clock menu more-arrow no yes mobile

Filed under:

If you thought solar was going to hurt utilities, get a load of solar+storage

Photorealistic rendering of solar+storage system.
Photorealistic rendering of solar+storage system.

As I wrote in my previous post, power utilities are pushing back against rooftop solar, but by doing so they are only accelerating the development of solar+storage, which may prove to be an even bigger threat.

Residential solar+storage (my main focus in this post, rather than commercial or utility-scale versions) refers to a) an array of rooftop solar panels connected to b) some means of energy storage, usually a battery or batteries, controlled by c) smart software that enables the system to communicate with the grid.

Adding storage to solar radically expands its value — the whole is more than the sum of the parts. The ability to store energy gives the solar homeowner more control over her energy; she can shift her time of use, storing power when it's cheap and selling it when it's more expensive. And she can potentially participate in larger markets, as we'll see later.

It's pretty cool.

But it's also expensive as hell at the moment. The question is, how fast will solar+storage get cheaper and scale up? And how can that process be accelerated?

the garage of the future
In the future, every garage will have a mountain outside it.

The value of solar+storage

When people think about solar+storage, their first question tends to be how much it costs. But the focus on cost — a single "levelized cost of energy" from solar+storage, in the jargon — can be very misleading, as this World Energy Council report emphasizes.

It makes more sense to approach it from the other end. Appreciating the value of solar+storage, the many different value streams it produces, will help us understand why the cost of solar+storage varies so much from place to place.

The different value streams solar+storage produces are worth more in some places than others, more at some times than others, and more under some market structures than others. All those context-sensitive variables go into "the cost" of solar+storage.

So let's get super-nerdy for a minute and separate out those value streams, some of which are valuable only to the owner, some are also valuable to the utility and the larger grid.

Price arbitrage (aka buy low, sell high)

For the individual solar homeowner, this is the big one. She can store power when it's cheap and use it (or sell it to the utility) when it's expensive, pocketing the difference.

Importantly, this does not work under net metering, the policy that's in place in most states today. Under typical net metering, the utility is obliged to buy her surplus power at retail rates. That's as much as she'll ever be able to get for it, so there's no incentive to store it for later.

Only if net metering rates fall substantially below retail rates do the economics of price arbitrage start to work out. That's why solar+storage is spreading so fast in Germany and Japan — in addition to direct subsidies for storage, tariffs for solar energy have fallen below retail prices, opening the opportunity for arbitrage.

Grid defection (going "off grid")

Some people like the idea of being independent of the grid entirely. A large enough solar+storage system can do that.

But it has to be a very large system, and it rarely makes economic sense at today's prices. (See this analysis and this analysis for extremely math-y demonstrations of the point.)

Going entirely off-grid only makes sense in very particular circumstances: Retail electricity prices are extremely high, the grid is unreliable, and/or the solar resource is extremely good. That's beginning to be true in Hawaii, but that's about the only place in the US.

off grid
Off-grid livin'.

Backup power

The notion of keeping the electricity on during power outages or natural disasters plays a big role in the popular conception of batteries, an impression encouraged by solar installer marketing.

But there are problems with using batteries this way, the main one being that even an unusually large home battery pack would at best provide two or three hours of peak electricity use in typical US homes.

It is certainly better to have a modest stream of power, enough to keep the refrigerator cold and charge a few cellphones, than to have none. But aspiring solar+storage customers should keep their expectations on this score in check.

Consuming more clean energy

Solar is zero-carbon, while most utility grid energy is far from it. Some customers just don't want dirty energy on their conscience, so they get batteries to help them consume more of their self-generated zero-carbon energy. (This doesn't make much sense, honestly, but people are funny.)

Demand management

Here we get into stuff that can be useful for utilities in managing the grid.

A distributed network of home batteries amounts to electricity demand that can be controlled by utilities — "dispatchable" demand, in the argot. That can help the utility smooth out fluctuations and spikes in demand and ease grid congestion, reducing the need for peaker plants and power lines.

This can potentially benefit the homeowner as well. She can sign up with a "demand response aggregator," who will control the batteries of dozens or hundreds of homeowners and businesses in order to bundle the controllable demand and sell it in wholesale power markets.

(The ability of demand response to participate in wholesale markets was recently affirmed by the Supreme Court. It's way sexier than it sounds.)

Virtual power plant

The usual problem with solar is that it's not dispatchable. The sun is shining, or it isn't; panels are producing electricity, or they aren't. Utilities can't control it or rely on it.

But if solar panels can either send electricity to the grid or store it and send it later, and if you aggregate a bunch of those solar+storage systems together under a single controller, then you have something much more like dispatchable power. You have a "virtual power plant," which, again, can bid into wholesale markets.

Load leveling

I mentioned this above, but it's worth emphasizing.

The electricity system is sized to supply the highest projected demand, plus a substantial buffer.

If daily peaks in demand can be "shaved" or leveled down, the system as a whole needs less capacity. This serves to displace "peaker plants" that produce the most expensive electricity in the system.

Solar and storage can both "peak shave" on their own, but together they do it better.

load leveling
Load leveling in action.

Lower transmission and distribution (T&D) costs

Relatedly, by reducing demand, especially peak demand, solar+storage reduces the need for T&D investments and helps lengthen the life of existing T&D infrastructure.

Voltage regulation and spinning reserves

Electricity supply and demand must be matched second by second, with alternating current (AC) frequency held within tight bounds. So grids need power resources capable of "spinning up" quickly in the event of, say, the sudden loss of a power plant.

For a power plant to play this role, it must be kept running at a constant low level, wasting fuel and emissions, and it takes minutes for it to spin up fully, wasting time.

Batteries can do it better: They start from zero and leap into service almost instantaneously.

The economics of solar+storage amount to conditions+cost

Okay, so those are all the value streams potentially produced by solar+storage. That was a pretty long and technical list, so I think we all deserve to pause and enjoy a puppy.

puppy (Shutterstock)

Ah, much better.

Now. With the value streams in mind, we can get our heads around the economics of solar+storage.

For solar+storage at a given place and time, the economics will depend on two big considerations.

The first is: How do the value streams described above add up in a given environment, under a given retail price of electricity, a given market structure, and a given amount of sunlight?

The market structure matters a ton. Only in a few places are all the value streams offered by solar+storage assigned a monetary or market value. In many places, many of them go uncompensated. (For instance, not all areas of the country have wholesale power markets where demand response can be bid in. Not all have "ancillary services" markets where customers can bid in their frequency regulation. Etc.)

In fact, it's safe to say most areas of the US currently lack the market structures that would fully compensate solar+storage for its value.

We'll get into when and where solar+storage is economical in the US in a moment.

The second consideration is: How cheap are solar panels and batteries?

Let's take a quick detour to look into that.

Solar and storage are both getting cheaper, quickly

On this subject, I recommend two posts from energy analyst Ramez Naam, who has done yeoman's work pulling together the latest data and expert opinions.

The first is "How Cheap Can Solar Get? Very Cheap Indeed." It concludes, "If current rates of improvement hold, solar will be incredibly cheap by the time it’s a substantial fraction of the world’s electricity supply."

Naam shows that new solar electricity prices decline roughly 16 percent for every doubling of global solar capacity — that is, solar's learning rate. (Virtually every industrial product shows improvement of this kind.)

Taking that as his baseline, he projects solar costs out through 2035, 20 years out.

future solar costs (Ramez Naam)

Now, of course there are caveats out the wazoo here. It's only a model. It assumes the 16 percent learning rate will hold, and that capacity will continue regularly doubling, though there's no way to know either of those for sure. (An asteroid could hit.)

But it's not a wacky projection. It just shows what will happen if what's been happening keeps happening.

If this holds, [within 20 years] solar will cost less than half what new coal or natural gas electricity cost, even without factoring in the cost of air pollution and carbon pollution emitted by fossil fuel power plants.

As crazy as this projection sounds, it’s not unique. IEA, in one of its scenarios, projects 4 cent per kwh solar by mid century.

Fraunhofer ISE goes farther, predicting solar as cheap as 2 euro cents per kwh in the sunniest parts of Europe by 2050.

Estimates for solar's future costs range widely. But it's worth remembering that a) virtually every past projection of solar costs has underestimated how quickly they will fall, b) there's every reason to believe that carbon pollution will be priced in most places in the world within the next 20 years, and c) changes in technology and market structure could easily make it so that a given solar module produces more value for the same cost. All these considerations militate in favor of optimism.

Long story short: Solar PV's current prices, which so strongly color both popular perception and expert analysis of solar power, are ephemeral. Solar is likely going to be the cheapest power source in the world, within the lifetime of geezers like me.

How about storage?

Naam's follow-up post is called, as you'd guess, "How Cheap Can Energy Storage Get? Pretty Darn Cheap."

Lithium-ion batteries, which is what you'll find in electric cars and most home storage, are a little tricker to pin down than solar, because they are earlier in their technology development curve. Some estimates have lithium-ion learning rates at a roughly 15 percent, some have it higher.

Bloomberg New Energy Finance has it at 21.6 percent, remarkably similar to its estimate for solar PV:

cost declines for solar and lithium-ion batteries (BNEF)

So Naam takes 15 percent as the low end, 21 percent as the upper end, and does a similar projection for lithium-ion batteries out to 2035.

battery prices to 2035 (Ramez Naam)

Again, this is just a crude model projection. Anything can happen. But if current trends continue, this is where prices are headed.

Natural gas electricity costs around 7 cents per kWh. Let's say that by drawing some power demand away from peaker plants, solar and storage make it more expensive, around 10 cents.

That means that to beat new natural gas electricity, the per-kWh price of solar and the per-kWh price of storage added together must be lower than 10 cents.

That sounds crazy. But according to Naam's projections, it could happen by 2030 or so — 15 years from now, well within the life of any new power plant.

And it can't be emphasized enough: Solar and storage, especially if they are intelligently bundled, are going to cut into utility revenue in a big way well before they reach such crazy-low prices.

When, though? Let's (finally) turn to that question.

Solar+storage is going to be cheaper than utility power for millions of customers within a decade or two

Last year, the Rocky Mountain Institute released a report on "the economics of grid defection."

It looked closely at five representative metro areas in the US — one each in Hawaii, California, Kentucky, Texas, and New York — and tried to determine when the cost of solar+storage would dip below the cost of utility power, thus, at least in theory, making grid defection a financially viable option.

They modeled four scenarios, one a baseline scenario using "generally accepted cost forecasts for solar and battery systems that can meet 100% of a building’s load," one showing faster technology improvement, one showing faster demand-side improvements in efficiency, and one showing combined improvements.

Here's the base case for commercial systems. (Unfortunately, there's no graph for residential systems; their economics trail a little behind commercial's.) The dotted lines are projected retail electricity costs. The solid lines are the levelized cost of energy (LCOE) from solar+storage. When the solid line dips below the dotted line, that means solar+storage is cheaper than grid electricity and grid defection is, at least in theory, viable.

grid defection base case (RMI)

As you can see, Hawaii is a bit of an outlier. Its retail electricity prices are so high that it's already cost-effective to bail, as many customers already are. But tens of millions of ratepayers in New York and California will be there by 2030 or so.

And under the high-end, optimistic scenario RMI modeled, the moment of "grid parity" is pushed decades earlier — New York and California by 2020, even Kentucky and Texas by 2030.

This is bad, bad news for utilities. RMI highlights just one example:

In the Southwest, across all MWh sold by utilities, for example, our conservative base case shows solar-plus-battery systems undercutting utility retail electricity prices for the most expensive one-fifth of load served in the year 2024; under more aggressive assumptions, off-grid systems prove cheaper than all utility-sold electricity in the region just a decade out from today.

Let that sink in a bit. If technology accelerates faster than (typically conservative) forecasts, all of the US Southwest could find it cheaper to provide their own power with solar+storage than to buy power from their utility within a decade. That's not very far off.

As RMI acknowledged in a follow-up report on "the economics of load defection" (load defection simply means using less utility power, rather than leaving the grid entirely), few people are actually likely to choose to go off-grid, even when it's viable. There are lots and lots of advantages to being connected to the grid. It offers homeowners reliable backup and the ability to participate in potentially remunerative markets (for capacity, demand response, etc.).

But even if they don't go off-grid, lots of customers are going to be a lot less dependent on utility power.

In the second report, RMI took the same five geographies and compared the total costs of three configurations over time — grid only, grid+solar, and grid+solar+batteries — trying to find the lowest overall cost.

It found:

The economically optimal system configuration from the customer’s perspective evolves over time, from grid only in the near term, to grid-plus-solar, to grid-plus-solar-plus-batteries in the longer term. ... Smaller solar-only systems are economic today in three of our five geographies, and will be so for all geographies within a decade. New customers will find solar-plus-battery systems configurations most economic in three of our geographies within the next 10–15 years. [my emphasis]

Here's the rough timeline for residential systems:

economics of load defection (RMI)

(Worth noting: These projections assume no net metering payments at all.)

So to sum up RMI's conclusions: Grid parity, the moment when solar+storage is as cheap or cheaper than utility power, will arrive on different schedules to different places. But in some places, it's coming soon — and no place escapes it for more than a few decades.

A few decades may seem like a long time, but it's much shorter than the life of a power plant. It is well within the time frame in which utilities and regulators need to start planning for it.

Solar+storage is coming, and most of us will live to see it

There are lots of other reports on solar+storage (I'll put a few in "further reading," at the bottom), and while they're not all as optimistic as RMI's, none are pessimistic. They all begin with the acknowledgement that solar+storage systems are hovering on the edge of commercial viability and are inevitably going to grow.

The fact is, most behind-the-meter storage currently deployed on the US grid is heavily subsidized. There's been a lot of public money spent trying to drive storage down the tech curve. Happily, it's working; batteries are within sight of the point that they will no longer need subsidies. Other forms of storage (designed for large-scale applications) are even closer.

Tesla just revealed that the Powerwall 2.0 is coming in 2016, which the company will build at its battery "gigafactory." It already has numerous competitors. SolarCity now sells a solar+storage system, and other solar companies are following its lead.

Solar+storage is likely to take off in the commercial market first, where batteries can help reduce "demand charges" (don't ask). It will spread to residential thereafter.

And remember the value streams. Solar+storage will tend to arrive sooner to places with one or more of:

  • lots of sun
  • high retail electricity prices
  • low grid reliability (or no grid)
  • low net metering (or feed-in tariff) compensation rates
  • high fixed utility bill charges
  • time-of-use pricing
  • access to demand-response, capacity, or ancillary services markets

Some places have lots of these. I mentioned Hawaii earlier. Other islands, or constrained electricity systems, will also be good candidates, as will remote or off-grid areas and states like New York with modernized electricity markets.

Australia (to take one example) is a perfect incubator for solar+storage. According to Morgan Stanley, if home batteries to store solar power reach about a $10,000, 10-year-payback price range, the market could expand to 2.4 million households — half of all households in the country. That would make it a $24 billion industry in Australia alone.

As for the US grid, well, weather-related outages are on the rise, as this NREL report shows:

grid outages (NREL)

Every minute customers spend without power is a minute they'll spend thinking about how to take more control over their energy fate, as it were.

Long story short: Widespread residential solar+storage is coming to the US, and it's coming soon, within the next decade or two. Utilities can't stop it.

And they will be destroyed by it unless they proactively work with regulators to reform the utility business model.

Further reading:

If you were not sufficiently bored by this near-Infinite Jest-length post, there are several other good reports digging into the prospects for solar+storage.

solar and storage (Lux Research)

Learn why climate change is not about saving the planet