Hacker News Reader

Comment by baybal2 | Hacker News Reader

Not sure where Boomberg is getting its digits.

Wholesale price for LFP cells in China has been stably below $100 per kwh for like 5 years.

The only problem is getting to those wholesale quantities.

All major LFP cell makers seem to run a cartel to cut off small distributors from preferential pricing. Anything you can buy in small wholesale is at least 2x the price of what companies making EVs are getting.

Once upon a time good offers like these https://item.taobao.com/item.htm?id=599076665397 come up, and then vanish.

View on HN

It's not a cartel or conspiracy but a simple matter of demand being consistently larger than the available production capacity. Right now owning a battery factory is a good business because there are a lot of manufacturers that are not able to secure batteries. Several car manufacturers are being blocked by this or have reduced their production ambitions because of this.

I think people are also confusing manufacturing cost vs. price here. Obviously, those are not the same. The article is mostly talking cost.

The packaged products that e.g. Tesla sells (cars, battery packs for domestic/grid usage, etc.) include more than just the battery cells.

In any case, Bloomberg is being a bit vague. On one hand, they are predicting mid 2020s for the cost to drop below 100$ and on the other hand you have people guestimating that Tesla is close to that right now. You could of course argue average price, i.e. other manufacturers are not there yet.

Demand for batteries is going to continue to increase for a while. Tesla is talking about TW factories now. The giga in giga factory used to refer to GW. That's a good indication of what they are planning to do. At that scale, the average industry cost and their cost are going to be very similar unless others also grow production capacity at the same pace.

> The giga in giga factory used to refer to GW. That's a good indication of what they are planning to do.

Haha, I never thought about that. That's actually a really clever way of quantifying the output of a battery factory. Battery capacity can be measured in watt-hours, so battery factory output could be measured in watt-hours of battery capacity produced over time, and power*time/time just reduces back to power (measured in watts).

Therefore, you could say that a factory which produces an average of 1 Gigawatt-hours of battery capacity every hour is producing 1 Gigawatt of battery capacity. I like it.

And yeah, a TW factory sounds like it'd be operating on a pretty ridiculous scale if it lives up to its name.

The gigafactory produces 20GWh of batteries per year [0], or 0.002GWh/hour. So, its more like a megafactory by your system.

I think it was just meant to mean "a factory which produces 1-999 GWh of batteries per year" - not per hour.

[0] https://www.tesla.com/gigafactory

edit: this newer 2019 source suggests a 35GWh/year as a "theoretical output". Still orders of magnitude away from a GWh/h. https://electrek.co/2019/04/14/tesla-gigafactory-1-battery-c...

Another way to look at it would be to ask "how much power would the factory need to charge all the batteriers as it produces them?" A gigawat factory would need a gigawatt electical connection dedicated to charging. That logic suggests that the largest factories will most certainly not be charging thier batteries.

You don’t sell rechargeable batteries at full power though right? At least I have never got a new iPad at 100%. More like 30%

Since world electricity production is about 3TW, a battery factory with TW capacity would saturate the world’s charging capacity within three years? (In other words, way too big even for the world’s entire battery needs...)

Musk is quoted as saying you need 100 Gigafactories. I think there’s room for error, considering the batteries needed to replace all ICE vehicles with EVs, the amount of utility scale battery storage needed to go 100% renewables, and any storage needed beyond those two primary use cases. GF1 in Sparks is also producing above anticipated capacity.

Only if your goal is to have enough capacity to store a full year's worth of energy at once. Otherwise you'd saturate the market far sooner than that.

I don’t see why that requires a cartel. Even business-to-business, the administrative/legal overhead of smaller orders implies that, for fixed per unit profits, prices of smaller orders go up.

Business-to-consumer, order size gets smaller, and consumer protection laws come into the picture, further increasing prices.

Is there even _anything_ that you can get in small wholesale/retail at prices lower than double of what companies ordering by the shipload pay?

Edit: consumer electronics might be an exception.

I heard plenty of first party accounts.

A salesman from the cell company comes to the maker of electric busses/trucks/LSEVs and says "resales of cells is strictly, strictly, strictly forbidden," and that they will unleash all torts imaginable if they found out.

And this is the prime reason why good offers popping on online auctions don't last long.

In comparison to margins of wholesale distributors of just any other industrial commodities in China, margins of large battery cell sellers look unbelievable.

In comparison, 18650 cells can be easily found in retail at under 125% of the wholesale price in China.

> A salesman from the cell company comes to the maker of electric busses/trucks/LSEVs and says "resales of cells is strictly, strictly, strictly forbidden," and that they will unleash all torts imaginable if they found out.

Because what invariably happens is the automaker rebins the cells based on their own QA and sells the crappy cells as automotive grade instead of sending back to the manufacturer. Next thing you know, Joe Hacker's total electric conversion uses all the imperfect cells in one poorly balanced battery and burns down a city block. I don't know if it has resulted in anything dramatic like that yet but most of the surplus cells I've bought on eBay that looked like they came from automotive stock were subtly out of spec, enough to be dangerous in high power/capacity configurations.

> In comparison, 18650 cells can be easily found in retail at under 125% of the wholesale price in China.

China doesn't have much in the way of consumer protections or a tort system that's as accessible to the general public as in the US/EU.

Next thing you know, Joe Hacker's total electric conversion uses all the imperfect cells in one poorly balanced battery and burns down a city block.

And you think manufacturers in China give a shit if that happens?

Yes, if they are the well-known large manufacturers, they certainly do to an extend. But mainly it probably depends on the deal the car manufacturers are getting. Supply prices for things not priced by open markets, are a very big secret. So the manufacturers might give car factory especially good conditions, which they don't want to disrupt the market entirely. Which is fine, as long as the batteries all end up in cars, but they don't want the car manufacturers become battery sellers and disrupt the market. And coming back to safety: they batteries might be specially configured for cars, where proper packaging and electronic control can be guaranteed, but you don't want to sell these configurations to non-manufacturers.

I would hope auto companies buying batteries would pay attention to the number of house burned down by batteries from that supplier, yes.

Its not the auto companies restricting resale

No, which is why I think they can sell them so cheap. They're largely immune to the consequences so they don't have to invest in risk mitigation which tends to be expensive.

If they are knowingly reselling dangerous out-of-spec parts, shouldn't that be an issue for law enforcement?

Usually when a corporation is involved, criminal enforcement goes out the window. That aside, it would be an issue for regulators if they were marketing to consumers, but the "spec" in this case is usually part of the vendor contract and isn't legally binding except between the original supplier and customer. Putting up a bunch of rebinned "automotive grade" battery cells on eBay doesn't count because for many items it is assumed by default that the customer is qualified to use the equipment and raw battery cells fall squarely into that category.

You can still do criminal enforcement against individuals working for the company.

I think the problem may stem from the economics of winding machines. They're a huge investment, and require comparable investments in clean room space, and it's hard to increase capacity.

If Dell or Tesla is going to commit to a large number of cells per quarter, Panasonic or Samsung or whoever can install and operate that many winding machines (or, in Tesla's case, install an entire factory). If somebody wants less, they're buying excess stock that's not earmarked and it makes sense to pay more for that.

> I think the problem may stem from the economics of winding machines. They're a huge investment, and require comparable investments in clean room space, and it's hard to increase capacity.

Biggest LFP cell makers have many factories the size of Elon's ones, and they all quietly omit their output figures for non-consumer cell businesses because... taxmen.

LFP manufacturing has been ridiculously profitable from 2010 to 2015, so even hand made cell shops were making millionaires.

Now, the party is over. Big manufacturers have crashed the margins using giant state handouts.

You could get a turnkey LFP cell line running in a garage for around $2600k-$3000k, but the time of good margins is over.

To make LFP under $100 per kwh, and still make money, you need 20-30+ lines, but with such volumes the only buyers will be EV companies, and we are back to square one with this.

You can still operate on garage scale, and try push your cells on online auctions, like many, many small makers still do, and live off small bites of distributor margins.

But you will be doing this knowing that you truly live in the shadow of biggest manufacturers, who, or whose distributors will start poaching your clients once you get big enough for them to get wind of you.

> All major LFP cell makers seem to run a cartel to cut off small distributors from preferential pricing. Anything you can buy in small wholesale/retail is at least 2x the price of what companies making EVs are getting.

That sounds like an easy way to make money. Buy large quantities, sell them in smaller batches at only 1.5x the price.

> That sounds like an easy way to make money. Buy large quantities, sell them in smaller batches at only 1.5x the price.

They will figure out who you are, and you say bye bye to your EV business.

Why would they do that? What's the incentive?

You could even pay them slightly larger prices and still make a profit.

They want to keep small wholesale prices high. The whole non-EV battery market is still quite, quite sizeable.

There's more than one producer. Wouldn't any one of them benefit from breaking rank?

Pack or cell? These days the $100/KWhr target typically refers to pack level costs, which is also what is supposed to enable price parity with ICE cars.

You're confusing the cost for cells with the cost for packs.

I'm not clear why this should be? Effectively 50% discounts for the largest customers? Market share acquisition strategy?

The largest customers probably have contractually obligated volumes that give the factory the ability to plan capital expenditures better and take less risk. I don’t know about the specifics of this industry, but it doesn’t seem ridiculous, in general. In my industry (speaker wire; closer to pure commodity than batteries), we charge our highest paying customers 2x what we pay the factory.

> contractually obligated volumes that give the factory the ability to plan capital expenditures better and take less risk

Bingo not the clown oh. That's exactly how you get better pricing. That's also how you make sure you get product when supplies get tight.

Where have you ever seen a 50% discount for an industrial commodity?

If the open wholesale price is 2X and the few big customers are paying X, that's equivalent to the ""true"" open market price being 2X and the favoured customers getting a 50% discount.

This isn't unusual in enterprise software sales, but it does seem odd for industrial commodities. What I'm asking is why, if supply is so tight, the favoured customers get favourable prices rather than all prices creeping up.

There are several reasons to give big discounts.

If a customer commits to buying enough over that you expect to break even over the next N years after paying for equipment, labor and other expenses I'll give them that price because I know for sure that I have a factor for N years. I can make a lot of money now selling any extra batteries I make at market price. Even if I can't sell in any more for a profit next year, I keep my employees paid for all of N years (giving me opportunities to find new markets) which is worth a lot in itself.

If I know a customer is well known I might give them a discount just to advertise to others they are my customer.

If a customer is large enough that them leaving would affect my bottom line they get a discount to ensure they stay with me.

If they provide me something other than money in return they get a discount. Auto makers sometimes extend employee discounts to their suppliers employees. Auto makers often do supply management tasks many levels down (my materials are certified not coming from slave labor, and GM did the research to prove it, GM will do it anyway so it isn't a big deal to let me know).

If I expect someone to be a bigger customer I will give them a discount now to ensure they can develop there. If batteries never come down in price from 20 years ago electric cars are non viable so I'll give a discount now because the volume cars represent on a small profit is worth more than all profit at volumes of 20 years ago.

That is just a few honest reasons I can think of now. There are also dishonest reasons that I'll ignore, but sometimes they are real.

I just specced a chip from digikey. going from 1 to 500 units halved the price. going from 500 to 3000 halved the price again. These kinds of price reductions are not uncommon for certain types of manufactured goods. Keeping factories running longer gives better margins. Also higher volume may change margins further up the supply line as well.

Does the discount vary monotonically? E.g. first 500 cost X, next marginal 2500 cost X/2?

Or do you end up in the situation where ordering 3000 costs less than ordering 2500?

I am pretty sure you can end up in the sort of situation where ordering 3000 costs less than ordering 2500 on Digikey, and in my experience the same goes for other major component distributors like Mouser.

It seems to me like it kind of promotes waste since one ends up ordering components that they may have no use for, but maybe it is not that big of a deal since it is common to want to have some amount of spare components when doing a production run. Personally, even if I only need a single unit of some little surface mount resistor or ceramic capacitor I usually buy 100 since the price tends to be like $0.10 each for <10 versus $0.50 for 100, and I can just add the excess to my spare parts collection.

You seem to be talking about battery cell price and the article is talking about battery pack price, which last I checked tends to be at least 30%-50% higher. For EVs it might be more if they also add cutting edge cooling systems, etc.

This was my read too; cell price has long been under $100, pack price not yet.

Reply on news.ycombinator.com