Plunge in lithium-ion battery costs | Hacker News Reader

Plunge in lithium-ion battery costs

news.mit.edu

Profound work. Sad to see the MIT PR hacks polluting it with a chartjunk headline image -- spurious shadow of the data line, no tick marks on scales, Y-axis values begging for a log scale but not getting it, superimposed photo of a metal box labeled "Battery".

https://en.wikipedia.org/wiki/Chartjunk

Might be better to link directly to the paper (which this press release does not bother to do):

https://pubs.rsc.org/en/content/articlehtml/2021/ee/d0ee0268...

Here is their better chart: https://pubs.rsc.org/image/article/2021/EE/d0ee02681f/d0ee02...

Looks to me like price decreases are substantially slowing on a log scale.

> Looks to me like price decreases are substantially slowing on a log scale.

Based on my rudimentary understanding of the industry, I wonder how large an impact the expiration of the patents on LiFePo (LFP) cells [1] will have on continuing the drop the in price.

I believe that the fundamental resources involved in LFP production are much lower cost. Other characteristics make LFP seem to be superior for stationary storage which is projected to be larger than mobility.

I hope that the superior economics of LFP, from the POV of production, will help the industry scale at the insane rate that is required.

[1] https://en.wikipedia.org/wiki/Lithium_iron_phosphate_battery

Here is a different page with deep link to the patent details:

https://en.wikipedia.org/wiki/Lithium_iron_phosphate#Intelle...

I hope so. I think about 95% of LFP cells are made in China right now. Good for them (at least on this one issue), but I think we really need more supply diversity, and maybe more non-China companies will start pumping out LFP cells with less licensing issues. It doesn't make much sense for one country to have a monopoly on this one thing that's likely going to be critical to transportation, energy infrastructure, and the world economy in general in the near-term (at least until something better/cheaper than LFP comes along).

> I think about 95% of LFP cells are made in China right now.

This is due to the patent licensing being ignored by the rights holders, but only for products sold in China. That was the deal they made.

That licensing advantage gave Chinese manufacturers a 10 year head start on LFP at scale.

Honestly, it seems like if US hawks had their minds on the future then this would have been a major national security priority - like oil has been for the last century.

It's clear that "national security" was always a euphemism for "the financial interests of certain rich capitalists who not coincidentally spend considerable sums on political lobbying".

Yes, their forecasts for $75 & $20/kWh (e.g. 2027) seem highly optimistic. The rate of decrease for small cells have declined precipitously. Based on the last 10 years of improvement, I'd expect something like $100/kWh in 2030.

However, this doesn't include LiFePO where a lot of manufacturing focus has turned in the last few years. Car maker and fixed battery install are driving demand and lower costs. Looking at $100/kWh next year. and $50 in 2030.

https://climatebiz.com/lifepo4-battery/

Thank you for this. Reading the OP article, I kept wondering when they would get to the numbers. Then it ended.

Looks like about 0.5 decades (factors of 10) per decade.

Still exponential, though, just with a smaller exponent.

Thanks for posting this, I wasn't previously familiar with the term "Chartjunk".

It's probably my favorite thing to rail against Robinhood for - none of their charts have values on the y-axis (or at least they didn't last time I looked), which means their charts are totally useless beyond being dopamine-triggering visuals in a video game played with real money.

It's an illustrative decoration that communicates the key idea.

None of the "chartjunk" (of which there is very little) matters as much as the likelihood that it's fake data.

Real numbers are lacking in this article. It’s way more interesting to talk about real dollar numbers. We are looking at $100/kWh. While most projections by NREL said it would be bottoming out around 200 in 2021. Having a 50% drop compared to projects is insane. At $100/kWh battery connected solar panel power plants become very lucrative.

Are lithium batteries becoming cheaper than lead batteries?

I was under the impression that lead was the leader for fixed installations (like solar) because weight isn't much of an issue if it doesn't have to move.

Is there a different reason why lithium is preferable to lead for these sorts of applications?

Lead for fixed installation was a general rule a few years back but it's not really advantageous anymore. There were several concerns with lithium (especially price, availability and safety) that are now addressed making it much more interesting.

For example a powerwall built with LiFePO4 cells will be good for at least 10000 cycles (+15 years) and afterwards the degradation curve is much nicer than lead-acid so you can easily get 25 years if you can cope with some reduction in capacity. It also has no risk of thermal runoff (so it won't start a fire is a cell has a catastrophic failure) which is much better than traditional cobalt or manganese batteries. You can get grade A cells for about ~400$/kWh today when buying only one and obviously bulk discounts are widely available if you are getting something bigger.

It still burns if your house catches fire though so that's still a concern.

> Is there a different reason why lithium is preferable to lead for these sorts of applications?

AFAIK lead-acid batteries have a number of downsides compared to LiFePO4. They're heavy, they require a ventilated area due to hydrogen gas production, they don't last nearly as long (fewer charge/discharge cycles) and they have significantly lower effective capacity at higher discharge rates.

At this point the main benefit seems to be upfront cost.

https://www.victronenergy.com/blog/2015/03/30/batteries-lith...

https://relionbattery.com/blog/agm-vs-lithium-batteries-whic...

One of the interesting things about battery production cost is that it is very much tied to the price of energy and the cost of the raw materials.

Both battery production and sourcing of the raw materials are energy intensive processes. E.g. refining lithium from either ore or brine, basically involves a long process of evaporating water, concentrating usable quantities, pumping stuff around, etc. So, you could simplify the statement above to the whole process being essentially about energy cost. It also uses water in large quantities and producing and cleaning that boils down to yet more energy intensive processes.

So, as energy gets cheaper, the whole production process gets cheaper. A lot of automotive companies are announcing plans to move their production processes to be carbon neutral. That includes using clean energy for battery production. For many manufacturers that means sourcing their own energy and turning that into a fixed cost rather than a variable cost. As it is such a big component of their overall cost, you can bet they'll work hard to lower that. And as they do, everything energy intensive they do, benefits.

Of course another way batteries can get cheaper is by improving the production process to e.g. require less rare earth materials (e.g. Kobalt), using less water, etc. That too is going to happen. But bringing energy cost down is already big deal. Long term alternative battery chemistry, improved production processes, cheaper energy, all work together to bring cost down further. The question is not if but when. The key metrics here are simply $/kwh. That used to be quite high; it's now around 100$. What happens when it hits 50$ or 25$? How long will it take to get there? These are interesting questions considering that the battery is by far the most expensive component in an electrical car. Also interesting for ICE car manufacturers still hoping to keep on producing ICE cars for another decade or so. Having a lot of cheap EVs to compete with is not going to be good for them.

The price of lithium doesn't dictate the price of batteries, though.

Look at the price of lithium carbonate over the past decade: https://www.statista.com/statistics/606350/battery-grade-lit... vs https://www.statista.com/statistics/883118/global-lithium-io.... It's very hard to see any correlation that would support your thesis. It looks to me that your simplified view seems to miss the truth of the matter.

A correction for you: The element is cobalt, Kobalt is a brand of tools they sell at Lowes home improvement stores. Also I believe most modern lithium battery chemistries omit cobalt due to cost and supply chain issues.

> The element is cobalt, Kobalt is a brand of tools

Kobalt is the name of the element in German; @jillesvangurp's profile says he is based in Berlin, perhaps he also speaks German. (German also capitalizes nouns.)

Actually my Dutch and English are better than my German. But you are right ;-). I've been spending a bit of time in Berlin.

FYI: This is from March 2021 (old news) and is about the large multi-decade decline that anyone following the topic is already well aware of. There are no recent (i.e. this year) developments reported in the article.

I am personally a firm believer that a sub 100$/kwh retail price LFP battery with 10k cycle life will be a reality in < 10 years and change our relationship with energy. I wish there could be modular energy blocks that we can move around portably and charge/ discharge for different loads (not just EV)

Ten years is long. I'm pretty sure that Tesla is already buying them below that price right now for their China operation. Retail prices are of course a different matter. This market is supply constrained so unless you are negotiating hard for large volume orders, it's pretty hard to buy battery cells directly. That indeed might take ten years to get fixed.

Anyway, you can buy some Tesla power wall for your home at quite steep prices right now. Considering the difference in price for that and their actual cost, margins on those might be very lucrative for Tesla right now. They charge as much as 8500$ for a 13.5 KW powerwall. If they source the batteries for 100$ per kwh, that's a nice markup.

At this point the cool kids on the DIY forums have worked out exactly how to build your own powerwall equivalent using LFP cells bought directly from China. It's enough of a savings to make it a worthwhile project.

do you have links for that?

My personal favorite is DIY Solar Forum. I'm more into LFP batteries and solar for my RV than for my house, but here is a direct link to their DIY Powerwall subforum: https://diysolarforum.com/forums/diy-powerwalls.25/

As cool as that sounds in a certain sense, I would be shocked (no pun intended) if doing so wouldn't void one's homeowner's insurance. There's no way in hell I would put so much concentrated energy in an amateur pack anywhere near a home.

I think a lot of people are putting them in metal sheds in their gardens, which somewhat mitigates that risk.

Moved into a metal-rental house where the a prior owner did a lot of his own custom electric and hvac installation.

About half a million in medical bills from that.

So yea, keep it in the shed please.

What does this mean? Previous owner's medical bills? yours?

The cool kids who are a lot braver about their house burning down or local grid problems being blamed on them that I am.

Price is already there. I bought ~89$/kwh LFP batteries last year (shipping brought them over ~$100/kwh though). Some LFP batteries already have a 5000+ cycle count, I'd expect way less than a 10 year outlook.

I read about such prices in articles about EVs but have never seen them in practice in a completely assembled pack.

I think so too but I also think that the groundbreaking research that will underpin that change is not going to be trumpeted in the press ahead of time but instead will quietly show up as incremental improvements to the current processes and materials. The venue for announcement will likely be the WIPO.

I think this is less about ground breaking research and more about simply ramping up volume production to the point where the market isn't supply constrained anymore. With the EV industry basically growing as fast as batteries can be produced, that might take a while. But as that happens, cost will continue to improve for manufacturers and that should start reflecting out in unit prices. The important thing to realize is that production cost and product price are two things.

In a supply constrained industry full of EV manufacturers desperate to buy any batteries at all, the profit margins are going to be insane for battery manufacturers for quite some time to come. So, we might not see lower prices until production volumes start meeting demand. That demand is for existing batteries at their existing prices. A better battery at a higher price would likely do equally well. However, new and improved batteries should drive the prices of the older ones down. So, we might end up with some commodity good enough batteries produced in very high volumes that will actually become affordable at some point.

Let's hope that there will be a major effort to make batteries not only cheaper but cheaper to recycle. Right now we are treating ground up batteries as high grade ore rather than to be able to retrieve their component parts in relatively pure state from the battery itself. There are probably very good reasons for this but it strikes me that we may be going from one environmental problem into another. A few hundred kilos of chemical waste per vehicle is not a great prospect to look forward to.

There is plenty of investment going into recycling. Basically recycling lithium from spent batteries is vastly easier compared to refining it from naturally occurring deposit that only have tiny trace amounts of it when measured in parts per million. That's why lithium is so expensive. There's no shortage of it but refining it from naturally occurring deposits is a very energy intensive process due to the low concentrations of it even in some of the richer deposits. Recovering it from a batteries that actually contain many kilos of it is going to be not a hard cost equation for a wannabe recycler. In the unlikely case that somebody were to actually dump EV batteries in a landfill that would just create a nice business opportunity for other companies. Much better to sell batteries directly to such companies.

But if you are genuinely worried about environmental impact, you should of course be very angry about anything ICE vehicle related where we convert fossil fuels into planet destroying quantities of CO2 without regard to how the oil is mined, how badly it destroys the environment or how many babies the exhaust fumes kill.

Exaggerating a little here but just making the point here that there is a bit of double standards here. When it comes to ICE vehicles, nobody gives a shit where their oil comes, how many children were involved mining the Cobalt involved in removing the sulfur from their diesel/petrol (yes, that's a thing). But as soon as we talk EVs people suddenly get hyper critical about how resources are mined and used and where stuff comes from. It's more than a little hypocritical.

Yes there are some challenges (and business opportunities) but it pales in comparison to the absolute dumpster fire that is the ICE vehicle industry and its apologists.

About the double standard: I'm guessing the cobalt used for replacing the ICE will be an order of magnitude or two more than what has been + will be used for portable electronic devices. Or am I full of balonie ?

Cobalt free batteries are already being produced and the cobalt used in the remaining batteries can be recycled.

It is perfectly possible to see another ecological catastrophe coming and to try to avoid a repeat of past mistakes.

If you want to really help the planet: don't use an ICE don't use a BEV, ride a bike.

If LFP does get to 10_000 cycles usable, that's 50 years at 200 cycles per year.

Cars are going to get much more durable, or it's going to have to be easy to move a pack to a new chassis three or four times. Car manufacturers would probably prefer the latter.

Either that, or we get my imagined rustpunk future, with multi-acre demolition yards full of rusting car bodies piled five high, all wired up and acting as grid storage.

The recycling problem is there, definitely; but it's not overwhelming. And it's much less of a problem than methane emissions and toxic chemical emissions from fracking for oil.

This would be ideal. But car manufacturers would hate to no longer sell fashion and status items but just transportation.

How much waste is left after the "high grade ore" is processed? I thought we were pretty good at extracting metals from stuff.

Anywhere from 5 to 75% depending on who you believe.

https://cen.acs.org/materials/energy-storage/time-serious-re...

Note that this is in flux, more recent articles may give different (hopefully higher) numbers.

“I think this is less about ground breaking research and more about simply ramping up volume production to the point where the market isn't supply constrained anymore.“

These happen hand in hand. It’s analogous to semiconductor improvements that take 1-10 thousand minor improvements (R&D) over the global supply chain during a given year to result in real gains.

I too believe that modular energy blocks are key to price and technology competition. But tech vendors really don't like competition, so I doubt that this will be a reality unless forced by regulators.

Regular batteries began to be standardized in 1928 by American Standards Association (predecessor to ANSI) - according to Wikipedia. We can buy standard batteries in a variety of sizes. I envision a "modular energy block" as something much bigger than these "C" and "D" batteries. But small enough to still be picked up. Perhaps 40 lbs.

So... normal car battery size?

Using Lithium-Ion and adding electronics to make it "modular" - I think it would closer to a cubic foot.

It may be a while until the KWH/kg are high enough to make a removable auxiliary battery that is light enough for the average person to pick up. Right now it looks like Teslas are going about a 4.5 lbs per mile, which is quite heavy. A 30lb battery would barely be worth pulling off the highway to swap. A pair of them wouldn’t get you 15 miles and add a bunch of complexity to your design. In ten years when density has doubled we might be in a very different place.

As the batteries get lighter the cars go farther on the same charge, so maybe we will hit an inflection point sooner than that.

For cars, you would want to automate the swap anyway, so you'd presumably have handling equipment to swap and move batteries between the remove-recharge-store-install steps. I don't think you'd want to use humans in that application.

Until the value of the battery is a tiny fraction of the vehicle, battery swapping will always be a risky proposition.

I as the battery swap station will try to keep the good batteries and swap into someone else's vehicle the terrible batteries. Even if most stations are good actors, it won't take many to destroy the ecosystem by pushing marginal (or worse) batteries into customer vehicles.

Nobody's going to blindly trust a swapping system where the swapped commodity is anywhere between very valuable and toxic waste.

You rent the batteries as a service.

https://insideevs.com/news/537644/nio-4-million-battery-swap...

I didn't spell it out in my top post but this is part of my concern as well. My primary battery is going to have to stay mine due to the market for lemons.

A swappable spare battery feels like something I don't need to own. And my hunch is it should be more tactile versus automated, for that reason and because humans are good at manipulating non-standard objects, machines are not. Probably require less weather proofing too.

But now you're limited by how much an average human can carry for fifteen, thirty feet. Propane tanks are 37 lbs new, and those are hard for many people to wrangle until they've burned off a few pounds of propane. 30 lbs, or maybe 15kg feels like a more workable number. But that's < 9 miles per 'cylinder'. I don't know what the threshold is for 'compelling' but I'm pretty damned sure it's a lot more than 8 miles. 25 miles for a pair might get someone's attention, but that's a few years out.

Or an automated replacement process, which still has other issues we haven't addressed in this thread.

Good points. But, there must be a way to test batteries in an automated way at the replacement station and presumably take them out of service. And, like gasoline distribution, a way to inspect and certify the workings of a replacement station so that the user has good confidence that they are getting a serviceable battery to carry along to the next replacement.

sub 100$/kwh retail price LFP battery with 10k cycle life will be a reality in < 10 years

Look on Alibaba. You can buy that right now.

You might already know this but anything labelled as grade A cells on Alibaba is a scam pure and simple (especially at 100$/kwh). You will get a grade B (used) cell with a somewhat diminished capacity that won't hold for 10k cycles.

To be clear those are usually still great value, but it's worth nothing because people are generally unaware and don't understand that discrepancy between Alibaba prices and what you'd get from a supplier in the US.

I would note that they were specifically talking about "retail price". Alibaba, where (based on experience) you have little recourse with refunds and accountability if the product isn't as described, is not a retail outlet.

Interesting metric. I recently came across this team [1] in Seattle that is building what you mention. Disclaimer: after coming across the team I decided to invest in their equity crowdfunding effort.

[1] https://joulecase.com/

I can see the charm, but they are more expensive per kwh than a powerwall.

However the form factor is a definite winner, along with the stacking. I;m not sure about the integration into solar. But for what it is, a portable power block, its very nice.

On a side note, I would love to see some weather proof alternative to a tesla powerwall. I dislike Tesla as a company, and it pains me to see that they only decent product out there being owned by them

Fascinating company. Not sure I see value proposition though for festival rentals replacing generating power as they aren't completely interchangeable products. Wish them luck though greening up the festival space.

Anker's isn't modular but they do have power devices that overlap with the lower end of their product line.

Allegedly Tesla is already at $130/kwh which part of the reason their margin is so high.

> LFP

I'm so thankful to see that acronym supplanting LifePO4.

Also let's not forget Wright's law:

https://twitter.com/skorusARK/status/1309225390494420996

https://news.ycombinator.com/item?id=24606382

link to the primary article: https://pubs.rsc.org/en/content/articlelanding/2021/ee/d1ee0...

Anyone who’s followed the price of a 12v 100ah battery on Amazon has seen the plunge over the last year. $900 -> $400.

I wonder if some of this is due to covid-driven inflation, possibly in china or other countries?

Covid driven inflation reduced the cost of a battery by $500?

> by $500

First, that would be $300, or a 33% reduction. Let me explain the reasoning.

If batteries are manufactured abroad and that foreign currency experiences sudden inflation, exchange rates for the local currency would make them quite a bit cheaper, especially in a volatile market if sellers are "desperate" to get some (perceived) more stable currency.

I wonder what kind of items that effect would be visible on first, and if that can be used as a tool to look at the market.

Now, this would be combined with regular year-on-year advances in battery tech and price reduction. But if the yearly price drop was higher than usual, I was offering a possible explanation.

Prices would have dropped faster without covid. Shipping is really problematic on LFP cells direct from China right now. Combined with the surge in RV purchases, prices are not going to drop much further this year. But in a couple more years when things have smoothed back out, it's going to be great.

Thank you, I can picture current shipping difficulties. Although, wouldn't that mostly affect big orders and/or guaranteed delivery? Customers on the average Chinese export marketplace are used to long delays (shipping > 1-2 months).

I remember a few years ago reading an article about a breakthrough in Lithium Metal batteries, as opposed to Lithium Ion. Supposedly have 2x the power density at half the size, and don't burst into flames when punctured.

Anyone else remember that? I want those batteries.

Apparently the main issue is that they don't last as long as Li-Ion.

The big competitors to lithium ion right now are LFP (lithium iron phosphate) and LTO (lithium titanate oxide). Neither have quite the power density of current lithium ion cells, but they're not that far off. LFP is quite popular right now for home storage and RVs. It is not nearly as prone to fire as lithium ion. LTO is even safer yet, it usually doesn't even pop the case if you abuse it to the point of destruction. LFP is improving density so it may be viable soon for EVs (in the US; in China they already use LFP), which manufacturers would love. Making the news for fires is really undesirable.

I might be wrong but it looks like neither of those are classified as a Lithium Metal Battery and are still a form of Lithium Ion.

https://en.wikipedia.org/wiki/Lithium_metal_battery

https://en.wikipedia.org/wiki/Lithium-titanate_battery

https://en.wikipedia.org/wiki/Lithium_iron_phosphate_battery

and here's an article about using pure lithium metal as the anode

https://arstechnica.com/cars/2021/11/lithium-metal-hybrid-ba...

You seem to be well informed, is this just hype?

From what you know off the cuff, do LFP or LTO batteries have advantages in number of cycles or discharge depth? Both of those being even marginally higher would make them superior for home batteries, since the power density wouldn't matter nearly as much.

E: Wikipedia says low cost, less rare earth metals and longer lifetime. Sounds like a big technology.

LFP have been popular as drop-in replacements for 12V batteries in motorcycles for years now, too, FWIW. Their cell voltage and resistance to charging damage make them excellent replacements. Their only real deficit is in the cold.

EV trucks are just coming on the market this year. They require at least 100 kwh to exceed 200 mile range. The leading models have 135 kwh. At $300 a kwh an EV truck was too expensive, but not $100 kwh.

Side note: this is from March 2021.

How much of this reduction in cost is due to improvements in the underlying technology (as opposed to a reduction in cost due to increasing the scale of production)?

its a good question, although I wonder whether the original tfa covered it already. In any case I suspect that cost reductions due to scaled production may be easier to model than technological advancements, so its probably the latter which caused projections to be exceeded.

Not trying to be snarky here, but wasn't this Elon Musk's entire thesis on Tesla's success by bringing the massive produced EV (aka Model 3) affordable and all-around better than conventional ICEs?

https://futurism.com/analyst-questions-elon-musks-10-year-ev...

This article was dated in 2017 but I am pretty sure Musk mentioned "economy of scale" will bring down the cost of lithium batteries which in turn make the massively produced EVs like Model 3 more affordable.

https://www.tesla.com/blog/secret-tesla-motors-master-plan-j...

Dated 2006 (not directly mentioning of cost of Lithium batteries though).

What would it take to start a factory in the United States making lithium-ion batteries to sell to consumers (for a PowerWall-like usage)?

Sorry but anything related to energy that has a climate impact should not be measured in dollars, but in co2 equivalent. Good luck doing a CO2 summary of lithium batteries from start to finish when you look at the whole chain and lifetime.

Not to mention lithium mines are those weird big artificial blue lakes which I really don't understand how they really work. Pretty weird there are no documentary on them.

Even mining Uranium does have a CO2 cost you can hardly avoid.

"anything related to energy that has a climate impact should not be measured in dollars, but in co2 equivalent..."

It depends on what you're trying to do with the data. If you're trying to estimate the economic competitiveness, then $ is absolutely the metric more likely to in fact be a predictor of that. For example, will renewables continue to get market share, or is that in danger of being held back by battery cost? If battery cost is falling exponentially, we're in good shape. If battery cost is _not_ falling exponentially, then it could be that we need to explore alternative ways to store the energy. Whether or not $ _should_ be the driving factor, the reality is that it _is_ the better predictor of market uptake.

so if coal is cheaper than renewables, does it matter?

why are you not measuring the co2 emissions of coal?

makes no sense at all.

It's not directly covered in the article but LiFePO4 which are the new popular cells for most things don't require manganese or cobalt so they are much cleaner than older lithium cells.

So I can't directly answer your question, but there is no doubt that CO2 production for lithium cell manufacturing is much lower now than it was in 1990 and it should keep dropping as car manufacturer flip to LiFePO4.

Reply on news.ycombinator.com