour battery’s maximum power as initial capacity [V]. In the 2024 adequacy balance, final revision of June 2026, final adequacy capacity was 64.3% of preliminary adequacy capacity in the subsystem that includes the north, the same for all plants [V]	USD 69/kW per year [D: 9.18 × 12 × 0.98 × 0.643], as a reference: assumes zero forced unavailability and that the 2024 factor remains [E]; project settlement is determined separately
Arbitrage margin	Final actual marginal cost at Crucero 220 kV, every 15 minutes, over 2 years: October 2024 to September 2025 and October 2025 to September 2026 (Coordinador Eléctrico Nacional). Strategy: charge between 8:00 and 17:00 and discharge from 17:00 until the end of the reported day, including the additional hour at the clock change, up to 1 daily cycle, with 89% round-trip efficiency and charging and discharging power of 1 kW per kW	Up to USD 137/kW per year in the first year and 105 in the second [D: maximum for that strategy with perfect foresight and energy balance; discharge sales minus charging purchases; no degradation, no unavailability, and no charges on charging energy; assumes the battery does not move prices]
System benefit	Operator’s study, 2023: efficient interval of 1,000–4,000 MW in the north, with 6–8 hours, for 2026–2030, additional to capacity declared under construction in March 2023 [V]	Unquantified with public data: requires simulating the system with and without the battery, against a defined alternative. This is the calculation performed by the Decision System in section 8

What these data allow investment evaluators to state:

• There is a remuneration opportunity, and it shrank in 1 year.  The illustrative sum of the capacity payment and arbitrage margin is USD 174/kW per year using the second-period references and 206 using the first [D: 69.4 + 105.1 and 69.4 + 136.5]. Charging energy is already deducted from arbitrage. Other costs remain to be included, and it must be verified that the battery can simultaneously meet the conditions for the capacity payment and arbitrage dispatch. Investment cash flow is evaluated separately

• System benefit is calculated separately.  An explicit alternative must be defined and charging costs and losses considered. The effect on operation and the contribution to reliability must also be considered. That evaluation determines the relative weight of energy, capacity, and other services

• The 2 evaluations come from different instruments.  Historical prices allow the financial evaluation to begin and the expansion study provides a reference for the economic evaluation, but the public evidence examined does not reconcile them

7 · Profitability changes during the asset’s life, and the decision is updated with it

The business case that justifies committing capital describes the system on the decision date, and the asset operates for 15 years or more in a changing system.  At Crucero, 3 changes are already visible in public data.

The arbitrage margin has already compressed.  Using the same calculation, the strategy’s margin at Crucero fell from USD 137/kW per year between October 2024 and September 2025 to 105 between October 2025 and September 2026: the margin fell in 12 months [D], with the underlying calculation available for inspection. The average price of discharged energy fell from USD 95/MWh to USD 76/MWh, while the price of charging energy rose from 1.5 to 3.8 [D]. In that period, system storage power rose from about 1.2 to 3.2 GW during 2025, and in April 2026 the operator reported a portfolio of around 4.6 GW expected to come online between March and December 2026 [V]. These data cannot isolate how much of the decline is attributable to batteries: the average price across all hours also fell, from USD 51/MWh to USD 43/MWh [D], and fuels, hydrology, and demand may influence that [H]. Charging increases demand in low-price hours and discharging increases supply in high-price hours, so the direction is expected; determining how much revenue a battery entering today will retain requires simulating the system with the incoming capacity.

Competition for revenue has already been announced.  Projects reported in operation, testing, and construction in Chile totaled 9.94 GW and 42.7 GWh as of October 2025, a power-weighted average of 4.3 hours [V; D: 42,729 / 9,941]; 5.44 GW of that capacity is in the Antofagasta region, where Crucero is located [V]. Those figures combine different stages of progress; the overlap in node, timing, and date with a specific project is evaluated using the updated portfolio. The 2023 study identified 1,000–4,000 MW additional to what was then under construction: the included capacities and cutoff dates are not comparable, so the difference requires updating the evaluation and does not support a conclusion of overinvestment.

Capacity additions may reduce the capacity payment, and the transitional regime ends before the asset’s assumed life.  With peak demand and other conditions held constant, new recognized capacity coming online may reduce the factor that converts preliminary adequacy capacity into final adequacy capacity: in a sensitivity exercise by Systep, a Chilean electricity consultancy, on the 2024 calculation, it falls to about 50% with 5 GW of additional batteries, which would take Crucero’s capacity reference from USD 69 to some USD 54/kW per year [H: third-party exercise; D: 9.18 × 12 × 0.98 × 0.502]. The transitional regime for calculating initial capacity lasts 10 years from July 2024 [V; D: until June 2034], against the assumed 15-year life: a battery entering in October 2026 would have some 7.75 years under the transitional regime. The regulation’s permanent methodology applies thereafter and determines remuneration for those years. Duration criteria also differ: the transitional rule reaches 98% recognition at 4 hours and 100% at 5 [V], while the study identified 6–8 as efficient. A recognition table is not a profitability function, and the study’s interval comes from system optimization rather than evaluating a specific asset. The difference warrants evaluation before duration is chosen.

Timing is also part of the return.  The storage project awarded in 2021 in Colombia’s Atlántico moved its in-service date 3 times, about 1,100 days in total [V; D]. Based on published information, that delay cannot be attributed to the configuration. It is the kind of event that requires asking again whether the asset, with its new date, still earns revenue from the need for which it was awarded and whether its business case still holds.

An evaluation performed only once, when bidding, leaves investors without a rule for knowing when their business case has ceased to hold.  This piece’s governance proposal is to set, before committing capital, the indicators that trigger review — a cost that moves more than expected, a network delay, a capacity addition, or a remuneration change — with their value, responsible person, and planned action, case by case. Once the asset is built, review compares future cash flows with avoidable costs: retiring the asset does not recover capital already invested.

8 · InsightIA turns configuration selection and review into a Capital Projects Decision System

A good development team already runs sensitivities, compares scenarios, and documents assumptions. We propose adding a capability that keeps the asset’s physical, operating, and economic decisions connected. Who makes each decision and with what evidence must be recorded. Investors must decide again when the system changes.  InsightIA builds that capability within the Capital Projects function, as a Decision System. The asset’s financial model and the system simulation are assessed together.

Figure 1 · The Decision System for a storage investment: 2 decision cycles and the rights governing them.  In cycle 1, alternative configurations — storage and other solutions — and system scenarios are simulated together, hour by hour and node by node. The simulation produces 3 quantities. The economic evaluation uses system benefit. The financial evaluation uses capturable revenue, together with the financeability test. The 2 evaluations use total cost, with its economic and financial treatment. If the configuration fails to meet the criteria, it is reconfigured or discarded. If it benefits the system and requires different remuneration, a viable contractual or regulatory mechanism is evaluated and the calculation is repeated. Only a configuration meeting the economic and financial criteria is selected for capital commitment. Cycle 2 runs during the asset’s life. When a material change crosses a threshold set in advance, the responsible person decides whether to continue, modify, replace, sell, or retire. They decide using future cash flows and avoidable costs, including closure costs, contractual commitments, and sale or salvage value. Cost is reconciled between the 2 evaluations: resources in the economic evaluation and applicable outflows in the financial evaluation. What is already included in net operating benefit or the arbitrage margin is counted once. Acceptance criteria, the discount rate, and risk constraints are set before comparison. Evidence, responsible person, decision, and date are recorded at each review. The top row names those decision rights. Each project assigns a responsible person, approver, evidence, and effective period.

The 5 questions the Decision System answers for investors:

Question	What is delivered
Is storage preferable to other solutions for the same need?	The need derived from an hourly, nodal representation of the system, and comparison with generation, transmission, demand management, and curtailment
Which technology and configuration, compared with the other resources earning revenue in the same market?	Comparison under the same system representation, with technologies, durations, nodes, dates, and hybrid configurations
How much does each configuration cost here?	Localized investment, operation, and replacement cost, with explicit degradation; taxes and financing represented in the corresponding cash flows
What benefit does it create, how much does it earn, and what return does it produce under different futures?	The 3 quantities and the 2 evaluations; cash flows, net present value (NPV), internal rate of return (IRR), debt coverage, and exposure to adverse scenarios
What event triggers a review of the configuration and investment?	Indicators and thresholds set before committing capital, with a responsible person and planned action, monitored during development, contracting, construction, and operation

This piece argues that the decisive risk in a storage investment is established before construction: whether storage is appropriate, with which technology, how many hours, and at which node.  Avoiding it requires comparing complete configurations, representing their operation within the system, separately measuring the benefit they create and the revenue they earn, and deciding again when conditions change. The Decision System is designed to turn that evaluation into a permanent capability for investors: what to build, where, when, with which technology, and under what conditions to commit capital.

The question for every investment committee is direct: can it substantiate today, with traceable assumptions and adverse scenarios, that the configuration to which it will commit capital meets return and risk criteria against the alternatives evaluated, in the system in which it will compete?

Sources

1. Coordinador Eléctrico Nacional · Estudio de Almacenamiento de Energía en el SEN (Energy Storage Study in the SEN), August 2023: https://www.bienesnacionales.cl/wp-content/uploads/2023/12/Coordinador-Electrico-Estudio-de-almacenamiento-de-energia-en-el-SEN-Agosto-2023.pdf

2. Secretaría de Energía de México · Convocatoria para la atención de proyectos estratégicos de generación y almacenamiento de energía eléctrica (Call for strategic electricity generation and storage projects), DOF, May 15, 2026, paragraph 18.2: https://sidof.segob.gob.mx/notas/docFuente/5787666, and its third amendment (SIDOF 5797703); PLADESE 2025–2039, chapter 3, section 3.4.1 and tables 3.5 and 3.7: https://www.gob.mx/cms/uploads/attachment/file/1042893/PLADESE_cap_3.pdf

3. Ministerio de Minas y Energía de Colombia · Resolución 40208, Diario Oficial 53.471, April 24, 2026, articles 7 and 9: https://gestornormativo.creg.gov.co/gestor/entorno/docs/resolucion_minminas_40208_2026.htm, and its memoria justificativa (explanatory memorandum); CREG · documento 901 225 (document 901 225)

4. CNE de Chile · Fijación de Precios de Nudo Promedio, Informe Técnico Definitivo (Average Node Price Setting, Final Technical Report), July 2026, table 17: https://www.cne.cl/wp-content/uploads/2026/06/2026.06-ITD-PNP-julio-2026.pdf; Biblioteca del Congreso Nacional · Decreto 70, published on June 5, 2024, first transitional article: https://www.bcn.cl/leychile/navegar?idNorma=1204012; Coordinador Eléctrico Nacional · Reliquidación del Balance de Potencia de Suficiencia 2024 (Resettlement of the 2024 Adequacy Capacity Balance), final revision R04, June 2026, file InformeSUF24_DEF_v4.3, sheet Resumen: https://www.coordinador.cl/wp-content/uploads/2026/06/Reliquidacion-PSUF-2024_rel_def_R04.zip; Coordinador Eléctrico Nacional · Costo Marginal Real (Actual Marginal Cost), monthly REAL-DEF files from October 2024 to September 2026: https://www.coordinador.cl/costos-marginales/; Coordinador Eléctrico Nacional · Informe de Monitoreo de la Competencia en el Mercado Eléctrico 2025 (2025 Electricity Market Competition Monitoring Report), April 2026: https://www.coordinador.cl/wp-content/uploads/2026/04/Informe-Monitoreo-2025.pdf; Generadoras de Chile · Boletín de octubre de 2025 (October 2025 Bulletin): https://generadoras.cl/wp-content/uploads/2025/12/BoletinGeneradorasdeChile_30112025.pdf; Systep · Reporte mensual del sector eléctrico (Monthly Electricity Sector Report), May 2025, figure E.3: https://systep.cl/reporte-mensual-de-mayo-2025/

5. Superintendencia de Electricidad de República Dominicana · Resolución SIE-092-2025-LCE, section 4.3: https://sie.gob.do/wp-content/uploads/2025/08/SIE-092-2025-LCE.pdf

6. ERCOT · Ancillary Services Study, September 2024: https://www.ercot.com/files/docs/2024/10/07/ERCOT-Ancillary-Services-Study-Final-White-Paper.pdf

7. United States Department of Energy · Pathways to Commercial Liftoff: Long Duration Energy Storage, page 13: https://www.energy.gov/documents/pathways-commercial-liftoff-long-duration-energy-storage-opportunities

8. NREL · Denholm and Frazier, Storage Futures Study: Four Phases Framework and Modeling, NREL/PR-5C00-80366, 2021, page 5 of the PDF: https://docs.nlr.gov/docs/fy21osti/80366.pdf

9. PNNL · Energy Storage Cost and Performance Database v2024, sheet Database: https://www.pnnl.gov/projects/esgc-cost-performance/download-reports

10. Lazard · Levelized Cost of Energy+, July 2026, page 38: https://www.lazard.com/media/kcfconhf/lazards-lcoeplus_vf.pdf

11. CNE de Chile · Informe de Costos de Tecnologías de Generación y Almacenamiento (Generation and Storage Technology Cost Report), May 2026: https://www.cne.cl/wp-content/uploads/2026/06/ICTGA-2026.pdf

12. Damodaran · Country Default Spreads and Risk Premiums, January 5, 2026: https://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/ctryprem.html

13. United States Energy Information Administration · Electric Power Annual, table 8.2: https://www.eia.gov/electricity/annual/html/epa_08_02.html

14. Riverstone · LinkedIn post by a Riverstone director on Zora Power’s 4 acquisitions, November 2026; Memoranda · “Carambola de tres bandas”, October 19, 2026: https://memorandamx.substack.com/p/carambola-de-tres-bandas; Latin Lawyer · “Mijares and Ritch Mueller seal BlackRock-Riverstone solar deal”, September 22, 2026: https://latinlawyer.com/article/mijares-and-ritch-mueller-seal-blackrock-riverstone-solar-deal; DLA Piper · October 26, 2026: https://www.dlapiper.com/en/news/2026/10/dla-piper-advises-banco-sabadell-on-financing-of-riverstone-acquisition; Power Technology · Pollux Solar PV Park: https://www.power-technology.com/data-insights/power-plant-profile-pollux-solar-pv-park-mexico/

Evidence markers: [V] verified in the source · [D] derived with the calculation shown · [E] estimate with stated assumptions · [H] hypothesis or illustrative example. The supporting evidence book identifies the source, calculation, and status of each claim. The arbitrage calculation is reproduced using the procedure, file manifest, and input checks described there.