Special Report: The truth about energy storage

For as long as renewable energy has been a threat to the conventional power sector, it has been plagued by warnings that too much wind and solar would destabilise the grid.

Intermittent renewable energies could never power advanced economies unless they were backed up by large-scale affordable battery storage — a technology that had always seemed a distant dream, a science-fiction fantasy.

Yet in what seems like the blink of an eye, the dream seems to suddenly be not only achievable, but imminent and inevitable. And the implications for renewables are profound.

In the past 18 months, the cost of lithium-ion (Li-ion) batteries has fallen 50%. And with vast new manufacturing plants under construction, including Tesla’s Gigafactory in Nevada, costs will continue to fall, with no price bottom in sight.

In Silicon Valley and elsewhere, storage start-ups are closing ever-larger funding rounds. Swiss start-up Alevo has attracted a staggering $1bn in funding without having a single battery in commercial operation, while California-based Stem has drawn millions from major players such as Iberdrola, Total and GE. At the same time, renewables heavyweights including SunEdison and SunPower are acquiring or forming partnerships with storage companies, and industrial giants such as Siemens and GE are overhauling and sharpening their own storage strategies.

Nearly every renewables conference now seems to include sessions on storage, and they are often bigger draws than the main events. And when Tesla wunderkindElon Musk unveiled his company’s new stationary storage business in May, it made the same kind of headline-grabbing impact as the launch of a new iPhone.

Add to all this the fact that California has mandated that its three big utilities procure a total of 1.3GW of battery storage by the end of this decade, and it seems that storage is the biggest thing to hit renewables since sliced polysilicon.

“It’s a total cliché, but storage really is going to be game-changing,” says Tom Starrs, vice-president for market policy and strategy at SunPower. “It will contribute very substantially to solar, wind and other zero-carbon technologies becoming the anchor energy technologies of the future.”

Yet for all the excitement and the limitless potential, it is important to keep things in perspective. Aside from a few small, isolated renewables-hungry power markets — such as Hawaii, Puerto Rico and the Japanese island of Okinawa — battery storage is unlikely to play a major role in the energy sector for many years to come.

“Storage isn’t necessary over the next three or five or even ten years to enable substantial growth in solar and wind,” Starrs points out.

The global market for stationary storage is currently estimated at around $1bn a year — “so tiny it’s difficult to count”, as one analyst puts it. And no-one knows what the storage market will look like in ten years — how large it will be, which companies will dominate, which technologies they will use, even what their business models will look like.

There are a few things, however, that nearly all the experts agree on. Stationary storage has shifted from a research-driven subsidised niche market into a high-growth phase. In the short term, storage presents interesting business opportunities for forward-thinking renewables companies; and in the long term, it will be critical in enabling grid operators to integrate ever greater levels of wind and solar energy.

“The question is not how we get to 33% or even 50% renewables, but 80%, where most people agree we need to be to address the climate challenge,” says Starrs.

There also seems to be a consensus on how the storage market will develop over the next few years — and it may surprise many in the wind and solar industries.

Grid-scale storage

When people in the renewables industry think of battery storage, they often picture banks of batteries at wind and solar farms, allowing power producers to pump out a steady flow of green energy around the clock.

After all, on-site storage projects have already attracted a great deal of attention, notably AES’s 32MW Li-ion battery system at its 98MW Laurel Mountain wind farm in West Virginia, and Duke Energy’s 36MW battery system at its 153MW Notrees wind farm in Texas, which was “repowered” in June by German start-up Younicos, replacing the underperforming lead-acid batteries with Li-ion.

Yet these storage facilities are not being used to supply so-called “firmed” wind energy 24 hours a day — and experts are uncertain if such a model will ever make economic sense, outside of island networks. Both AES and Duke use their on-site storage to provide grid frequency-regulation services.

Experts believe it will usually make the most sense — economically and practically — to build storage assets closer to the load (ie, the place where the electricity is consumed), allowing batteries to be used for multiple grid-related purposes.

“If you see 200MW of batteries going up in the middle of nowhere, somebody missed the point,” says Johannes Rittershausen, managing director of New York-based storage start-up Convergent Energy + Power.

In simple terms, storage is unlikely to transform the way renewable energy is supplied to the grid — it will transform the way grid operators manage electricity, allowing far more intermittent energy onto the network.

Batteries will not revolutionise the electricity sector overnight, but they will eat away at it from the inside, steadily stealing work previously performed by other technologies.

Ancillary services such as frequency regulation, voltage control and operating reserves — the behind-the-scenes operations that keep the grid humming as power demand and supply fluctuate — is where battery storage will come into its own.

Take frequency regulation, for example, which helps maintain the grid at 50Hz or 60Hz (depending on the network) by injecting or absorbing small amounts of power on a second-by-second basis — effectively smoothing it out. There’s always been a need for frequency regulation and, historically, slower-reacting gas- or coal-fired plants have managed the job. But as more variable renewables come on line, balancing the grid becomes more challenging, and the need for such smoothing grows. It turns out that batteries, which can charge and discharge power virtually instantaneously, are almost perfectly suited for the task — and may be far cheaper than building a new fossil-fuel plant.

In recent years, a steady stream of companies that develop, build and operate renewables plants have diversified into battery-based ancillary services, including RES, Canadian Solar, EDF, and AES. The latter operates the world’s largest fleet of grid-scale batteries, with nearly 200MW on line or under construction (still a drop in the ocean compared to the 35GW of power plants it oversees).

After ancillary services, the next big target for storage developers is providing back-up power within so-called capacity markets — in which companies, often independent power producers (IPPs), are paid by grid operators to keep power plants on the sidelines, ready to spring into action whenever additional power is needed.

Historically, this back-up power has come from pumped hydro or gas-fired peaker plants. And in purely economic terms, gas peakers still hold a sizeable edge over grid-scale batteries in most capacity markets. But the landscape is changing quickly, especially in places keen to curb their carbon emissions, as Southern California Edison’s (SCE’s) groundbreaking procurement round last November shows.

In that auction for back-up power, necessitated by the sudden closure of the San Onofre nuclear plant near Los Angeles, SCE invited IPPs to offer 2.2GW of capacity reserves. Assuming SCE would need its arm twisted to include batteries, California’s regulator mandated that at least 50MW of the 2.2GW come from stationary storage. But SCE caught the industry off guard by awarding storage contracts for 261MW — a testament to the improving economics of storage and utilities’ growing acceptance. Contracts include a 100MW Li-ion battery system to be built by AES, which will become the world’s largest; and a 50MW fleet of batteries to be installed on “hybrid-electric” buildings by Advanced Microgrid Solutions.

No-one expects peakers to vanish overnight — after all, in addition to its battery project AES bid and won a contract for a 1.28GW gas plant in the same auction round. But SCE’s decision to include so much storage was “a clear turning point in the market”, says Brett Galura, vice-president for solution development within AES’s storage unit.

The average US peaker plant is used just 7% of the time, and as batteries get cheaper, the economics of such plants will increasingly be called into question, notes Galura. “To spend hundreds of millions of dollars on a power plant and use it less than 10% of the time seems very wasteful.”

GE, which is marketing itself as a builder of turnkey storage projects after scaling back its sodium-nickel battery production business, today makes far more money supplying equipment for gas peaker plants than for storage, but believes that batteries will increasingly steal market share. “Our view is that you’ll continue to see that happen,” says Jeff Wyatt, GE Renewables’ general manager for energy storage.

Batteries have many natural advantages over peaker plants, including their ability to absorb as well as discharge power, their quick-to-build nature, and their modularity — the latter two reflecting advantages wind and solar have over conventional generation.

Much of the expected price decreases for storage will simply come from greater economies of scale, as battery factories are built and expanded. But another trend that will greatly improve the economics is a push towards “value stacking” — tapping into multiple revenue streams with a single storage asset.

Presently, most storage systems are built with just one purpose in mind, such as frequency regulation. But in the future, storage will be more like a smartphone, says Ali Nourai, director of DNV GL’s storage unit.

“Nobody can justify spending $500 just to make calls,” he says. “They buy smartphones because they do so many other things.”

Battery arrays could simultaneously perform a variety of increasingly sophisticated functions — from load shifting to ancillary services to simple back-up — “but nobody offers the software to do that yet”, Nourai says.

“That’s the missing element. What’s needed today is smart software, not cheap hardware. We’re waiting for that Apple-like company to offer four or five values at the push of a button.”

Nearly all of the big markets for grid-scale storage today are the result of regulatory mandates. California’s is the largest to date, but other places, such as Ontario, Canada, and South Korea, are taking similar if smaller steps.

Over time, however, analysts say that the grid-scale storage market will spread more organically, with regulators crafting battery-friendly rules and then stepping out of the way.

California provided the spark, Nourai says. “Once the rest of the country sees utilities are doing this, and it’s working — it’s not serial number one anymore — they will be encouraged to do it. And the ripple will spread globally.

“It’s a matter of time; everybody will need storage.”

Stationary storage is “going to be the next big pipeline of energy assets”, Rittershausen says. “Five years ago, that was a completely crazy thought. Now it’s pretty much accepted fact across the industry.”

Reinforcing that point, SunEdison announced in May it will buy the 50MW of storage being built by Advanced Microgrid Solutions for SCE, and then drop it down into TerraForm Power — the first time storage assets have been added to a renewables yieldco.

Behind the meter

Stationary storage is also taking off behind the meter — in people’s homes or businesses, either with or without accompanying solar panels.

And one of the first places this makes clear-cut economic sense is on US commercial and industrial (C&I) buildings — anything from warehouses to hospitals to malls.

“The fastest-growing market for energy storage is C&I in North America,” says Anissa Dehamna, storage analyst at Navigant Research, explaining that this is largely due to “demand charges”.

In the US, the electricity bill for many C&I buildings comes in two parts: a charge for total energy usage, plus a separate demand charge, which corresponds to a building’s maximum power use (in any 15-minute period) in any given month. In places such as California, demand charges can account for half a building’s monthly power bill. So even relatively pricey batteries can be worth it if they help to “flatten” the building’s consumption over the course of a day — soaking up power at night, for example, and then offsetting peak demand.

Many of the most-hyped companies targeting C&I storage, like Stem, do not even offer solar packages. Batteries, after all, can be charged just as easily with dirty power as with renewables.

But solar is a natural complement to storage systems. And over the past year or so, rooftop solar companies have been climbing over one another to get into the C&I storage business — for example, SunEdison’s recent acquisition of Philadelphia-based Solar Grid Storage

The economics of solar-plus-storage are not yet as clear in the US residential market. Net-metering policies, the bedrock of the US home solar market, encourage PV owners to flow their excess generation back onto the grid rather than saving it for future use.

The residential storage market has been quicker to take off in places such as Germany, where the combination of a low feed-in tariff and high retail electricity prices does more to encourage self-consumption of solar power, boosting the case for at-home storage. While the price of storage is still too high for many, 20,000 storage systems were installed between June 2014 and June 2015, according to the German solar association, BSW. And with prices falling fast, the BSW expects rooftop solar-plus-storage projects to be cost-competitive with utility-supplied power within the next three years. In the sunnier parts of Germany, such as Bavaria, they may already be cost-competitive over a 20-year period. Dehamna predicts that the fastest-growing residential storage markets will be Germany, Austria, Switzerland, Italy and Japan.

Even in the US, many companies are betting big on residential storage, perhaps assuming that net-metering policies will eventually disappear. Tesla, for one, is targeting both the C&I and residential markets with its new Powerwall battery, which Musk describes as like having a “beautiful sculpture on the wall”.

Among the big home solar companies to announce plans to integrate Powerwall systems into their offerings are Sunrun and, of course, SolarCity, of which Musk is the largest shareholder.

In August, Musk said that Tesla — which has not even started supplying Powerwalls — is “basically sold out of what we could make” until the end of next year, and added his belief that sales will likely continue “at quite a nutty level”.

And for solar companies eyeing the storage market, Tesla, which uses batteries made by Panasonic, is far from the only game in town. Germany’s Sonnenbatterie, which uses Sony Li-ion batteries, has partnered California’s Sungevity, while SunPower has teamed with storage start-up Sunverge, which uses batteries from South Korea’s Kokam.

Compared to the utility-scale storage market, which will probably remain dominated by big utilities and IPPs, the emerging behind-the-meter storage market looks like a wide open field.

Will stand-alone storage companies come to dominate? Or rooftop solar companies? The utilities themselves? “I think it’s still a really open opportunity,” says Cosmin Laslau, a senior analyst for Lux Research.

Even the eventual business model to be employed by distributed storage companies is anyone’s guess. Few would have predicted a decade ago that the solar lease model would become central to the leading US rooftop PV companies such as SolarCity and Vivint Solar (which themselves did not exist ten years ago).

Small though it still is, distributed storage is a bigger market today than grid-scale storage — and will probably remain that way for the next five or so years, Laslau predicts. Utilities have a slower adoption cycle, and the utility-scale power market is more price-sensitive, making batteries a tougher sell.

“But as we get into the 2020s, the picture could well reverse,” he says.


The fever pitch around storage risks masking two important realities.

First, the market for stationary storage is vastly smaller than the market for renewables today — and probably always will be. The simple truth is the world needs far more wind and solar capacity than it does storage.

GE expects the global market for stationary storage to grow to around $6bn annually by 2020, says Wyatt. By comparison, the world spent more than $300bn on clean energy last year.

Tristan Grimbert, chief executive of EDF Renewable Energy, predicts storage will only ever be a tenth the size of the renewables market.

EDF, among North America’s largest wind owners, has diversified into storage in recent years, targeting frequency regulation. Storage is “a nice market for us”, Grimbert says. “It’s a market we know because we understand very well the intermittency created by renewables. But it’s not going to be as big a market in terms of capital deployment as solar or wind.”

For renewables companies, then, the biggest opportunity from storage is not diversification. It’s the chance to build more renewables — lots more. And there the conversation jumps back into the realm of the theoretical.

Nearly all experts agree that at a certain level of renewables penetration, storage becomes necessary. One commonly cited illustration is the “duck curve” made famous in energy circles by California ISO, the state grid operator, which shows the power load it must follow in an average day as greater amounts of solar capacity are brought on line. To follow the load, generators must ramp their baseload output between mid-afternoon — when there tends to be a demand lull — and early evening, when people get home from work and demand surges.

But as more solar capacity comes on line — effectively collapsing the afternoon load because PV panels are busy cranking out lots of power — grid operators face an ever-steeper daily climb.

The challenges of meeting this daily “ramping range” are formidable for utilities, but they can be greatly eased by storage.

That brings up a second reality about storage. All the warnings that too much renewables would destabilise the grid without storage have been greatly exaggerated.

So far, grid operators have proved much more adept at integrating variable renewables than anyone expected, and they have ever better tools at their disposal for doing so.

Improved forecasting of wind and solar resources, combined with longer operating histories at existing projects, has taken much of the guesswork out of predicting how much power a wind or solar plant will generate.

Meanwhile, the inherent variability of renewables is mitigated as projects are built across wider geographies — if the wind is not blowing in one place, then it is blowing somewhere else.

Increased interconnectedness between grids also means that if one network needs a power boost, it can often import it from a neighbouring grid — which is particularly true in Europe, where the need for storage is consequently not so great.

And other sectors are emerging that will help to integrate renewables, perhaps competing with storage. One example is “demand response”, which is part of the growing trend of “IT meeting the electricity system”, says Sonia Aggarwal, director of strategy at Energy Innovation, a clean-energy consultancy.

Demand response is kind of the inverse of storage. Storage allows the banking of excess renewable power when it is not needed, for later use. Demand response promises the ability to adjust overall power demand — in a home, even a city — to align it with how much renewable power is being generated at a given time.

Nest Labs, a home automation company acquired last year by Google, offers a service that allows utilities to turn down air-conditioning units within homes when local power demand is high. The homeowner is compensated, but for the utility it is still cheaper than firing up an idle peaker plant.

None of this suggests that storage will not be essential for renewables — one day.

“If you want to get to 100% renewables, then you definitely need to store the energy at some point,” says Dehamna. “But people have been saying we’d need storage at 5% renewables, 10%, 20%. None of those predictions panned out.”

There is no magic number for renewables penetration at which storage becomes necessary, says Wyatt. “That breaking point is going to vary.”

In other words, storage will not do one big thing to save renewables from their own runaway success. Rather, over years and decades it will take on increasing roles throughout the electricity system — some of them brand new, each of them making it a little cheaper and a little easier to integrate more wind and solar.

Anyone expecting a eureka moment will be disappointed, says Laslau. “It’s very early days, and we’re already seeing a lot of crowding in the market, with people sensing the long-term potential and looking to stake a claim. But this is going to be a long game.”