Let's Connect
Contact us
Thank you for reaching out, we will be in touch shortly!
Oops! Something went wrong while submitting the form.
As the energy transition advances and renewables make up growing portions of supply stacks, many assumptions traditionally used by production cost models no longer hold. Many fundamental forecasts suffer from a backwards-looking paradigm, in which gas always sets the marginal price and inexorably increasing gas prices keep the market price of power rising as well. In reality, weather will serve as the critical underlying driver for both demand and supply, driving price formation. Renewables and storage will increasingly become price setters in energy, capacity, and ancillary markets. Government policies and corporate clean energy goals will drive clean energy demand beyond what purely economic models might predict.
Amid these rapidly changing market dynamics, choosing the right energy forecast that captures the key driving forces of the energy transition has never been more critical. Energy market forecasts serve as critical foundations for developers, utilities, independent power producers (IPPs), energy project investors, asset owners and operators, and electric retailers. The right forecasts anticipate the changing market dynamics of the energy transition, helping to maximize investment returns, minimize risks, and optimize project siting, sizing, valuation, origination, and operation. Choosing the right forecast can be the difference between profit and loss.
Recent rises in capacity prices in PJM and elsewhere have generated significant pushback from load-serving entities, ratepayer advocates, politicians, and other stakeholders. However, the fast-approaching era of load growth and energy transition will necessarily lead to high capacity prices being the new normal.
The past 20 years have seen largely flat electricity demand growth across most of the country, driven by a combination of improving energy efficiency, increasing behind-the-meter (BTM) solar generation, offshoring of manufacturing, and an economic shift toward service industries with low energy intensity. While ERCOT and SPP serve as notable exceptions due to a combination of increased population, economic growth, and territory expansion in SPP, the remaining US markets have all seen minimal growth in peak demand since the early 2000s, as Figure 1 shows. MISO’s peak demand even remains close to where it was in the early 2000s despite a jump when the southern portion joined the system in 2013.
While load growth has been limited over the past two decades, a new era of demand growth is beginning. Data center buildout has emerged as a focus point for many recent demand forecast revisions, and will serve as an important driver for load growth. However, electrification of transportation, heating, and industrial demand were already expected to drive load growth on their own. As Figure 2 shows, industrial and transportation sectors are largely served by fossil fuels with minimal contributions from the electrical sector. An energy transition in which these end uses become served by electricity would necessitate substantial demand growth since the end-use energy in these two sectors alone is larger than the total current US electricity production.
In addition to the general resumption of load growth, many loads will experience significant differences in timing, flexibility, and reliability impacts relative to historical patterns. Because shortages will be unlikely to occur when renewables are producing at high levels, net load (total load minus renewable production) will become more important than gross load peaks. The availability of duration-limited resources such as batteries will also play a critical role in determining when reserves run the lowest. As peak conditions widen, duration-limited resources will struggle to further reduce peak demand.
Amid growing renewable and storage buildout, the degree of available load flexibility can play a significant role in managing demand peaks and reliability risks. Correlations between heat, cooling loads, and solar production mean that periods of system tightness will be relatively short during summers in solar-heavy grids, allowing pre-cooling of buildings to provide substantial load flexibility. Much EV charging can also be flexible, although this requires mechanisms to encourage and support smart and/or coordinated charging. As more EV charging and heating demand moves onto the grid, winter overnight periods will increasingly drive the largest reliability risks, exacerbated by cold-driven thermal power plant outages. The long duration of cold weather events makes pre-heating buildings less helpful than pre-cooling them during the summer. AI data center demand is largely inflexible, having a sufficiently high revenue value that the power demand is relatively inelastic even when wholesale power prices spike to their max.
New capacity resources will be needed that are specifically available during those conditions when the largest reliability risks occur. If those high-risk conditions occur relatively infrequently, then those capacity resources will be utilized infrequently and will need to recoup capital costs via capacity markets. Load-serving entities and their customers have been largely insulated from capacity costs for the past 20 years due to oversupply, but will face a rude awakening when capacity prices rise high enough to signal the need for new entry.
This paper will lay out the options for meeting capacity needs and explain why capacity prices must go high, discussing:
Capacity markets are designed to provide missing revenues that generation resources need to cover costs. When new generation has a higher efficiency (i.e. lower heat rate) than existing generation units that set prices, that new generation will earn some net revenue by having variable costs below the market price for power. New generation resources can earn additional revenues by providing ancillary services. Together, these two revenue sources reduce the ‘missing money’ needed from the capacity market to cover the capital expenditures required to build a project. However, in an energy transition, both the energy and ancillary revenues for new units will be eroded.
Because renewable generation has negative or near-zero variable costs, renewables shift the entire generation supply stack to the right and move prices down for a given load, as shown in Figure 3a. As renewable penetrations grow, prices first decline as increasingly efficient thermal generation becomes the price setter. Eventually prices approach and ›cross zero as renewable curtailment becomes an increasingly common price-setting condition. As Figure 3b shows, at high renewable penetrations average prices not only approach zero, but also exhibit very little variation. Prices are nearly always low at high renewable penetrations. With increasingly efficient price setters, thermal generation will struggle to earn margins in energy markets.
Ancillary service revenues also get eroded in an energy transition. As batteries get deployed as capacity resources, they are also perfectly suited for performing ancillary services because they are effectively always on with no limits to their ramp rates. Most batteries are also limited to roughly one cycle per day, so a four-hour battery often has around 1 hours per day with mostly nothing to do other than provide ancillary services. Because ancillary services markets are shallow relative to the quantity of storage that is expected to be built, batteries quickly become price setters, resulting in ancillary prices that reflect the opportunity costs for the energy arbitrage that batteries would otherwise perform. This results in near-zero ancillary prices when energy prices are not volatile and batteries have nothing else to do, and prices reflect the lost energy arbitrage revenues during volatile periods. As Figure 4 shows, these price dynamics have already emerged in markets where batteries have been deployed at scale. Ancillary prices started declining relative to energy prices in ERCOT and CAISO (the two markets with significant storage buildout by the end of 2024) as battery deployment grew. With storage depressing ancillary service prices, thermal generators cannot earn meaningful revenues from those ancillary services that can also be provided by batteries.
In addition to declining net revenues from energy and ancillary services, new thermal generation also faces declining capacity accreditation in most markets. As system operators improve their approaches to modeling reliability contributions for renewables and storage, they have also begun to re-evaluate the reliability contributions of thermal generation. Weather-correlated thermal plant and fuel supply outages during severe cold weather served as a major factor in blackouts during Winter Storm Uri in Texas in February 2021 and Winter Storm Elliot in PJM in December 2022, as shown in Figure 5. As a result, most thermal generators are now being accredited significantly below their nameplate capacity. In PJM, a natural gas combustion turbine receives just 63% accreditation without dual fuel, and 79% with dual fuel. A natural gas combined cycle unit receives 78% accreditation. Because diminished accreditation will reduce the revenue these resources receive from the capacity market, the prices they bid will need to be proportionally higher.
Prices will have to rise to offset declining revenue from the energy market, minimal revenue from ancillary services markets, and declining accreditation. At the same time, renewed demand for gas capacity is driving supply shortages in the gas turbine market, leading to installed costs rising by 2-3x. With 20+ year asset lifetimes and financial contracts, high capacity prices will have to persist through the entire financial life of the project to cover these escalated costs and reduced revenues.
As battery storage deployment increases, the ability for batteries to serve peak demand declines. As shown in Figure 6, demand peaks widen as they get shaved by peaking resources. This widening of the peak presents a challenge for batteries, which inherently have a finite duration. For example, a four-hour battery would have to operate at half of its nameplate capacity to reduce an eight-hour wide peak, effectively cutting its accreditation in half.
As Figure 7 shows, batteries are generally well-suited to provide capacity value during summer peaks, which have relatively short durations in systems with plentiful solar generation and a relatively narrow period of time between sunset and declining demand as people go to bed and the weather cools.
Winter peaks, however, present a much larger challenge for storage, as illustrated in Figure 8, which shows demand in PJM during Winter Storm Elliot. Each daily and sub-daily demand peak is six hours long, but once these peaks are shaved off, the storm represents a 30-hour peak. With heightened thermal outages during periods of severe cold, and with batteries getting depleted by long-duration peaks, these conditions will represent the periods of lowest reserves on the grid even before electrified heating creates a winter-peaking system. Systems where winter conditions drive the largest reliability risks will give particularly low accreditation to storage. In PJM, which has winter reliability risks, capacity accreditations are 55%, 65%, and 68% for four-hour, six-hour, and eight-hour storage, respectively.
Adding to the accreditation challenge for storage, most system operators are moving to a marginal effective load carrying capability (ELCC) accreditation structure rather than average ELCC. ELCC measures how much reliability benefit a given resource class provides, with ‘average ELCC’ accrediting capacity according to the reliability benefit of a given resource class and ‘marginal ELCC’ assessing the reliability benefit of the next incremental addition in the resource class.
Solar generation provides a straightforward illustration of the difference between average and marginal ELCC: while solar generation can improve reliability by generating during peak summer afternoon demand, once enough solar is built to move the net load peak into sunset hours, additional solar provides no further reliability benefit. At this point, solar has an average ELCC above zero but marginal ELCC equal to zero. In the case of storage, as storage begins shaving and widening the demand peak, each additional unit of storage must serve an increasingly wide peak. As a result, marginal ELCC decreases much faster than average ELCC, as Figure 9 shows.
Storage will see declining accreditation across markets, and the decline will be particularly precipitous in markets with winter reliability challenges and marginal ELCC accreditation. While each additional hour of battery duration increases capacity accreditation, each additional hour of arbitrage sees declining value, leaving more to be covered in capacity markets. As capacity accreditation declines, capacity prices must rise accordingly to cover the costs of building storage that are not offset by energy arbitrage.
Given its accreditation challenges, storage can be deployed with increasing duration to compensate. However, each added hour of duration incurs an incremental cost while realizing declining value for each additional hour of energy arbitrage. This declining value limits the benefit of added duration for storage with expensive cells, such as those associated with lithium-ion chemistries.
Alternatively, many long-duration energy storage (LDES) technologies aim to minimize the cost for additional duration. However, current options for LDES technologies typically have a low round-trip efficiency (RTE), which presents a generally underappreciated revenue challenge. In addition to the declining value of each hour of arbitrage, price spreads also need to be sufficiently high to offset the losses that occur in the round-trip charge-discharge process. When prices are relatively flat, the arbitrage revenue opportunity disappears as price spreads are insufficient to cover the losses. For example, a battery with 50% RTE (which is generally reflective of many current LDES options) must charge two hours for every hour of dispatch, requiring dispatch prices to be at least double the charge prices just to break even. As Figure 10 shows, this type of LDES may only be able to dispatch for a few hours per day in a market like ISO-NE with relatively flat prices (top). Even in CAISO when prices are at their most ‘ducky’ in the spring, several hours will still have an insufficient price spread to justify battery dispatch (bottom).
When prices become negative, low RTE can actually present an economic benefit as batteries receive payment to consume, effectively monetizing the increased production tax credits and renewable energy certificates (RECs) that must be generated to charge inefficient storage. These conditions provide a limited depth of opportunity, however, since significant renewable overbuild would be required to support these storage technologies at scale.
Due to the limitations of low RTE LDES, many of the available hours of duration will be unused throughout much of the year and unable to generate revenue. For multi-day storage options, many of the available hours of duration may only be utilized once every few years (or longer) during unusual and infrequent weather patterns. Because these hours of duration will be difficult to monetize in energy markets, LDES capital costs will have to be largely covered by capacity market revenues.
With declining energy and ancillary margins for thermal generation, declining capacity accreditation for short-duration storage, and limited arbitrage opportunities for LDES, each new capacity resource will have large revenue gaps that must be met through capacity markets (or other capacity revenue structures in regions that lack a capacity market) if it is to attract the new entrants that are needed to accommodate load demands.
Figure 11 shows an illustrative revenue stack in a market that has undergone an energy transition for currently inflated thermal generation due to supply constraints, future thermal generation after markets normalize, four-hour storage, and LDES, using estimates for costs and revenues that are consistent with Ascend Analytics’ forecasts. To reflect current uncertainty in the future of the clean energy tax credits under the Inflation Reduction Act, storage estimates are shown both with and without the investment tax credit (ITC). Revenue requirements that are not met by energy and ancillary markets will have to be met in the capacity market, with an additional increase to offset accreditation below 100%.
With recent increases in gas turbine costs, four-hour lithium-ion storage has much lower gross CONE and Net CONE required from the capacity market. After turbine costs normalize, lithium-ion storage should have a slightly lower gross CONE than thermal capacity, but will have a much lower Net CONE because of ITC as well as energy arbitrage revenues that significantly exceed the energy and ancillary margin that a thermal unit can earn.
LDES aspirational target costs are higher than those of four-hour lithium-ion batteries, and their low RTE significantly reduces their energy arbitrage opportunity, resulting in a much higher Net CONE than both new thermal entry and four-hour storage. All resources have significantly higher Net CONE than previous thermal entry, which was both lower in cost and able to offset capital costs via meaningful energy margin and ancillary revenues.
Due to accreditation differences between resources, however, Net CONE does not equate to the capacity prices each resource would need to support its entry. In markets with high storage deployment, even severely declining capacity accreditation for four-hour storage pushes the capacity price required for new storage to be on par with that required for new gas, as shown in Figure 12. Without the ITC, however, storage will become significantly more expensive than new gas capacity.
Uncertainties in capacity accreditation, capital costs, stranded asset risk, and state/federal clean energy policy could all easily tilt the resource economics in either direction between new gas and four-hour storage. Because the impact of accreditation on capacity prices is an inverse function, the sensitivity to storage accreditation becomes particularly acute at high storage penetrations and low accreditations. Low energy arbitrage revenues and higher installed cost for LDES push required capacity prices even higher, despite a high accreditation. Even new thermal generation will require capacity prices to be multiples of historical new thermal entry due to increased costs, reduced capacity accreditation, and declining energy and ancillary revenues. Because these assets have financial lives of 1 years, capacity prices will need to remain perpetually high in order to support the new entry.
While the numbers in Figure 11 and Figure 12 are intended to be illustrative rather than precise, they are indicative of how implied capacity prices vary across resource types, and why prices are likely to be set by new thermal generation in markets undergoing an energy transition. In states with policy mandates for clean energy, difficult choices must be made between:
Ascend expects capacity prices to rise into the vicinity of price caps tied to the cost of new entry in most US markets by the early 2030s, with the timing depending on current reserve margins, expected load growth, peak load conditions, planned retirements, and clean energy policies. In CAISO and SPP, which lack capacity markets, this rise will appear in the cost of bilateral capacity contracts. While MISO has an optional capacity market, Ascend expects the price rise to appear primarily in bilateral capacity agreements rather than in the capacity market, with most utilities preferring to contract with resources directly. In CAISO, where new gas is largely impossible to build, the high capacity price will appear through a need to contract with multiple resources to serve widening peak demand before eventually driving new renewable fuel capacity. ERCOT’s capacity incentives are likely to be spread among multiple mechanisms, including scarcity pricing, targeted ancillary services, and subsidized financing costs for thermal generation.
After PJM’s capacity market prices rose in the 202 /202 Base Residual Auction, backlash emerged from a variety of stakeholders, including load-serving entities, ratepayer advocates, regulators, and politicians. This backlash came despite PJM giving numerous warnings prior to this auction about the reliability risks of generator retirements in the face of years of low capacity prices. Even when capacity price rises should have been anticipated and recognized as necessary to incentivize new generation and discourage retirements, opposition and discomfort with high capacity prices was fierce. Opposition was so strong that PJM proposed a capacity market price cap and floor in early 202 , distorting the capacity market in a way that keeps both generators and consumers unhappy while undermining trust in the market and disincentivizing the new generation that PJM needs.
States with clean energy mandates will experience even greater opposition to high capacity prices, which will be needed to incentivize new clean generation but will also keep incumbent thermal generation online with windfall revenues. In the face of these high capacity prices, Ascend expects stakeholders to express growing interest in restructuring capacity markets such that needed capacity revenues can flow to new entry without paying extra revenues to incumbent generation, especially in states with clean energy policies.
Market restructuring could take several forms. States could subsidize clean capacity resources to reduce the amount of missing money that would need to be earned in capacity markets, thus enabling clean new entry to set lower capacity prices that displace incumbent thermal generation. ISOs could introduce new long-duration ancillary service products that would flow revenues specifically to long-duration resources and reduce the needed revenue from the capacity market. Capacity markets could also bifurcate into separate procurements for clean and fossil resources, decoupling the capacity prices needed to incentivize new clean resources from the prices needed to support incumbent fossil generation and allowing clean energy procurement to not cross-subsidize fossil generation. States may also consider leaving their ISO and re-regulating their utilities, which would enable different cost recovery or bilateral capacity procurement contracts for each generation asset. Other options, such as pay-as-bid market structures or multi-year fixed capacity payments for new entry, are unlikely to be approved by FERC.
High capacity prices are a necessary byproduct of an energy transition, are coming, and will persist. This will almost certainly drive a backlash among key stakeholders but is unavoidable: load growth requires new capacity, and high buildout of renewables and storage will erode offsetting revenues. To prepare for the implications of sustained high capacity prices, stakeholders and investors must evolve their investment and risk management strategies to reflect the new market dynamics of the energy transition.
Developers, investors and independent power producers seeking to build and operate generation resources will need to understand the timing and dynamics of the evolving outlook for capacity prices, capital costs, accreditation, energy markets, and ancillary markets. Load-serving entities and retailers will increasingly need to manage the financial risks from exposure to capacity prices rather than energy prices. Accordingly, capacity purchase agreements may either complement or replace power purchase agreements as a key tool for hedging market exposure risks and present a new opportunity for capacity resources to earn contracted revenue. Demand-side management and BTM options to reduce peak demand may become increasingly valuable and attractive to large load customers. Utilities and retailers will need to evolve their retail rates and efficiency programs to disincentivize and reduce consumption during peak periods to reduce their exposure to high capacity prices. States and market operators will need to anticipate consumer backlash and consider plans for responses and reforms. Ascend Analytics will be producing several targeted strategy briefs that expand on these topics, providing actionable guidance to stakeholders as this new paradigm arrives.
New market dynamics require new financial decision-making, and Ascend Analytics provides the analytics needed to make prudent financial decisions.
As the energy transition advances and renewables make up growing portions of supply stacks, many assumptions traditionally used by production cost models no longer hold. Many fundamental forecasts suffer from a backwards-looking paradigm, in which gas always sets the marginal price and inexorably increasing gas prices keep the market price of power rising as well. In reality, weather will serve as the critical underlying driver for both demand and supply, driving price formation. Renewables and storage will increasingly become price setters in energy, capacity, and ancillary markets. Government policies and corporate clean energy goals will drive clean energy demand beyond what purely economic models might predict.
Amid these rapidly changing market dynamics, choosing the right energy forecast that captures the key driving forces of the energy transition has never been more critical. Energy market forecasts serve as critical foundations for developers, utilities, independent power producers (IPPs), energy project investors, asset owners and operators, and electric retailers. The right forecasts anticipate the changing market dynamics of the energy transition, helping to maximize investment returns, minimize risks, and optimize project siting, sizing, valuation, origination, and operation. Choosing the right forecast can be the difference between profit and loss.
Recent rises in capacity prices in PJM and elsewhere have generated significant pushback from load-serving entities, ratepayer advocates, politicians, and other stakeholders. However, the fast-approaching era of load growth and energy transition will necessarily lead to high capacity prices being the new normal.
The past 20 years have seen largely flat electricity demand growth across most of the country, driven by a combination of improving energy efficiency, increasing behind-the-meter (BTM) solar generation, offshoring of manufacturing, and an economic shift toward service industries with low energy intensity. While ERCOT and SPP serve as notable exceptions due to a combination of increased population, economic growth, and territory expansion in SPP, the remaining US markets have all seen minimal growth in peak demand since the early 2000s, as Figure 1 shows. MISO’s peak demand even remains close to where it was in the early 2000s despite a jump when the southern portion joined the system in 2013.
While load growth has been limited over the past two decades, a new era of demand growth is beginning. Data center buildout has emerged as a focus point for many recent demand forecast revisions, and will serve as an important driver for load growth. However, electrification of transportation, heating, and industrial demand were already expected to drive load growth on their own. As Figure 2 shows, industrial and transportation sectors are largely served by fossil fuels with minimal contributions from the electrical sector. An energy transition in which these end uses become served by electricity would necessitate substantial demand growth since the end-use energy in these two sectors alone is larger than the total current US electricity production.
In addition to the general resumption of load growth, many loads will experience significant differences in timing, flexibility, and reliability impacts relative to historical patterns. Because shortages will be unlikely to occur when renewables are producing at high levels, net load (total load minus renewable production) will become more important than gross load peaks. The availability of duration-limited resources such as batteries will also play a critical role in determining when reserves run the lowest. As peak conditions widen, duration-limited resources will struggle to further reduce peak demand.
Amid growing renewable and storage buildout, the degree of available load flexibility can play a significant role in managing demand peaks and reliability risks. Correlations between heat, cooling loads, and solar production mean that periods of system tightness will be relatively short during summers in solar-heavy grids, allowing pre-cooling of buildings to provide substantial load flexibility. Much EV charging can also be flexible, although this requires mechanisms to encourage and support smart and/or coordinated charging. As more EV charging and heating demand moves onto the grid, winter overnight periods will increasingly drive the largest reliability risks, exacerbated by cold-driven thermal power plant outages. The long duration of cold weather events makes pre-heating buildings less helpful than pre-cooling them during the summer. AI data center demand is largely inflexible, having a sufficiently high revenue value that the power demand is relatively inelastic even when wholesale power prices spike to their max.
New capacity resources will be needed that are specifically available during those conditions when the largest reliability risks occur. If those high-risk conditions occur relatively infrequently, then those capacity resources will be utilized infrequently and will need to recoup capital costs via capacity markets. Load-serving entities and their customers have been largely insulated from capacity costs for the past 20 years due to oversupply, but will face a rude awakening when capacity prices rise high enough to signal the need for new entry.
This paper will lay out the options for meeting capacity needs and explain why capacity prices must go high, discussing:
Capacity markets are designed to provide missing revenues that generation resources need to cover costs. When new generation has a higher efficiency (i.e. lower heat rate) than existing generation units that set prices, that new generation will earn some net revenue by having variable costs below the market price for power. New generation resources can earn additional revenues by providing ancillary services. Together, these two revenue sources reduce the ‘missing money’ needed from the capacity market to cover the capital expenditures required to build a project. However, in an energy transition, both the energy and ancillary revenues for new units will be eroded.
Because renewable generation has negative or near-zero variable costs, renewables shift the entire generation supply stack to the right and move prices down for a given load, as shown in Figure 3a. As renewable penetrations grow, prices first decline as increasingly efficient thermal generation becomes the price setter. Eventually prices approach and ›cross zero as renewable curtailment becomes an increasingly common price-setting condition. As Figure 3b shows, at high renewable penetrations average prices not only approach zero, but also exhibit very little variation. Prices are nearly always low at high renewable penetrations. With increasingly efficient price setters, thermal generation will struggle to earn margins in energy markets.
Ancillary service revenues also get eroded in an energy transition. As batteries get deployed as capacity resources, they are also perfectly suited for performing ancillary services because they are effectively always on with no limits to their ramp rates. Most batteries are also limited to roughly one cycle per day, so a four-hour battery often has around 1 hours per day with mostly nothing to do other than provide ancillary services. Because ancillary services markets are shallow relative to the quantity of storage that is expected to be built, batteries quickly become price setters, resulting in ancillary prices that reflect the opportunity costs for the energy arbitrage that batteries would otherwise perform. This results in near-zero ancillary prices when energy prices are not volatile and batteries have nothing else to do, and prices reflect the lost energy arbitrage revenues during volatile periods. As Figure 4 shows, these price dynamics have already emerged in markets where batteries have been deployed at scale. Ancillary prices started declining relative to energy prices in ERCOT and CAISO (the two markets with significant storage buildout by the end of 2024) as battery deployment grew. With storage depressing ancillary service prices, thermal generators cannot earn meaningful revenues from those ancillary services that can also be provided by batteries.
In addition to declining net revenues from energy and ancillary services, new thermal generation also faces declining capacity accreditation in most markets. As system operators improve their approaches to modeling reliability contributions for renewables and storage, they have also begun to re-evaluate the reliability contributions of thermal generation. Weather-correlated thermal plant and fuel supply outages during severe cold weather served as a major factor in blackouts during Winter Storm Uri in Texas in February 2021 and Winter Storm Elliot in PJM in December 2022, as shown in Figure 5. As a result, most thermal generators are now being accredited significantly below their nameplate capacity. In PJM, a natural gas combustion turbine receives just 63% accreditation without dual fuel, and 79% with dual fuel. A natural gas combined cycle unit receives 78% accreditation. Because diminished accreditation will reduce the revenue these resources receive from the capacity market, the prices they bid will need to be proportionally higher.
Prices will have to rise to offset declining revenue from the energy market, minimal revenue from ancillary services markets, and declining accreditation. At the same time, renewed demand for gas capacity is driving supply shortages in the gas turbine market, leading to installed costs rising by 2-3x. With 20+ year asset lifetimes and financial contracts, high capacity prices will have to persist through the entire financial life of the project to cover these escalated costs and reduced revenues.
As battery storage deployment increases, the ability for batteries to serve peak demand declines. As shown in Figure 6, demand peaks widen as they get shaved by peaking resources. This widening of the peak presents a challenge for batteries, which inherently have a finite duration. For example, a four-hour battery would have to operate at half of its nameplate capacity to reduce an eight-hour wide peak, effectively cutting its accreditation in half.
As Figure 7 shows, batteries are generally well-suited to provide capacity value during summer peaks, which have relatively short durations in systems with plentiful solar generation and a relatively narrow period of time between sunset and declining demand as people go to bed and the weather cools.
Winter peaks, however, present a much larger challenge for storage, as illustrated in Figure 8, which shows demand in PJM during Winter Storm Elliot. Each daily and sub-daily demand peak is six hours long, but once these peaks are shaved off, the storm represents a 30-hour peak. With heightened thermal outages during periods of severe cold, and with batteries getting depleted by long-duration peaks, these conditions will represent the periods of lowest reserves on the grid even before electrified heating creates a winter-peaking system. Systems where winter conditions drive the largest reliability risks will give particularly low accreditation to storage. In PJM, which has winter reliability risks, capacity accreditations are 55%, 65%, and 68% for four-hour, six-hour, and eight-hour storage, respectively.
Adding to the accreditation challenge for storage, most system operators are moving to a marginal effective load carrying capability (ELCC) accreditation structure rather than average ELCC. ELCC measures how much reliability benefit a given resource class provides, with ‘average ELCC’ accrediting capacity according to the reliability benefit of a given resource class and ‘marginal ELCC’ assessing the reliability benefit of the next incremental addition in the resource class.
Solar generation provides a straightforward illustration of the difference between average and marginal ELCC: while solar generation can improve reliability by generating during peak summer afternoon demand, once enough solar is built to move the net load peak into sunset hours, additional solar provides no further reliability benefit. At this point, solar has an average ELCC above zero but marginal ELCC equal to zero. In the case of storage, as storage begins shaving and widening the demand peak, each additional unit of storage must serve an increasingly wide peak. As a result, marginal ELCC decreases much faster than average ELCC, as Figure 9 shows.
Storage will see declining accreditation across markets, and the decline will be particularly precipitous in markets with winter reliability challenges and marginal ELCC accreditation. While each additional hour of battery duration increases capacity accreditation, each additional hour of arbitrage sees declining value, leaving more to be covered in capacity markets. As capacity accreditation declines, capacity prices must rise accordingly to cover the costs of building storage that are not offset by energy arbitrage.
Given its accreditation challenges, storage can be deployed with increasing duration to compensate. However, each added hour of duration incurs an incremental cost while realizing declining value for each additional hour of energy arbitrage. This declining value limits the benefit of added duration for storage with expensive cells, such as those associated with lithium-ion chemistries.
Alternatively, many long-duration energy storage (LDES) technologies aim to minimize the cost for additional duration. However, current options for LDES technologies typically have a low round-trip efficiency (RTE), which presents a generally underappreciated revenue challenge. In addition to the declining value of each hour of arbitrage, price spreads also need to be sufficiently high to offset the losses that occur in the round-trip charge-discharge process. When prices are relatively flat, the arbitrage revenue opportunity disappears as price spreads are insufficient to cover the losses. For example, a battery with 50% RTE (which is generally reflective of many current LDES options) must charge two hours for every hour of dispatch, requiring dispatch prices to be at least double the charge prices just to break even. As Figure 10 shows, this type of LDES may only be able to dispatch for a few hours per day in a market like ISO-NE with relatively flat prices (top). Even in CAISO when prices are at their most ‘ducky’ in the spring, several hours will still have an insufficient price spread to justify battery dispatch (bottom).
When prices become negative, low RTE can actually present an economic benefit as batteries receive payment to consume, effectively monetizing the increased production tax credits and renewable energy certificates (RECs) that must be generated to charge inefficient storage. These conditions provide a limited depth of opportunity, however, since significant renewable overbuild would be required to support these storage technologies at scale.
Due to the limitations of low RTE LDES, many of the available hours of duration will be unused throughout much of the year and unable to generate revenue. For multi-day storage options, many of the available hours of duration may only be utilized once every few years (or longer) during unusual and infrequent weather patterns. Because these hours of duration will be difficult to monetize in energy markets, LDES capital costs will have to be largely covered by capacity market revenues.
With declining energy and ancillary margins for thermal generation, declining capacity accreditation for short-duration storage, and limited arbitrage opportunities for LDES, each new capacity resource will have large revenue gaps that must be met through capacity markets (or other capacity revenue structures in regions that lack a capacity market) if it is to attract the new entrants that are needed to accommodate load demands.
Figure 11 shows an illustrative revenue stack in a market that has undergone an energy transition for currently inflated thermal generation due to supply constraints, future thermal generation after markets normalize, four-hour storage, and LDES, using estimates for costs and revenues that are consistent with Ascend Analytics’ forecasts. To reflect current uncertainty in the future of the clean energy tax credits under the Inflation Reduction Act, storage estimates are shown both with and without the investment tax credit (ITC). Revenue requirements that are not met by energy and ancillary markets will have to be met in the capacity market, with an additional increase to offset accreditation below 100%.
With recent increases in gas turbine costs, four-hour lithium-ion storage has much lower gross CONE and Net CONE required from the capacity market. After turbine costs normalize, lithium-ion storage should have a slightly lower gross CONE than thermal capacity, but will have a much lower Net CONE because of ITC as well as energy arbitrage revenues that significantly exceed the energy and ancillary margin that a thermal unit can earn.
LDES aspirational target costs are higher than those of four-hour lithium-ion batteries, and their low RTE significantly reduces their energy arbitrage opportunity, resulting in a much higher Net CONE than both new thermal entry and four-hour storage. All resources have significantly higher Net CONE than previous thermal entry, which was both lower in cost and able to offset capital costs via meaningful energy margin and ancillary revenues.
Due to accreditation differences between resources, however, Net CONE does not equate to the capacity prices each resource would need to support its entry. In markets with high storage deployment, even severely declining capacity accreditation for four-hour storage pushes the capacity price required for new storage to be on par with that required for new gas, as shown in Figure 12. Without the ITC, however, storage will become significantly more expensive than new gas capacity.
Uncertainties in capacity accreditation, capital costs, stranded asset risk, and state/federal clean energy policy could all easily tilt the resource economics in either direction between new gas and four-hour storage. Because the impact of accreditation on capacity prices is an inverse function, the sensitivity to storage accreditation becomes particularly acute at high storage penetrations and low accreditations. Low energy arbitrage revenues and higher installed cost for LDES push required capacity prices even higher, despite a high accreditation. Even new thermal generation will require capacity prices to be multiples of historical new thermal entry due to increased costs, reduced capacity accreditation, and declining energy and ancillary revenues. Because these assets have financial lives of 1 years, capacity prices will need to remain perpetually high in order to support the new entry.
While the numbers in Figure 11 and Figure 12 are intended to be illustrative rather than precise, they are indicative of how implied capacity prices vary across resource types, and why prices are likely to be set by new thermal generation in markets undergoing an energy transition. In states with policy mandates for clean energy, difficult choices must be made between:
Ascend expects capacity prices to rise into the vicinity of price caps tied to the cost of new entry in most US markets by the early 2030s, with the timing depending on current reserve margins, expected load growth, peak load conditions, planned retirements, and clean energy policies. In CAISO and SPP, which lack capacity markets, this rise will appear in the cost of bilateral capacity contracts. While MISO has an optional capacity market, Ascend expects the price rise to appear primarily in bilateral capacity agreements rather than in the capacity market, with most utilities preferring to contract with resources directly. In CAISO, where new gas is largely impossible to build, the high capacity price will appear through a need to contract with multiple resources to serve widening peak demand before eventually driving new renewable fuel capacity. ERCOT’s capacity incentives are likely to be spread among multiple mechanisms, including scarcity pricing, targeted ancillary services, and subsidized financing costs for thermal generation.
After PJM’s capacity market prices rose in the 202 /202 Base Residual Auction, backlash emerged from a variety of stakeholders, including load-serving entities, ratepayer advocates, regulators, and politicians. This backlash came despite PJM giving numerous warnings prior to this auction about the reliability risks of generator retirements in the face of years of low capacity prices. Even when capacity price rises should have been anticipated and recognized as necessary to incentivize new generation and discourage retirements, opposition and discomfort with high capacity prices was fierce. Opposition was so strong that PJM proposed a capacity market price cap and floor in early 202 , distorting the capacity market in a way that keeps both generators and consumers unhappy while undermining trust in the market and disincentivizing the new generation that PJM needs.
States with clean energy mandates will experience even greater opposition to high capacity prices, which will be needed to incentivize new clean generation but will also keep incumbent thermal generation online with windfall revenues. In the face of these high capacity prices, Ascend expects stakeholders to express growing interest in restructuring capacity markets such that needed capacity revenues can flow to new entry without paying extra revenues to incumbent generation, especially in states with clean energy policies.
Market restructuring could take several forms. States could subsidize clean capacity resources to reduce the amount of missing money that would need to be earned in capacity markets, thus enabling clean new entry to set lower capacity prices that displace incumbent thermal generation. ISOs could introduce new long-duration ancillary service products that would flow revenues specifically to long-duration resources and reduce the needed revenue from the capacity market. Capacity markets could also bifurcate into separate procurements for clean and fossil resources, decoupling the capacity prices needed to incentivize new clean resources from the prices needed to support incumbent fossil generation and allowing clean energy procurement to not cross-subsidize fossil generation. States may also consider leaving their ISO and re-regulating their utilities, which would enable different cost recovery or bilateral capacity procurement contracts for each generation asset. Other options, such as pay-as-bid market structures or multi-year fixed capacity payments for new entry, are unlikely to be approved by FERC.
High capacity prices are a necessary byproduct of an energy transition, are coming, and will persist. This will almost certainly drive a backlash among key stakeholders but is unavoidable: load growth requires new capacity, and high buildout of renewables and storage will erode offsetting revenues. To prepare for the implications of sustained high capacity prices, stakeholders and investors must evolve their investment and risk management strategies to reflect the new market dynamics of the energy transition.
Developers, investors and independent power producers seeking to build and operate generation resources will need to understand the timing and dynamics of the evolving outlook for capacity prices, capital costs, accreditation, energy markets, and ancillary markets. Load-serving entities and retailers will increasingly need to manage the financial risks from exposure to capacity prices rather than energy prices. Accordingly, capacity purchase agreements may either complement or replace power purchase agreements as a key tool for hedging market exposure risks and present a new opportunity for capacity resources to earn contracted revenue. Demand-side management and BTM options to reduce peak demand may become increasingly valuable and attractive to large load customers. Utilities and retailers will need to evolve their retail rates and efficiency programs to disincentivize and reduce consumption during peak periods to reduce their exposure to high capacity prices. States and market operators will need to anticipate consumer backlash and consider plans for responses and reforms. Ascend Analytics will be producing several targeted strategy briefs that expand on these topics, providing actionable guidance to stakeholders as this new paradigm arrives.
New market dynamics require new financial decision-making, and Ascend Analytics provides the analytics needed to make prudent financial decisions.
Products
Consulting
©2024 Ascend Analytics. All rights reserved. Privacy Policy.
.png)
-1-ercot%20image.png)
%20(1).avif)
.svg){kind=small}
.webp){kind=small}