- We expect total capital investment in North American battery storage to exceed $3 billion annually starting in 2024. For context, capital investment in U.S. wind power averaged only $2 billion annually between 2000 and 2006.
- The firming of renewable power (making it less interruptible) is a primary use case across the country as investment tax credit (ITC)-advantaged solar plus storage is becoming a cost-effective resource option to meet capacity requirements in most regions.
- Solar plus storage is now a disruptive factor in the industry because of cost reductions, not so much for storage but in solar capital and operating costs.
- We present our estimates of the levelized cost of energy for solar photovoltaic (PV) projects and also for solar plus storage investments.
- Investments in co-located solar PV and storage provide significant growth opportunities for utilities and power companies that have the advantage of established interconnections.
- Companies that have substantial conventional generation are being forced to quickly pivot strategies. While we have heard about declining coal-fired generation for some time, co-located storage with wind and solar is the decisive nudge.
We recollect a conversation back in 2006 with the then CEO of a large diversified utility. In response to a question, he offered "I don't want to sound spiritual because I will likely not see it in my lifetime, but solar power and storage would be the potential gamechanger whenever it comes. Duration is the hurdle it will have to overcome but when that comes is tricky." We distinctly recollect the conversation because of what he said next, "I ain't often right, but I've never been wrong."
We never ignore anyone who can quote the Grateful Dead. Also, he was right. Solar plus storage technology has advanced much since then. Technologically, it still has a duration problem, but now it is disrupting conventional power.
We Expect Storage Deployments To Increase Substantially
As of December 2020, the U.S. had about 25 gigawatts (GW) of power in energy storage, a number that is still de minimus relative to the renewables already on the grid. Moreover, over 90% of this is pumped hydroelectric storage (23 GW), a tally that is not likely to grow much at all, due to high permitting and construction costs. In contrast, we estimate front-of-the-meter battery storage installations at about 1.3 GW in 2020. Overall, the U.S. has installed just over 1 GW of battery storage in the past five years, with a third of it in the Pennsylvania-Jersey-Maryland (PJM) Interconnection and a quarter in the California Independent System Operator (CAISO) region.
While growth has been robust over the past five years because the deployment of global installed battery capacity has increased at a 25% compound annual growth rate, battery storage solutions in the power sector have flattered to deceive. Each year, mass scale deployment appeared to be on the horizon only to be pushed three years into the future. However, we now believe that battery storage solutions in the U.S. are at an inflection point and that large scale deployment is imminent (see chart 1).
S&P Global Platts Analytics, an affiliate of S&P Global Ratings, expects the U.S. storage market to increase nearly nine times in 2020-2023, fueled largely by advanced battery energy storage, with cumulative deployment approaching 10 GW by 2024 (see chart 2). This is arguably an optimistic projection given installations have typically lagged expectations and will likely be influenced by the pandemic in the near term. We note that the Energy Information Administration (EIA) and SNL Energy project lower aggregate numbers through 2023 of the order of about 6.5 GW while Wood Mackenzie Power and the Energy Storage Association recently published a joint report putting their estimates at 7.5 GW (26.5 GWh) by 2024. Regardless, we think the secular trend is undeniable; the upward growth curve will likely get more convex even if the pace turns out somewhat slower than currently expected. In contrast, other forms of storage have seen their global installed capacity hold relatively flat over this same time period, demonstrating the attractiveness of modular and distributed solutions that advanced battery energy storage offers.
In a January 2018 commentary, we had focused on the economics of stand-alone battery operations in a power generation application (peaking unit) (see "Going With The Flow: How Battery Storage Economics Are Changing Power Consumption," published Jan. 11, 2018). We had also compared the less conspicuous, but more economically attractive, applications like commercial and industrial (C&I) peak demand shaving against the high-profile, yet less economic, applications like residential rooftop solar.
In this follow-up, we update our calculations for stand-alone battery peaker applications based on current cost-curve economies. Our focus now shifts to the economics of utility scale solar plus storage solutions, which we expect to be a gamechanger for the power industry.
What's Driving The Deployment Trend?
Battery installations in the past five years ranged between 150 megawatt (MW) and 225 MW a year, with the largest capacities located in California, the Electric Reliability Council of Texas (ERCOT), the PJM states, New York, and Hawaii. However, even before the pandemic struck, the near-term forecasts for battery installations had declined as uncertainty increased from the trade dispute with China and as some projects faced delays. We suspect there will be some uncertainty around the impact of the pandemic on 2021 deployments because states and the U.S. Congress are still discussing possible economic measures that could benefit storage and renewables more broadly. However, we expect annual deployments to accelerate starting in 2022 because several large projects already contracted will be commissioned.
A combination of reasons will likely spur battery deployments over the next three years, in our view. These include:
- State targets and support schemes, including utility procurement through integrated resource planning processes,
- Growth in storage capacity in the interconnection queue, and
- The economics of storage and the potential for revenue stacking.
State mandates and targets
Similar to renewable portfolio standards that ushered in massive renewable deployments over the past decade, key drivers for the expected surge in battery deployments are the surging mandates and targets. Increasingly, more states--now seven--have implemented an aggregate 11 GW of battery deployments (see table 2). These incentives are particularly important in regions with limited current economic potential for storage, either due to cheaper incumbent technologies or a limited need for new peaking capacity.
|State||Mandate or target||Remark|
|California||1.325 GW by 2024||Per AB 2514; AB 2868 (2016) authorized an additional 0.5 GW of behind-the-meter deployment|
|New York||1.5 GW by 2025 and 3.0 GW by 2030||AB 6571; The PSC authorized $310 million of market acceleration funding for storage, on top of the $40 million announced earlier for pairing storage with solar PV. NYSERDA launched in May 2019 its Market Acceleration Bridge Incentive Program with $150 million in incentives for systems over 5 MW and $130 million retail incentive for smaller systems.|
|New Jersey||0.6 GW by 2021 and 2.0 GW by 2030||AB 3723; No specific implementation inacted. State likely to miss its 2021 target.|
|Massachussetts||1 GWh by 2025||HB 4857; Clean Peak Standard|
|Virginia||3.1 GW by 2036||SB 851. The bill split the deployment into 400 MW for Phase I utilities (Appalachian Power) and 2.7 GW for Phase II utilities (Dominion). About 10% of the energy storage projects should be behind the meter, and about 35% owned and operated by third parties.|
|Nevada||1 GW by 2026||SB 204 and Order 44671. In March 2020, the PUC adopted the rulemaking implementing the law with a procurement target starting at 100 MW in 2020, 200 MW by 2022, and growing by 200 MW every two years until reaching 1 GW by 2030. While the regulation does not indicate a target storage duration, projects previously procured by NV Energy were all of four-hour duration.|
|Oregon||10 MWh by 2020||HB 2193; 5 MWh each for PacifiCorp and Portland General Electric|
|Arizona||3 GW by 2030||Proposed, not final.|
Storage in the interconnection queue
Anecdotal evidence of a potential tsunami comes from the rapid increase of battery projects in the interconnection queue (see chart 2). This capacity is largely speculative as developers line up projects and reserve slots in the queues in anticipation of utility requests for proposals. Projects with an executed interconnection agreement have a higher likelihood of eventually commissioning. For comparison, the battery capacity in regional transmission organization (RTO)/ISO interconnection queues had increased to 85 GW by June 2020 from 14 GW in June 2018. The three RTO/ISOs with the highest increase are CAISO, PJM, and ERCOT.
In CAISO, the speculative nature of the interconnection queue is underscored by the decline in battery capacity in 2020. Many counties have seen stand-alone or solar plus battery contracts signed with community choice aggregators (CCAs) and one factor that results in a decline in capacity is when a developer, with a project in the queue, fails to secure contracts with those CCAs. Several developers plan to install batteries at existing wind or solar PV assets to benefit from the existing facilities and interconnection capacity while deriving additional revenue for their assets through participation in ancillary service markets.
In PJM, the median capacity MWs of stand-alone projects ranges between 25 MW and 50 MW. Further, only about 55% of stand-alone projects have requested interconnection as a capacity resource, which would allow these units to participate in PJM's capacity market. In contrast, almost all co-located solar PV plus storage solutions have applied as capacity resources. However, the current 10-hour duration limitation to qualify in the capacity market limits the potential revenue for stand-alone batteries. Shorter duration batteries would have to de-rate their capacity to qualify for the capacity market, receiving the capacity market clearing price only for a fraction of their capacity.
Finally, in ERCOT, most battery projects are located in counties that expect large renewable solar PV projects. Moreover, most projects co-located with solar PV installations have a battery capacity ranging from 50%-150% of the PV nameplate capacity. This indicates that developers are seeking to contract with utilities for the provision of peaking capacity.
The potential for revenue stacking
Perhaps the most significant development since we last published on storage in 2018 has been the continually declining cost curve. We believe lower capital costs for storage over the next few years, together with declining PV solar installed capital costs, as well as lower operating and maintenance (O&M) costs, will result in solar plus storage significantly disrupting conventional generation.
But before we get into the cost dynamics, we'll revisit some basics:
Storage has many uses. In contrast to generating units, energy storage supports a wide range of applications, many of which do not directly correspond to delivering energy to the grid. Storage can be deployed both on the electric grid and at the consumer's home. Its economics are shaped by the customer type, location, customer load shape, rate structure, and nature of the application. For instance, storage can have many different applications, like energy arbitrage, rate arbitrage, generation capacity, or transmission/distribution capacity deferral, etc. It is also uniquely flexible in its ability to change its dispatch to serve different needs over the course of a year or even an hour.
Ancillary services. Battery application in the ancillary services market was enabled by the Federal Energy Regulatory Commission (FERC) through Orders 890 and 755. These FERC orders effectively allowed non-generator resources to compete with generating units by selling ancillary services as a stand-alone product. These applications typically require storage of 15 minutes to an hour.
Ancillary services are increasingly required as electricity grids experience continuous imbalances between power generation and consumption as millions of devices are turned on and off in an uncorrelated way. These imbalances cause electricity frequencies to deviate, which can hurt sensitive equipment and, if left unchecked and allowed to become too large, even affect the stability of the electric grid. To avoid disruptions and blackouts, power has to flow within the grid at quasi-constant frequency (or voltage). Historically, thermal plants would provide that by quickly varying the kinetic energy of their generation units. Now, storage systems are particularly well suited for frequency regulation applications because of their rapid response time and ability to charge and discharge efficiently.
So, in the near term, it has all been about short-duration opportunities (frequency regulation and spinning reserve markets), which continue to be key drivers behind fully or quasi-merchant projects. Of these, the most economic opportunity remains the frequency response market. In PJM, this is Regulation D, i.e. the ancillary services market. While almost all batteries in PJM have been commissioned to provide frequency regulation, it is interesting to see that even in CAISO most batteries currently commit most of their capacity to the provision of frequency regulation. In New York ISO and ISO-New England as well, batteries have been commissioned mostly to perform frequency regulation. While this use for batteries is limited, we see compensation as attractive and see the bulk of near-term storage development activity oriented toward earning revenue in this $100 per kilowatt-year (/kW-year) market from short duration 15-minute to 1-hour battery products. Near term, this service will likely support early deployments in more markets--such as in the Southwest Power Pool and the Midwest Independent System Operator (MISO). However, ancillary services, in general, are shallow markets, prone to rapid saturation. Yet, ancillary services can provide some initial growth in volumes that can help drive down battery prices.
Front-of-the-meter. Large-scale installations used by utilities for a variety of on-grid applications.
Behind-the-meter. PV solar systems that do not provide power back to the utility through net metering.
Feed-in-tariffs (FIT). Guaranteed price adders (the premium that regulators allow over market prices) for renewable power. FIT is less employed in the U.S., which has typically provided incentives through subsidies like the ITC, the production tax credit, and bonus depreciation etc.
DC/AC ratio. Because the PV array rarely produces power to its standard testing condition capacity, it is common practice and often economically advantageous to size the inverter to be less than the PV array. This ratio of PV to inverter power is measured as the DC/AC ratio. A healthy design will typically have a DC/AC ratio of 1.25. The reason for this is that only about 1% of the energy produced by the PV array throughout its life will be at a power above 80% capacity. Thus a 10 kilowatt (kW) PV array paired with an 8.5 kW AC inverter would have an ideal DC/AC ratio with minimal power loss.
Clipping losses. When the DC/AC ratio of a solar power system is too high, the likelihood of the PV array producing more power than the inverter can handle increases. If the PV array outputs more energy than the inverter can handle, the inverter will reduce the voltage of the electricity and drop the power output. This loss in power is known as "clipping." For example, a DC/AC ratio of 1.5 will likely see clipping losses of 2%-5%.
Firm generation. Currently, firm generation capacity refers to electric energy resources that can produce energy on demand during extended periods of time during which wind and solar energy are not available.
Effective load carrying capability (ELCC): The ELCC of a generating resource is a measurement of that resource's ability to produce energy when the grid is most likely to experience electricity shortfalls. ELCC is typically expressed as a percentage of a resource's capacity, for example, a 100 MW solar plant that has an ELCC of 30% could make a 30 MW contribution toward reliability requirements.
Peaking capacity. Revenue stacking remains key to achieving a positive return on investments with long-duration batteries. New deployments will have to tap into new revenue streams (price arbitrage, spinning reserve, capacity, and transmission and distribution deferrals) as we have seen happen in the U.K and Germany (see chart 4). Just as one would expect of nascent technologies, we've seen storage solutions enter the market as 15-minute products that have now grown to economical peak-shifting four-hour applications. The total value from storage is a blend of several benefit streams from different sources and might not always be driven purely by economics.
We believe the provision of peaking capacity is the next big use because utilities are starting to recognize the value of storage in their integrated resource planning, particularly in the vertically integrated markets. Actually, we expect most of the large projects recently contracted by utilities to provide peaking capacity (with or without solar PV). Generally, the first utilities contracting for storage were located in regions with high solar PV insolation but now it is increasingly across all RTO regions. However, most projects will have a four-hour duration and not come online before 2022.
Low reserve margins in ERCOT will likely drive the first wave of merchant projects. ERCOT has the highest revenue potential due to a scarcity pricing mechanism that increases power pricing at times of low reserve margin. For instance, ERCOT saw more than 100 hours in 2019 when energy prices exceeded $1,000 per megawatt-hour (/MWh). As reserve and energy are co-optimized, spinning reserve prices reached very high levels, making it an attractive service for batteries to provide. ERCOT procures between 2,500 and 3,000 MW of spinning reserve and about 250 MW of upward and downward regulation, on average.
Stand-alone battery economics
Battery costs have declined enough to replace gas-fired peaking generation in certain markets such as Hawaii and California. With continually declining capital costs, the economics of a four-hour grid-level battery storage unit is within the neighborhood of a gas peaker plant. But with no moving parts, the battery bank needs less maintenance (yet, battery asset life is shorter). It also requires no fuel, making its long-term cost of operation more stable and predictable. A battery bank can respond to power demand almost instantly--less than a millisecond rather than several minutes. Where a gas turbine is strictly an energy generator, a battery bank can also store surplus energy. Finally, a battery bank is scalable: more units can be added as needed without dramatic cost increases. We discuss how these factors factor into our ratings assessment in the commentary Credit FAQ: What Factors Power A Rating On A Utility-Scale Battery Storage Project?, published Jan. 12, 2018
Since we have the highest visibility of Lithium-ion economics, we'll take that as an example. First, when we think in terms of capital costs for batteries, units are in kilowatt-hours (kWh) of operation--i.e., it's in $/kWh because we expect batteries to be duration products for peak shifting (or peak shaving) solutions. Stated differently, since a battery is used to store energy, the capital cost of a battery is expressed in dollars per unit of energy stored, or $/kWh. To be clear, this is different from energy prices, where the cost of electric energy is expressed in cents/kWh.
Utility scale battery economics are currently at about the $240/kWh price point, or about $950/kWh for a battery peaker plant that provides a four-hour peak shift. Costs for the balance of the system are about $300/kW for equipment like inverters/rectifiers, transformers and power control equipment, and various safety equipment. So a utility scale battery would currently cost about $1,250-$1,300/kW (240 times 4 plus $300-$350)--we think those costs are comparable with the cost of building a natural gas-fired peaker plant in California. We present a simplified calculation (see table 2) of the installed cost of a battery park that could replace a peaking unit.
|Cost Economics Of A Battery Unit That Could Substitute For A Gas-Fired Peaker|
|Size of the battery system deployed (MW) (A)||100||Peaking capacity that the battery installation can deliver instantaneously. This is the typical size of the LM6000 installed in California.|
|Peak-shifting application hours (B)||4||Typical shift required; however, batteries can now deliver six hours too.|
|Total battery energy (MWhrs) = (A X B)--('C)||400|
|2021 battery cost that we have assumed ($/kW-hour) (D)-- Inc.301 tariffs with tax equity||240||Utility scale battery costs have been declining. But without ITC still at $310/kWh. Residential battery costs (Tesla Powerwall etc.) are still much higher at $800/kWh.|
|AC to grid connection one-way efficiency loss (E)||4.0%||Just an inevitable loss but getting more efficient. Historically, this was as high as 7% one-way.|
|Two-way aggregate loss (F) = (2 X E)||8.0%||At charging and while discharging.|
|Maximum depth of discharge (G)||90.0%||Li-ion batteries are typically recharged before they fully discharge. However, they have low memory effect, i.e. unlike nickel-cadmium batteries, they do not lose their maximum energy capacity if they are repeatedly recharged after being only partially discharged.|
|Minimum depth of discharge (H)||10.0%|
|Battery capacity use (I) =(G-H)||80.0%||Only 80% of the battery is effectively used between recharges.|
|Combined losses (J)||23.20%||Combines 4% charging inefficiency loss and 20% of capacity that is not used between battery charge and recharge. It just means that this 400 MWh battery actually needs to be sized to 520 MWh to be able to deliver 400 MWh (400/(0.8*0.96).|
|Cost of the battery ($/kW) (K) = (C X D)/(1-J)||1,250.0||Cost over life of the battery. S&P Platts' Analytics' 2021 cost estimate (with ITC and 301 tariffs) is about $1,233/kW. To be clear, stand-alone batteries do not attract ITC. This is illustrative.|
|Cost of the 100 MW battery (Mil. $) (L) = (K X A) X 1000||125.00||Conversion to MW from kW and converting to Mil. $.|
|PV of cash flow stream for $1 in future over 15 years (M).||$8.83||NPV (rate, cash flow stream of 15 1s). Discounted at 7.5% required return over 15-year battery life. We've decided, somewhat randomly, that 7.5% is the required return. Please use your assumption.|
|How capital costs will be recovered (N)= (1/M)||11.33%||Inverse of M. This number indicates that $1 invested today requires about 8.8 cents in revenue over the next 15 years for full capital recovery of the $1, discounted at a hurdle rate of 7.5%.|
|Annualized O&M costs through life||$1.48||Combines O&M @ 1.75% of capex every year; includes mainanence capex every four years.|
|Annual revenue (O)= (L X N) + O&M Costs||15.64|
|Required capacity price ($ /MW-day) (P) = O/(A X 365) X 1000000||428.54||Merchant markets do not support a capacity price of $ 425 /MW-day. This needs to be supported through a offtake contract.|
|Converted to $ /kW-Mo. = P X 365/(1000*12)||13.0||An installation without ITCs will be about $16/kW-mo. As battery cost will be about $310/KWhr|
|Now let's assume additional energy margins at $15/MWh||I.e. The battery is able to glean a $25/MWh 'charge' spread, arbitraging between on and off peak power.|
|Battery output in MWh (same as C)||400||From C above|
|Battery input in MWh (Q)||543.5||This time we factor a two-way efficiency loss (i.e. 8%) as the battery will be actively called and not merely charged and on stand-by.|
|Power price at on-peak delivery ($/MWh) ('R)||$45||The net spread is $25 /MWh, which is the arbitrage opportunity.|
|Power price at off-peak charging (S)||$20|
|Charge' spread ($) (T) = (C X R)- (Q X S)||$7,130|
|# of charging cycles per year (U)||250||Let's assume no charging requirements on weekends|
|Aggegate annual 'charge' spread (Mil. $) (V) = (U XT) X 1000000||1.78|
|Charge spread offset on Total revenue (W) =(O-V)||13.86|
|Required capacity price ($ /MW-day) (X)- W/(A X 365) X 100000||379.7||Still not competitive in merchant markets but competitive when battery costs approach $200/kWh. S&P Platts estimates that by 2025 if ITCs continue.|
|Converted to $ /kW-mo.||11.6||At a battery cost of $200/kWh, this will be $9.5 /kW-mo.|
From market sources, we are given to understand that a $11-$12/kW-month fixed all-in tolling payment (i.e. capacity, energy, etc.) is roughly in line with contracts with stand-alone battery projects. We think batteries as capacity and energy resources are still some time away from being economically viable in a merchant market compared with both ancillary services and C&I peak-shaving applications. However, as a natural transition until they become competitive as peaking applications, such investments are increasingly being considered as part of a utility's rate base or supported by a contract in fulfillment of state initiatives.
Stand-alone solar PV economics
Due in part to the capital cost declines, the economics of pairing storage with solar to provide a near firm power product are now favorable in many regions. Utility resource plans are including co-located storage as a new resource option; we have seen increases in announced storage projects as a result.
However, before we talk about co-located solar plus storage solutions, we'll discuss how solar PV economics have recently trended. We calculated the levelized cost of energy (LCOE) of a competitive solar unit--one whose capital and operating costs are in the 25th percentile. Our base case assumptions (see table 3) for overnight capital costs and capacity factors are $1,050 per kW and 28%, respectively--with a cautious note that capacity factors do vary significantly by region (i.e. irradiance/insolation). Capacity factors are lower for projects in the PJM and MISO, where 25% is more appropriate, for instance.
|Base Case Assumptions|
|Solar PV system||Supporting rationale|
|Capital costs ($/kW)||1050||NREL is using $950/kW. S&P Platts estimates $1,045/kW for 2021, declining to $850/kW by 2024. In its Nov. 2002 report, Berkley Labs reports that the median installed price of projects that came online in 2019 fell to $1.4/WAC ($1.2/WDC), down 20% from 2018 and down by more than 70% from 2010.|
|Capacity factor (%)||28%||Average appears to be 25%-28% now. Berkeley Labs' upper end is 35% but reports that average capacity factors range from 17% in the least-sunny regions to 30% where it is sunniest.|
|Asset life (years)||30||NREL assumes 25-40-year range. Berkeley Lab notes that project life increased to 32.4 years in 2019 from 21.5 years in 2007 . We have assumed last 15-year merchant.|
|Operating expenses ($/kW-yr)||18||From Berkeley Labs Nov. 2020 update: lifetime OpEx estimates have declined to an average of ~$17/kWDC-yr in 2019 from an average of ~$35/kWDC-yr for projects built in 2007. Across 13 sources, the range in average lifetime OpEx for projects built in 2019 is broad, from $13-$25/kWDC-yr. O&M costs--one component of OpEx--have declined precipitously in recent years, to $5-$8/kWDC-yr in many cases. Property taxes and land lease costs are highly variable across sites, but on average are--together--of similar magnitude. Other OpEx line items include security, insurance, and asset management.|
|ITC (%)||26%||According to the stimulus bill, the Solar ITC extension will keep 26% ITC through 2022 (vs current 2020), stepping down to 22% through 2023 (vs current 2021), and fall to 10%/0% in 2024 for large/small scale solar projects.|
|Contract price ($/MWh)||29||LevelTen PPA Index reports that lowest P(25) prices increased to $31/MWh from $24/MWh by Q4, 2020.|
|Contract life (years)||15||Average solar contract|
|PPA and O&M cost escalator||2%||Typical escalator seen in contracts|
|Captal costs ($/kWh)||240||Availability of batteries from a supply chain perspective has been a critical bottleneck in 2020 and that could persist through 2022.|
|PPA Premium price adder for co-located solar||55%||Effectively a firming premium. It depends on battery/PV capacity. Typically, the adder is about 15%-20% for battery/PV up to 25%; 40% for battery /PV of 50%, and 50%-60% for battery/PV of 75%-100%. If premium becomes too high, there are risks of economic curtailement, unless the product is seen as firm.|
|Incremental O&M costs of co-located solar system ($/kW-year)||2||Typically, battery annual O&M is about 2% of its installed cost.|
Our base case LCOE estimate for stand-alone solar PV, including an ITC, is about $23.00/MWh. A no-ITC case increases our estimates by about $5.50/MWh to $8.50/MWh, depending on how low (or high) the base case O&M and capital costs assumptions are (see table 4). These LCOEs appear low because they represent the LCOE of the lowest cost units, also because of the prevailing financing environment. We have presented a matrix that provides a snapshot of how LCOEs would move if operating costs, or utilization, were higher or lower than base case expectations. For example, in urban areas near major cities, property taxes will make operating costs higher than our assumptions.
|Levelized Cost Of Energy ($/MWhr)|
|Capital costs-$1,050/KW; ITC-26%|
|Operating costs ($/kW-year)||15||16||17||18||20||24||28||30|
|Operating costs-$18/KW-year; ITC 26%|
|Capital cost ($/kW)||850||900||950||1000||1050||1200||1300||1350|
|*Bolded region is current LCOE band for P(25) solar projects. Source: S&P Global Ratings.|
Our results are consistent with LevelTen Energy's P25 power purchase agreement (PPA) price index (see chart 5). This PPA index reports the prices that solar project developers have offered for standard, flat, as-generated (i.e. unit-contingent) power from their projects. The P25 refers to the most competitive 25th percentile PPA offer prices. The P25 of levelized PPA offer prices is $25.10 and $26.70 in CAISO and ERCOT, respectively, even as the aggregate prices were higher because of higher offers in the PJM Interconnection and MISO.
We note that the PPA index prices were lower still and have risen recently. Permitting and interconnection are becoming increasingly costly and difficult in many U.S. markets, which creates a bottleneck, constraining supply and putting upward pressure on prices. As the best locations get scooped up, identifying and securing permits and interconnection positions--in the absence of broad-based reform in these two areas--becomes more challenging for new projects. In addition, economically competitive projects that had secured permits, and safe-harbored equipment for the full ITC, found contracts with offtakers, leaving higher-priced projects available in the market.
Battery Levelized Cost Of Energy (LCOE): All You Need To Know For Now
Levelized cost of energy. LCOE is the net present value of delivering a unit of energy over the lifetime of the asset. However, in contrast to generation units, energy storage supports a number of applications, many of which do not correspond to delivering energy to the grid (peak-shifting etc). Consequently, there is greater inconsistency in calculating storage LCOEs.
Four inputs primarily determine the LCOE of a solar plus battery storage system. The vast majority of the costs are for the solar array and the battery pack. The third is the battery size (capacity) to reach the desired use and the fourth is the local solar production factor.
Battery size. The size of the battery and hours of storage installed would depend on the customer's load profile. The capital cost of a battery increases in proportion to the number of hours of storage. A developer will chose the size of the battery to offset the few hours of the day in which the load is significantly higher than the rest of the day. Most load profiles have a four- to six-hour peak period when demand is greater than during other hours.
A key limitation of LCOE is that it ignores the time effects associated with matching electricity production to demand. This happens at two levels: 1) Dispatchability, the ability of a generating system to come online, go offline, or ramp up or down quickly, as demand swings; and 2) The extent to which the availability profile matches or conflicts with the market demand profile. Another limitation of LCOE is that it does not consider indirect costs of generation. These include environmental externalities or grid upgrade requirements.
Peak Shifting Economics
The problem with solar power is that it is interruptible. Moreover, all solar units produce at the same time. This means that as solar installed capacity goes up, its effective load carrying capability (ELCC) goes down. This is because the ELCC values depend on when electricity shortages are most likely to occur, and the type and quantity of resources already on the grid have a significant impact on the likely timing of electricity shortfalls. As the grid adds more solar plants that are all producing electricity at the same time, it reaches a point where all those solar plants are preventing daytime reliability issues so effectively that the remaining reliability challenges move into the evening hours when solar can't produce. At this point, adding more solar does very little to prevent electricity shortages, and its ELCC often falls to 0%. This is why co-location of storage with solar has gained importance.
We see co-located solar plus storage projects of varying configurations. Broadly, the ratio between the solar PV capacity and the battery capacity can provide an indication of the use of the battery. While the ratio varies widely by developers, lately, we have seen this increase, suggesting that battery usage is slowly moving into longer duration applications like peak shifting and peaking capacity. Ratios above 50% indicate that batteries are used to provide some peaking capacity (by shifting midday generation to evening peaks). For example, National Grid USA commissioned a 6 MW/48 MWh battery in Nantucket, Mass. in October 2019, to be used to meet summer peak demand and defer the need for an additional underwater supply cable to the island. A 20 MW/100 MWh battery at the Lawai Solar and Energy Storage Project, commissioned by The AES Corp. (BBB-/Stable/--) in December 2018 in Hawaii, provides dispatchable solar PV. Similarly, Vistra Corp. has developed the world's largest battery energy storage project, the 400-MW/1,600-MWh Moss Landing energy storage facility, in California.
Note that utility-scale solar-plus-storage installations can take advantage of the ITC as long as the energy storage system is at least 75% charged by the on-site solar unit. Our assumptions (see table 3) for co-located batteries with solar are similar to those for a stand-alone unit but with additional ITC benefits from the storage component to offset initial capital expenditures. Also, the storage is typically priced either through a $/kW-month fixed tolling payment basis, as with stand-alone storage, or as a $/MWh adder to the total project PPA. However, as a result of the ITC benefit, the fixed toll is typically lower for co-located storage than for a stand-alone storage unit.
Based on our analysis, we arrived at a LCOE value that adds about $10-$11/MWh to the LCOE ranges of stand-alone solar, or a premium of about 50% to stand-alone solar levelized costs. However, an LCOE analysis in the context of storage is less relevant. What matters more in storage installation is the type of unit that the solar plus storage unit is displacing. In regions that are served by a preponderance of base load units and peaking units, without any meaningful intermediate units in the power supply stack, solar plus storage could provide a valuable firm power product. For instance, in a location like Hawaii, offtakers will be willing to pay a higher power price if the unit that is being mitigated is a high-cost oil-fired peaker. Thus, it is better to evaluate the internal rate of returns (IRRs) for co-located storage projects. We chose to model a solar plus storage unit that employs a $/MWh storage premium adder of 55%. (see storage premium adder rationale in table 3).
|Solar Plus Storage Leveraged Internal Rate Of Returns|
|Solar PPA ($/MWh)|
|Capital costs ($/kW)|
|*ITC-26%; 55% Storage PPA Premium Adder. Source: S&P Global Ratings.|
We believe that continued cost reductions support further penetration, with current all-in utility-scale storage costs trending around $240-$250/MWh for larger procurement. Reliability remains a key driver, with value in storage being located close to load in supporting grid stability in light of transmission constraints, especially in markets like California. However, given relative complexity compared with solar assets, managing the operating risks of storage assets will be key to enabling reasonable IRRs.
Implications For The Power Sector
We think that with substantially lower capital costs for both solar and storage, solar developers will increasingly deploy co-located projects. Not only does adding storage provide the potential for offering a firm power product, it has meaningful up-front benefits from qualifying for ITCs. Storage effectively doubles the capital costs of a solar project, 26% of which can then immediately attract ITC benefits and also avail modified accelerated cost recovery system depreciation. We believe adding storage to existing sites offers significant growth prospects to companies like NextEra Energy (A-/Stable/--) because it already has the advantage of established interconnections.
We see AES Corp. as another company that would benefit given its widespread adoption of storage at renewables development sites, as well as at existing thermal sites. In particular, Fluence, AES' battery energy storage joint venture with Siemens AG, has disclosed that it expects to make $500 million in revenue in 2020 and increase it by 45% each year over the next five years. The company recently announced the acquisition of Advanced Microgrid Solution's (AMS) software and digital intelligence platform. AMS' technology uses artificial intelligence, advanced price forecasting, portfolio optimization, and market bidding to ensure energy storage and flexible generation assets respond optimally. We think this will support Fluence's ability to provide competitive fully integrated storage solutions given the need to provide optimized "revenue stacking" around storage assets. This is particularly important in markets like ERCOT to capture merchant value given real-time price volatility and lack of a capacity market. We note that at year-end 2020 Fluence received a $125 million private investment from Qatar Investment Authority for a 12% minority stake. After the transaction, AES and Siemens will remain major shareholders, with each maintaining a 44% stake. The transaction implies a $1 billion valuation, or roughly twice 2020 sales for the company.
While storage creates opportunities for many companies, it also poses a threat to companies that have significant conventional generation. As renewable costs continue to decline, they weigh heavily on both peak and off-peak power prices. We think this changing market paradigm affects gas-fired assets in different ways depending on their location or their ability to respond quickly to market signals. We think that in several regions of the country gas-fired assets will continue to coexist with renewables. For instance, Energy & Environmental Economics (E3) believes that the least-cost electricity portfolio to meet the 2050 economy-wide greenhouse gas goals for California includes 17 GW to 35 GW of natural gas generation capacity for reliability. The market consultant concluded that firm capacity is needed even while adding very large quantities of solar and electric energy storage. We think power companies with gas-fired assets could witness cash flow backwardation as they cede market share slowly to renewables but have adequate time to pivot strategies.
However, companies that have coal-fired generation exposure could face substantial asset retirements. Given continued low power prices, increased renewable penetration, and higher costs to operate coal units, we see nearly 85 GW, or about one-third of existing capacity, at risk (see link to our November 2019 report below). It is no surprise then that companies such as Talen Energy Corp. have announced substantial capital plans. In a November 2019 strategic shift, Talen announced that its Montour generation facility located in Pennsylvania and its Brandon Shores and H.A. Wagner coal generation facilities located in Maryland will cease coal-fired operations by the end of 2025 and repower with alternative fuel. The company simultaneously announced that it would undertake about 1 GW of renewable energy (both solar and wind) and battery storage projects currently under development across its existing asset footprint.
Similarly, in September 2020, Vistra Corp. announced a comprehensive plan to accelerate its transition to clean power generation sources and advance efforts to significantly reduce its carbon footprint. The company's plans include the development of six new solar projects and one battery energy storage project that total nearly 1 GW and represent a capital investment of approximately $850 million in ERCOT. Vistra simultaneously announced its next phase of coal plant closures in Illinois and Ohio. The company plans to retire seven plants with a combined capacity of more than 6.8 GW between 2022 and 2027. In aggregate, Vistra and its subsidiaries have retired or announced the retirement of more than 19 GW at 23 coal and natural gas plants since 2010.
"Once in a while you get shown the light, in the strangest of places if you look at it right"
So continue the lyrics of that fan favorite Grateful Dead song that we alluded to in our introduction. Like many market participants, we have focused on the cost curve of storage for some time and have been watching the rapid increase of battery projects in the interconnection queue with some interest. However, we now believe that the market has been spending too much time monitoring the declining cost curve of batteries and, in doing, so missing the forest for the trees.
We think this is not just about storage cost curves. While that is an enabler, the real story is in the improving economics of solar power generation. In 2020, overnight capital costs of solar PV installations declined below that of onshore wind generation. Moreover, operating expenses for solar PV projects have also declined substantially over the past three years. As a consequence of these developments, the PPA index prices for solar power have declined below those of wind power (for the first time in the first quarter of 2020). Given batteries would provide the much-needed firming that solar power needs, we think large-scale co-located solar plus storage deployments are now imminent, which would be a gamechanger for the power industry.
Battery Storage --A Primer
Battery storage applications store electrical energy in the form of chemical energy, which is then converted back into electricity when needed. Batteries contain two electrodes (the anode and cathode), two terminals composed of different chemicals, and an electrolyte that separates the terminals. The electrolyte enables the flow of ions between the two electrodes and external wires allow for electrical charge to flow.
The U.S. has several operational battery-related energy storage projects based on lead-acid, lithium-ion, nickel-based, sodium-based, and flow batteries. Within this subset of batteries, lithium-ion (Li-ion) technology is about 85% of all new battery capacity and continues to be over 80% of aggregate capacity installed this year, followed distantly by sodium sulfide (NaS) chemistry batteries manufactured by NEC Corp. in Japan. Li-ion batteries are the fastest-growing energy storage markets due to their high energy densities, high power, near 100% efficiency, and low self-discharge. Their longer operating history and relatively large original equipment manufacturers (Samsung, Tesla, etc.) also offer a compelling difference to other technologies.
Most lithium-ion batteries use graphite as an anode. Graphite is traditionally supplied from mined natural flakes or derived from petroleum coke (synthetic graphite). Synthetic graphite is preferred due to its higher consistency and purity compared with natural graphite, but it is also more expensive. There are a wide variety of cathodes--with implications for performance and cost (see chart 6).
There is no commercially available ideal lithium-ion chemistry suitable for all applications. The choice of chemistry is typically a trade-off between energy density, power density, safety, life, and cost requirements.
Energy density. Chemistries differ in their ability to store electricity by unit of weight (measured in watt-hour per kilogram). The energy density has been increasing due to improvements in chemistries and battery pack engineering. Some of the potential improvements in pack engineering include the use of larger cells and the reduction of non-cell components. NMC and NCA batteries (see elements in the battery in chart 6) have the highest energy density of commercial batteries. Generally, partly replacing cobalt with nickel increases the energy density, but at the expense of lower safety and cycling ability.
Power density. This refers to the amount of power a battery can supply by unit of weight and is typically measured in watts per kilogram. Higher power chemistries include LCO (expensive for transportation or power storage applications), LMO (high degradation), LFP (low energy density) and LFP/LTO (lower energy density and higher cost but also highest safety).
Safety. Thermal runaway--the primary safety concern for Li-ion batteries--involves batteries reaching a specific temperature threshold and triggering rapid combustion. The reactions break down the electrolyte and release flammable and toxic gases such as carbon monoxide, hydrogen, or ethylene. Overcharging; external factors, such as heat or mechanical damage; or manufacturing defects can lead the battery to become thermally unstable.
Other storage technologies
Flow batteries. While many current energy storage projects employ Li-ion batteries (cellphones etc.), others are testing more esoteric chemistries and technologies. Prominent among these are liquid electrochemical systems known as flow batteries. These are made by pumping the electrolyte (a brew of metallic salts) through an electrochemical cell. For long-duration storage, flow batteries are the most likely candidates because they are constructed in tanks and have the theoretical advantage of being scaled to larger sizes. Their advantage is in their ability to decouple the power and energy capacity because the energy is a function of the quantity of electrolyte stored. To add battery capacity, one merely increases the size of the tank or links a series of tanks together. They also have few moving parts and a relatively longer working life. What's holding them back is cost: Most flow batteries use the element vanadium as an electrolyte, which is expensive. Flow batteries can also use zinc-bromide. To date, limited deployments have happened due to difficulty getting projects financed, as well as the rapid cost reduction of lithium-ion batteries. This in turn has limited flow batteries to a duration greater than six to eight hours, which currently remains a niche market. The industry is challenged by the need to scale rapidly enough to go down the cost curve. We note that RedT Energy and Avalon Vanadium Batteries, two flow battery manufacturers, merged in March 2020, to form Invinity Energy Systems, which might lead to a higher manufacturing scale.
Solid state. We note new battery technology involving solid state batteries could potentially provide many benefits over Li-ion batteries in the form of higher energy density (about 50% more per unit mass and about 2x per unit volume), faster charging times, increased range, less fire risk, longer life, and better low temperature performance. Besides, compatibility of solid state batteries with new cathode technologies like FeF3, NiF2, and CoF2 could lower costs (i.e. smaller battery packs).
Hydrogen. The technology could offer multiple days (if stored in tanks) or seasonal storage (if stored in geological storage). Hydrogen can be generated by electrolyzers at times of low demand/peak renewable generation and used in a gas turbine or a fuel cell to generate electricity when needed. The round-trip efficiency is low, and equipment remains expensive, but hydrogen could offer seasonal storage capabilities, charging during the summer at times of high solar PV generation and discharging in winter at the time of heating needs.
For now, Li-ion has taken over and presented itself as a solution for energy storage problems in multiple industries. We think that it will remain the market leader for grid storage for the next 10 years. However, that does not mean that other batteries, like lithium sulphur, magnesium-ion, or zinc battery solutions, which are in the research laboratory today, will not eventually dominate. There is now a range of battery manufacturers, suggesting their prospective cost curves on scaling the technology would beat out Li-ion. What appears certain is the trend toward scaling of a multitude of technologies, each with the ambition to construct reliable batteries.
- S&P Global Sharply Lowers Its Recovery Values For Coal-Fired Power Projects, Nov. 7, 2019
- Credit FAQ: What Factors Power A Rating On A Utility-Scale Battery Storage Project?, Jan. 12, 2018
- Going With The Flow: How Battery Storage Economics Are Changing Power Consumption, Jan. 11, 2018
- Uncharged: The Possible Market Impact Of Electricity Storage, Jan. 10, 2018
- Future Shock: Will Better Batteries Dim Electric Utilities' Prospects?, Nov. 3, 2017
- Storage: The Final Piece In The Global Energy Transition Puzzle, June 10, 2016
This report does not constitute a rating action.
|Primary Credit Analyst:||Aneesh Prabhu, CFA, FRM, New York + 1 (212) 438 1285;|
|Secondary Contact:||Jason Starrett, New York + 1 (212) 438 2127;|
|Research Assistant:||Sachi A Sarvaiya, Mumbai|
No content (including ratings, credit-related analyses and data, valuations, model, software or other application or output therefrom) or any part thereof (Content) may be modified, reverse engineered, reproduced or distributed in any form by any means, or stored in a database or retrieval system, without the prior written permission of Standard & Poor’s Financial Services LLC or its affiliates (collectively, S&P). The Content shall not be used for any unlawful or unauthorized purposes. S&P and any third-party providers, as well as their directors, officers, shareholders, employees or agents (collectively S&P Parties) do not guarantee the accuracy, completeness, timeliness or availability of the Content. S&P Parties are not responsible for any errors or omissions (negligent or otherwise), regardless of the cause, for the results obtained from the use of the Content, or for the security or maintenance of any data input by the user. The Content is provided on an “as is” basis. S&P PARTIES DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, ANY WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE OR USE, FREEDOM FROM BUGS, SOFTWARE ERRORS OR DEFECTS, THAT THE CONTENT’S FUNCTIONING WILL BE UNINTERRUPTED OR THAT THE CONTENT WILL OPERATE WITH ANY SOFTWARE OR HARDWARE CONFIGURATION. In no event shall S&P Parties be liable to any party for any direct, indirect, incidental, exemplary, compensatory, punitive, special or consequential damages, costs, expenses, legal fees, or losses (including, without limitation, lost income or lost profits and opportunity costs or losses caused by negligence) in connection with any use of the Content even if advised of the possibility of such damages.
Credit-related and other analyses, including ratings, and statements in the Content are statements of opinion as of the date they are expressed and not statements of fact. S&P’s opinions, analyses and rating acknowledgment decisions (described below) are not recommendations to purchase, hold, or sell any securities or to make any investment decisions, and do not address the suitability of any security. S&P assumes no obligation to update the Content following publication in any form or format. The Content should not be relied on and is not a substitute for the skill, judgment and experience of the user, its management, employees, advisors and/or clients when making investment and other business decisions. S&P does not act as a fiduciary or an investment advisor except where registered as such. While S&P has obtained information from sources it believes to be reliable, S&P does not perform an audit and undertakes no duty of due diligence or independent verification of any information it receives. Rating-related publications may be published for a variety of reasons that are not necessarily dependent on action by rating committees, including, but not limited to, the publication of a periodic update on a credit rating and related analyses.
To the extent that regulatory authorities allow a rating agency to acknowledge in one jurisdiction a rating issued in another jurisdiction for certain regulatory purposes, S&P reserves the right to assign, withdraw or suspend such acknowledgment at any time and in its sole discretion. S&P Parties disclaim any duty whatsoever arising out of the assignment, withdrawal or suspension of an acknowledgment as well as any liability for any damage alleged to have been suffered on account thereof.
S&P keeps certain activities of its business units separate from each other in order to preserve the independence and objectivity of their respective activities. As a result, certain business units of S&P may have information that is not available to other S&P business units. S&P has established policies and procedures to maintain the confidentiality of certain non-public information received in connection with each analytical process.
S&P may receive compensation for its ratings and certain analyses, normally from issuers or underwriters of securities or from obligors. S&P reserves the right to disseminate its opinions and analyses. S&P's public ratings and analyses are made available on its Web sites, www.standardandpoors.com (free of charge), and www.ratingsdirect.com and www.globalcreditportal.com (subscription), and may be distributed through other means, including via S&P publications and third-party redistributors. Additional information about our ratings fees is available at www.standardandpoors.com/usratingsfees.
Any Passwords/user IDs issued by S&P to users are single user-dedicated and may ONLY be used by the individual to whom they have been assigned. No sharing of passwords/user IDs and no simultaneous access via the same password/user ID is permitted. To reprint, translate, or use the data or information other than as provided herein, contact S&P Global Ratings, Client Services, 55 Water Street, New York, NY 10041; (1) 212-438-7280 or by e-mail to: firstname.lastname@example.org.