Financing distributed batteries and electric vehicle charging stations
Pricing for bank debt for behind-the-meter batteries has shown significant margin compression in the last 24 months, suggesting growing interest among banks in lending to the sector. Several different business models for deploying electric vehicle charging infrastructure are emerging.
A developer and two investors talked about opportunities and challenges with financing these types of assets at Infocast Storage Week Plus in San Francisco in late July. The following is an edited transcript.
The panelists are Douglas Staker, a vice president at Enel X, Dan Cary, a senior vice president at Macquarie Capital’s Green Investment Group, and Peter Nulsen, a director at Generate Capital. The moderator is Deanne Barrow with Norton Rose Fulbright in Washington.
BARROW: Bill Peduto, the mayor of Pittsburgh, said that we are fast approaching a time when we are no longer going to make our morning toast from energy that comes from a distant power plant. He is describing the evolution of a more decentralized energy grid. To get there, a key piece of the puzzle will be access to capital to finance distributed assets. Traditional project finance expects a fixed-price, long-term offtake contract with a creditworthy counterparty, but this is not always available with distributed energy assets.
Let’s kick off the discussion by delving into the unique aspects of financing distributed energy. Peter Nulsen, Generate Capital focuses exclusively on financing distributed rather than utility-scale assets. When you are looking at a distributed energy project, where do you probe? What are the key areas on which you focus?
NULSEN: There are a lot of places we probe. The first is the revenue stream. Distributed generation usually involves a combination of a fixed revenue stream together with one or more variable revenue streams that are dependent on the performance of a fleet of assets as opposed to a single asset.
A challenging aspect is that there may be several different underlying revenue contacts of varying lengths, and the lengths do not always match the underwritten term. For example, we could be underwriting a 10-year term on an asset with a 10-year life and a 10-year customer contract, but on top of the contracted revenue stream, we may be bolting on additional uncontracted revenue streams that are not fixed in future price or future term. We have to take a view on where those prices are going and how long we think the revenue streams will be around for.
Another area we focus on is volume. Generate Capital is set up to enter markets early and de-risk investments where other investors are not. We look for sponsors who can deliver at least $50 million of projects over the next two or three years. The sponsor does not need to be doing $100 million or even $10 million of volume today. We will focus on its growth trajectory. Not every small energy efficiency company is going to be able to deliver that kind of deployment growth over three years. It is really important to have a view on what growth in project deployment will look like for a given sponsor.
BARROW: You mentioned uncontracted revenue streams. Do you as a lender or equity investor give credit to merchant revenue when you size the loan or investment?
NULSEN: We do so on a case-by-case basis. When we underwrite a merchant revenue stream, which for folks in the audience generally means subject to fluctuations in quantity or price, we would have to take a view on how the price and quantity will vary over time. We analyze the factors that influence supply and demand and general and specific market conditions. This helps us identify overarching factors that may give us comfort that the revenue stream is going to grow or perform in a certain way.
Merchant revenue streams go against the grain of traditional project finance like you said, but that creates an opportunity for a company like Generate Capital to differentiate itself. We do not necessarily mark merchant revenue zero in a pro forma financial model just because it is uncertain. If a distributed asset has a significant revenue stream with a smart story behind it, we are open to investing in it.
There are gradations in the degree to which a revenue stream is merchant. A distributed asset that receives revenue due to demand-charge reduction can be variable due to the variability of tariffs or performance of the asset. That degree of variability, however, is less than payments for frequency regulation in PJM or energy payments in ERCOT or CAISO. Different valuation frameworks are required for each.
Distributed battery portfolio
BARROW: So the revenue analysis is done on a case-by-case basis. Let’s look at a specific case. Dan Cary, Macquarie invested in a fleet of behind-the-meter battery storage projects in Southern California Edison territory. The batteries provide 65 megawatts of capacity to SCE. They are installed at businesses where they reduce demand charges for the host customers. One fleet of batteries generates two revenue streams: one from the utility and another from the host customer. Are both payments fixed? On what other issues did Macquarie focus?
CARY: You are describing what we call the Electrodes portfolio, which is our partnership with AMS — formerly known as Advanced Microgrid Solutions — for deployment of almost 100 distributed battery storage facilities in the west Los Angeles basin. Before diving into Electrodes, let me talk first about our infrastructure investor philosophy at a higher level.
For background, Macquarie’s Green Investment Group is the platform through which Macquarie Group invests its balance sheet in assets and platforms that support the transition to a low-carbon economy. Our mandate is asset creation, and we partner with developers to de-risk assets for the most optimal financing terms.
The way we normally think about distributed projects in particular is to focus on three key concepts outside of project returns.
First, we look for a contractual backbone. We are not necessarily looking for a 100%-contracted, 25-year, availability-based power purchase agreement with a US utility. In today’s renewable space, investors and lenders are taking a view on their exposure to variable pricing. We review on a case-by-case basis this exposure and the market fundamentals involved, although we like to see at least some portion of the income derived from lower-risk, recurring revenue.
Second, we focus on bankability of the contracts, most importantly, the EPC contract, the O&M contract and the customer or utility agreements. We focus on whether those contracts are, or can become, non-recourse agreements with high-credit, experienced counterparties. We also try to ensure that the key tenors and exposures of the agreements are all aligned.
Third, we focus on the dynamics of the local market. Distributed resources are typically high-value when they address a specific local problem or need. We have to believe the assets we deploy are part of a long-term solution. If the distributed energy asset is providing specific “merchant” or “variable” services, then we are conscious of the saturation point for those services as more assets are deployed in the local market.
Let me illustrate those three concepts with the Electrodes story. Starting with the contractual backbone, when we invested in Electrodes three and a half years ago, AMS had successfully won an RFP with Southern California Edison to provide resource adequacy, which is California’s version of capacity. To fulfill the resource adequacy requirements, AMS went on to contract with the owners of 90 different commercial and industrial sites to provide them with six-hour, behind-the-meter battery energy storage systems that, collectively, provide a total of 65 megawatts, or 340 megawatt hours, of capacity.
SCE makes a payment for capacity, and the host customers pay a fee for the batteries, saving them money on their energy bills.
It was key that these revenue streams and the associated services were with quality counterparties and that all contract terms and tenors were aligned with the rest of the contracts. We worked with AMS on more than 3,000 unique contracts and documents across the portfolio.
In terms of the local market story, the west LA basin near Aliso Canyon suffers from significant transmission and distribution constraints. The constraints are evident today, but even more importantly, the area is expected to need an increasing level of support over the coming decades. This was a key factor for us to see value in the systems. We did extensive research to validate the local grid story to feel confident that the portfolio would be able to support the ongoing transmission and distribution concerns in the area.
BARROW: You said the batteries are installed at 90 sites. How many different customers make up those 90 sites?
CARY: Around 30.
BARROW: Were the 30 customers rated corporations? If they were not, how did you assess credit risk to ensure creditworthiness?
CARY: Good question because it is something we are having to do more and more now with distributed resources. Many of them were not rated. We spent a lot time on each host customer to understand each and to assess the critical nature of the site to its operations. Certain characteristics regarding the site and the business being conducted there were very important. For example, if the site is leased, the lease tenor should be aligned with the customer agreement term.
BARROW: Doug Staker, Enel X finances behind-the-meter energy storage projects on balance sheet. When you are putting that capital to work, where do you probe on diligence?
STAKER: It is a matter of learning to get comfortable being uncomfortable. We have to convince Italians to be comfortable being uncomfortable, which is not always an easy task.
One of the things we focus on is the disruption factor, especially from climate change. Hurricane Sandy changed the whole interconnection process in New York. All of a sudden, everybody wanted resilience. While that was happening, we would come out and talk to folks in California, and they would look at us and say, “Resilience in California? Are you serious? There is no value stream here.” Then there were wildfires. Resilience is now important in California. I try to keep my eyes open and think about what is the disruption factor that can occur.
If I only had a dollar for every time somebody asked me, “You are basing some of your revenue streams on demand-charge savings. What if demand charges go down?” But no one can give me an example of demand charges going down. The challenge for utilities right now is with managing the demand side, and they are putting more and more rate recovery on the capacity or demand side, and less on energy. Although we are starting to see more contracted revenue offerings from utilities for distributed energy resources and non-wires solutions, there is still a blend of contracted and merchant revenue streams to get comfortable with.
New York just announced a largescale offshore wind project. People ask, “What does that mean from the wholesale market perspective and the whole optimization of the delivery system?” Some people are scratching their heads over how all of that power will be integrated. That tells me that there is opportunity.
Not many people are thinking about the value of being able to create load. We always talk about the value of creating generation to reduce load, but not the value to create load at certain times in certain markets. As we start to look at beneficial electrification, where does that drive value?
BARROW: Enel X has been active in the New York distributed energy storage markets since 2012. You have more than a dozen behind-the-meter battery storage projects in New York providing services to Con Ed. Does it work the way Dan Cary described for the Electrodes project in California, where there are two revenue streams? Does Con Ed make a capacity payment and do the host customers pay for demand charge reductions?
STAKER: In the last year, the options in New York have multiplied.
The classic case has been behindthemeter demand – charge reduction. There has always been a nice incentive stream around that service, whether from providing demand-charge reduction for the end customer or providing demand response for Con Ed. Those programs evolved over time and have become more lucrative.
An interesting recent development is the New York Public Service Commission’s adoption of a replacement for net metering called the “value of distributed energy resources” or VDER. We have had revenue streams from solar systems, and now solar-plus-storage systems receive approval for VDER value. We perform an analysis of whether it is more lucrative to offer a non-wires solution to the utility under a contract, to participate in the demand-response program and look for the incentives available there, or to connect solar plus storage in front of the meter and export power to the grid to get the VDER rate.
We do that analysis all the time before we sign up for a non-wires solution contract where, like Dan described for Electrodes, the project has to make a commitment to provide the utility a certain level of load reduction in exchange for contracted revenue. We look at all three of those potential business cases and decide which is the better play.
BARROW: Let’s talk about the cost of capital for distributed energy resource infrastructure. Keith Martin produces a webinar in January of each year called the Cost Of Capital, where industry experts give their senses of where the market is headed for tax equity, debt and term loan B financing. Let’s compare the outlook for 2019 and see if you think the figures hold true for distributed projects. It was said that debt for plain-vanilla projects in the US is pricing at between 125 and 137.5 basis points over LIBOR in terms of the spread, and that debt service coverage ratios are 1.3 or 1.4.
Dan Cary, Electrodes raised additional project finance debt late last year. Where did it land? If you can’t speak to Electrodes specifically, give us a sense of pricing and DSCRs for distributed storage generally or another class of distributed assets.
CARY: I think the thing that can give the industry real comfort, and certainly has given Macquarie comfort, is that financing for distributed energy resources more generally, and distributed energy storage in particular, has changed markedly over the last 24 months in the right direction.
We raised external debt on a project finance basis for Electrodes twice. The first financing closed in March 2017 and the second at the end of last year. The second financing was done by a bank club made up of four traditional project finance lenders. The growth in appetite in the lender market over that period was clear and the terms of the financing reflected that. Within that relatively short period, I think lenders have become more comfortable with the industry and with the way the technology works and is paid for.
In terms of return thresholds, in general I think about this by comparing a portfolio of distributed assets to an operating solar-plus-storage microgrid that has a long-term PPA with a utility as the sole revenue stream being financed and that is fully wrapped for technology and performance risk. From a lender’s perspective, the debt will be priced at the sort of levels Deanne mentioned. From the investor’s perspective, the levered equity return might be at high single digits.
There is a premium for every conceptual risk that you add into this “fully-wrapped” project dynamic, and that goes for both lenders and investors. There will be a premium for construction risks if the financing or sale is occurring when notice-to-proceed with construction is issued rather than when the project is in operation. There will be a premium for technology exposure for new offerings not wrapped by a bankable technology counterparty for the whole operating term. There is typically operating software risk, especially if the project is using a tool that has not been tested or proven and is not backed by a performance guarantee. There is also a premium to account for taking on complexity with a portfolio of distributed assets. At the Green Investment Group, we work to minimize these premiums through a highly-structured set of contracts, so that the project is left with minimal risks that are added into the equity return premium for the eventual financier and owner.
BARROW: One follow-up question. You moved to four lenders in the second financing from one lender in the first?
BARROW: We hear that there are more lenders chasing deals than there are good projects to finance. Is that true even for distributed infrastructure projects? How difficult was it to line up those four banks?
CARY: There is certainly appetite in this space, but there is some reluctance to price at competitive levels before banks or investors have done their first one. I often see momentum in markets like this, in that the second and subsequent deals can be executed more easily than the first as lenders and investors get comfortable with the inherent risks involved. I think as we see more deals get done, we are going to see pricing get even more competitive as people are successful in entering the space and want to grow their books.
BARROW: Peter Nulsen, Generate Capital leverages most of its capital as project equity. What kind of internal rates of return are investors expecting in this space?
NULSEN: To take a step back, Generate is a permanent balance sheet investor, which means we have flexibility to do project equity, project debt, asset-backed deals, mezzanine capital and all of the above, to match what the developer or entrepreneur is looking for. In a lot of cases, we piece together different kinds of financing and investment solutions to help the entrepreneur scale up from, say, a $5 or $10 million investment to a large, $100 million distributed fleet.
From a returns perspective, we look at a buildup of risks. It is hard to put a number on one. Generate was one of the first investors in distributed storage in 2014. As you can imagine, the pretax return from a distributed storage deal in 2014 is significantly lower today. It is interesting to watch the lithium-ion battery storage industry come down the bankability curve and end up where Electrodes did. Other classes of distributed energy resources, like distributed fuel cells and behind-the-meter natural gas generators, can be conceptualized in terms of a risk premium buildup just like storage.
EV charging models
BARROW: We have talked a lot about distributed energy storage. Let’s shift gears to another kind of distributed asset, which is electric vehicle charging infrastructure. Doug Staker, Enel X has a business under its umbrella called eMotorWerks. eMotorWerks installs EV charging infrastructure at residential and commercial sites. What business model does eMotorWerks use to support electric vehicle charging infrastructure buildout?
STAKER: It is the same as what we do with storage and the same as what we do with solar plus storage. It is really about the value of flexibility. The ability to control and manage electric vehicle charging as a flexible resource has piqued our interest.
We have installed about 8,500 car chargers across Italy. We are going to do some large programs here in California and in the northeast because we can start to see more users of EVs showing up. This growth will lead to challenges around managing the increased load on the grid and putting value around load flexibility.
Enel X controls about 5,000 megawatts in demand response around the world through our subsidiary EnerNOC. Demand response is managing customers’ peak load, in a classic case, for maybe five cycles in a hot summer or 10 cycles in a season.
Customers have a pain threshold. At a certain point, they want to opt out of the demand response program. What is nice about charging, storage or load management through software control is that it is transparent from the customer’s point of view. The more transparent it is, the more people will be willing to participate in programs that can give the grid the flexibility it is going to need.
The opt outs are not just in the peak season. We are starting to see challenges in the shoulder season, when renewable production is high and load is low, so there are ramping conditions that have not been seen before. The grid needs flexible resources to help manage load effectively and with value.
BARROW: There are currently only about a million electric vehicles in the United States, but California has a goal of putting five million electric vehicles on the road by 2030 and a goal of building 250,000 electric vehicle charging stations, also by 2030. Dan Cary and Peter Nulsen, what do you perceive as the key opportunities and risks for third-party financing of EV charging infrastructure? In terms of opportunities, what business models are showing promise?
CARY: Macquarie Capital is able to act throughout the company and asset lifecycle as an early-stage venture investor or, alternatively, as an infrastructure investor in a non-recourse traditional project financing. We view basically any asset that is either generating electrons or using electrons and that is dispatchable as a potential product in this space. Electric vehicle charging fits that paradigm.
Several different business models are emerging. One of the companies we have invested in provides mobile EV fast charging infrastructure via a hardware sales platform.
Another business model is “EV infrastructure as a service,” which is where EV chargers are effectively rented.
A third play focuses on using the software involved in EV charging technologies to monetize the assets.
Another emerging model is to include EV charging infrastructure as part of a larger, onsite distributed infrastructure package that involves a combination of solar PV, storage, energy efficiency and combined-heat-and-power units, among others things. It can often make sense in terms of scale of the contract structure to make the EV charging stations part of that whole package.
For a traditional third-party project financing of small systems to work, real scale is needed to aggregate enough EV charging units into a financeable portfolio. There is an opportunity for somebody to work out how to create enough volume in an area where he or she can also manage the infrastructure to generate additional income.
NULSEN: We have looked at clean transport a fair amount, specifically EV chargers as well as the vehicles themselves. We formed a partnership with BYD to lease electric buses to municipalities in California. We remain bullish on the sector, specifically EV charging.
There are some fundamentals worth highlighting as we move toward a third-party-financed offering. The first issue on the revenue side is utilization. You have to take a long view on the electrification of vehicles. How do you get comfortable that, say, 200 vehicles per day will travel between LA and San Francisco? A unique analysis needs to be done to get comfortable with utilization.
It will help spread the utilization risk if a lot of EV charging stations are aggregated in a portfolio, but the flip side to that coin is cost. If you can get utilization to a somewhat predictable level, then the challenge becomes ensuring that electricity costs are manageable. This might require a special utility tariff for EV charging like we are moving toward in some parts of California. There could be an energy storage play to optimize the value stream, or even a solar-plus-storage play.
It is early days still, but we are looking at how to bring all of those elements together and are working with developers and entrepreneurs to build the right kind of third-party-financing offer. Hopefully you will read about something new in the next 12 months or so.