As we’re heading down a path of ever more intermittent energy, the need for energy storage will become critical. It’s early stages, but the solutions are becoming clearer: overcapacity of solar & wind generation will help, together with (ever cheaper) lithium-ion / sodium-ion batteries for short-term storage, plus pumped hydro and hydrogen & derivative gases for long-term storage. Various new mid-term storage technologies such as metal-air batteries have a chance to play a supplementary role.
Like many people, I’m interested in understanding the solutions to climate change, and I invest in young companies building solutions in this space through Climate Capital. Energy storage has long been one of the big question marks on my mind. I remember – many years back – watching a Breakthrough Energy Ventures video that suggested solar & wind wouldn’t help us decarbonize sufficiently unless we found some major storage technology breakthrough. Does that still hold true?
To get a sense of where things are heading, I spent many weeks digging through industry reports and academic research. Below are my findings.
Solar & wind as generation backbone
Our global electricity consumption is increasing. This is due to global standards of living improving plus the transition from fossil fuel combustion energy to electricity (e.g. gasoline to electric vehicles and gas heating to heat pumps).
To generate this electricity, the vast majority of new capacity we’re adding nowadays is solar & wind.
And the solar & wind expansion is accelerating: coming from a base of 12% of global generation in 2022, in the next 4 years we’ll likely double our installed capacity (adding another 1600 GW). In a few decades from now, many estimate that the majority of our energy generation will come from solar & wind, increasing perhaps 16-fold in the next 30 years (per DNV).
Beyond solar & wind being renewable, they’re also cheap and expected to get much cheaper still, solar in particular.
So: great? We’re transitioning our economies from fossil fuels to cheap solar and wind energy – a major leap forward. And a fairly unexpected one too: not many were able to predict it a decade ago (historic solar forecasts have been greatly surpassed). It’s a climate transition not relying on nuclear fusion or geothermal breakthroughs, but rather on tons of cheap, mass-produced panels and blades.
Unfortunately solar & wind’s intermittency does leave us with that gnawing problem: energy storage. While demand response (reducing energy consumption at cloudy, windless times) and long distance transmission (sending electricity from sunny, windy to dark, windless places) will help, the silver bullet is cheap, abundant energy storage.
The storage issue: a 20X gap
Across global grids, we have some 176 GW of stored energy generation deployed with 8500+ GWh of storage capacity, enough to power about 5% of the world for 2 days. While this is a good start, it’s a far cry from the 3000+ GW projected to be needed to fully decarbonize by 2050. We need about 20X of what we have today.
And the gap becomes wider when you look at it more closely: today’s storage is almost all (90%+) hydro power. Pumped storage hydropower (PSH) is a great clean storage solution, but its expansion potential is constrained by geography: you need specific mountain & water combinations to make it work, and the best sites have already been used. Estimates are that we can perhaps double or triple our PSH generation in the next 30 years, but it’s unlikely to fill our 20X gap.
The other grid storage technology that provides meaningful capacity today is batteries. These are similar to the ones in our smartphones and electric cars: lithium-ion (Li-ion), of various sorts and mineral compositions. In fact, grid storage is a small part of Li-ion batteries’ use and the vast majority of them are produced for transportation, primarily for electric vehicles (EVs).
Li-ion batteries are good for storing energy at high density, and their costs are going down due to the rapidly increasing production of EVs. Contrary to PSH, they’re best for storing a few hours worth of energy, while storing multiple days’ worth for the grid is still prohibitively expensive. If we want to fulfill our storage needs with Li-ion batteries, we need perhaps 200X the amount of batteries on the grid that we have today. Not only is that a costly path, together with surging EV battery production, it would also put a staggering demand on the mineral supplies needed for these batteries, such as lithium, cobalt, manganese, nickel and graphite, which could become a bottleneck.
Short-term and long-term storage
Before moving to solutions, it’s good to add some nuance on short-term vs. long-term storage. We need the former to bridge a cloudy afternoon and the latter to bridge a dark winter. The requirements for long-term storage are different than for short-term: you need to be able to store incredible amounts of energy, while your generation capacity should stay fixed (saving cost).
As an example, Li-ion scales with storage: 10 batteries both store (measured in kWh) and generate (measured in kW) 10 times as much as 1 battery. Conversely, a hydrogen combustion plant can have 10 or 100 storage cylinders (kWh) while operating only 1 turbine (kW): its storage scales separately from its generation. Those types of scalable storage solutions are ideal for long-term storage.
Solutions
So, as our need for short- & long-term storage will explode as we electrify with solar & wind, the potential to expand hydropower being constrained, and chemical batteries too expensive to cover our long-term needs, how will we store our energy?
1. Overproduction of solar & wind
One solution path is partially sidestepping the storage problem altogether by overbuilding solar & wind capacity. This path bets on solar in particular becoming so cheap that it makes sense to build it in overabundance, such that we have significant electricity overproduction on most days, and just enough on darker days. This significantly reduces the gaps we need to fill with storage. Many experts & analysts (e.g. Stanford, RethinkX) believe this is one of the key answers.
2. Cheaper chemical batteries
Another solution is betting on batteries getting cheap enough, and building them in huge quantities (ideally co-located with solar & wind resources). It’s a reasonable bet: like solar & wind, chemical batteries continuously get cheaper as their demand increases (Swanson’s law), driven by EVs.
And to sidestep mineral mining constraints, new battery chemistries can help. Sodium-ion (Na-ion) batteries look particularly promising, using (abundant) sodium as an alternative to lithium and requiring far fewer scarce minerals. They carry about 40% less energy per kilogram, but that’s not much of an issue for stationary grid storage. The good news with Na-ion is that they are starting to be used in small, lower-range EVs in China: if that trend continues, they’ll find their path to scale and a new, low-cost short-term grid storage solution emerges.
Other options in the chemical battery space are based on cheap materials such as iron or carbon. Their downside is that they’re not suitable for EVs, so they’ll need to rely entirely on the demand from utilities, limiting their potential to scale and get cheaper. For those reasons, Li-ion and Na-ion look like the strongest contenders.
3. Tapping into EV batteries
EVs are huge (Li-ion) batteries on wheels: the cheapest Tesla Model 3 has 3.7X the battery capacity of a Tesla Powerwall and could power a house for 2 days. The 26 million EVs on the road today already contain around 1000 GWh of storage and could (in theory) provide some 200 GW of generating capacity: that’s 10X more than the grid batteries we have. And that’s just a fraction of what we’ll have in the years to come: EV production is expected to surge, creating a gigantic pool of battery potential.
Tapping into this pool of existing batteries as a generating capacity is called Vehicle To Grid (V2G) and it is – unfortunately – pretty much non-existent today. It suffers from complex technological and incentive questions as cars, chargers, circuit breakers, utilities, etc. all need to cooperate to make it work. There are also worries around its effects on car battery longevity, though these can be kept at minimal levels, plus the benefits to the grid are large enough that car owners can be easily compensated to upgrade their batteries when needed. Still, we’re starting from zero here: there’s a long road ahead.
A “light” version of V2G – avoiding EV charging when energy is in short supply – is easier to achieve and already takes place in pockets (for example, we only charge our EV during cheap early sun hours and get additional ad hoc rebates for avoiding charging at certain times). Perhaps this simpler path serves as a gateway to true V2G. V2G holds enormous promise, but the ecosystem needs a kick-start or it won’t materialize.
4. Hydrogen storage
Hydrogen is another solution: it’s the big bet for long-term storage. The ideal way to use hydrogen for storage (from a climate perspective) is by running electrolyzers to produce it at times of solar & wind overgeneration, and then using fuel cells or turbines to generate electricity at times when energy is scarce.
A major benefit of hydrogen is that it has multiple purposes: beyond electricity production, it can be burnt for heat in industrial processes (e.g. for steel production) or used as fuel in transportation (e.g. for long distance trucking & shipping). These additional use cases help drive demand and scale, and with that reduce cost. In addition, hydrogen can be converted into derivative energy dense substances such as ammonia and methanol, which are easier to store and transport and open up more markets. Governments around the world have pledged to increase their use and production of clean hydrogen, spurring the industry.
Today clean hydrogen infrastructure is nascent and has hurdles to overcome. First off, electrolyzers are expensive, and running them only part-time (using excess solar & wind energy) makes that expense harder to justify. Luckily, these costs are expected to come down. Second, storing hydrogen is expensive. The best solution for this is to use existing salt caves and depleted natural gas reservoirs. They are a fairly abundant storage opportunity, though it does impose geographical constraints. Third is efficiency: round trip electricity-hydrogen-electricity conversion loses around 60% of energy, making it half as efficient as Li-ion batteries. But when cheap, overabundant intermittent energy is used to produce the hydrogen, lower efficiency looks like less of an issue. Fourth is demand: industrial and transportation use cases still need to ramp up.
Clean hydrogen is still in early stages and has a long road ahead. But given that a large, multi-faceted market for it is building, driven in large part by government decarbonization pledges, it is likely to become one of our largest long-term storage solutions over the next decades.
5. Natural gas + carbon capture
When talking about hydrogen stored in reservoirs, there’s an easy parallel to existing natural gas reservoirs. Of course, natural gas is not a low carbon fuel, producing both fugitive methane and carbon dioxide in combustion. The latter can be significantly reduced through carbon capture, and given the existing infrastructure and scale benefits of natural gas, stored natural gas with carbon capture will likely be a major storage solution for this decade at least – as imperfect as it is.
6. Novel storage technologies
The range of (often new) tech specifically well-suited for mid-term (few days) energy storage is large: compressed air and cryogenic storage, battery types such vanadium flow, iron-air, molten salt and liquid metal, gravity storage (lifting & descending heavy objects), artificial underground pumped hydro reservoirs, heat batteries of various kinds – the list goes on.
The technologies are incredibly diverse (and excitingly innovative) though all with a similar proposition: storing large quantities of energy at low cost, typically using abundant, cheap materials. Early stage investment in these is also increasing:
Many of these look promising and perhaps one (such as metal-air batteries, the current investor favorite) ends up scaling to become a quintessential mid-term energy storage solution over the next decades. They do also face tough hurdles:
- Strong competition: they all have to compete with both one another as well as with larger scale storage solutions (e.g. PSH, Li-ion, hydrogen)
- Lack of outside demand: their market is limited to grid storage so they don’t enjoy economies of scale from larger sectors (such as Li-ion batteries with EVs, or hydrogen with shipping and manufacturing)
Will one or more of these technologies achieve sufficiently low cost and scale to become a primary grid storage provider? The answer may be somewhat more muted: most will fade away, while a few will likely play a supplementary role.
Where it’s heading
In conclusion, my bet is that the future of storage will be dominated by technologies we’re already familiar with today. First off, the overproduction of solar & wind generation, which can significantly reduce the need for storage. Then for short-term storage, Li-ion and close alternatives like Na-ion. For long-term storage, more PSH where feasible, plus hydrogen and derivatives stored in caves & reservoirs. More nascent / smaller scale technologies like metal-air and vanadium flow batteries or compressed and cryogenic air storage will likely find a role for mid-term storage, though it might be a comparatively smaller one.
The path to zero carbon energy generation & storage is becoming clearer. Over the next decade, we need to make it happen.