Question for the EV folks here...
Can battery production economically scale to produce enough storage for the entire auto market to transition to BEV?
Background:
My own industry is data storage. The last decade have seen a sea change in data storage as SSDs have come down in price, and due to the many performance/durability/form factor advantages over the traditional HDD, have displaced it in many applications. I think the last I've heard is that notebooks are somewhere 90% SSD market share and only 10% HDD.
This has caused many prognosticators to extrapolate out and based on the cost reductions of the last decade in SSD, to predict that HDD should start writing its own obituary, and sooner rather than later. People saying HDD is dead and the market for HDD will
completely disappear in a matter of a few years. But it's wrong for a very simple reason; NAND flash simply cannot economically scale to meet the cumulative global storage demand.
The truth is that global storage demand is huge, and growing. NAND is an extremely capital-intensive business, prone (like DRAM) to boom/bust cycles as suppliers bring on new fabs or technology / process nodes and flood the market with product, which pushes pricing down to the point where suppliers pare back slightly on CapEx spending to keep in balance, which allows prices to recover and fund more CapEx. And the curves by which this economically occurs don't show any roadmap where enough flash can be produced to meet global storage demand.
The NAND industry could just throw down a couple hundred billion dollars, build a bunch of fabs, and do it. But they'd be forced to sell their product at SO far below cost that they'd all go bankrupt as soon as the floodgates opened. Which is why they don't do it. Or more accurately, they don't have hundreds of billions of dollars to do it and if they tried to borrow it, whoever looked at that business plan would immediately short their stock!
This is something almost nobody talks about. They just look at the price curves and say "HDD is dead, man! Who wants spinning rust anyway!"
Digression over:
I can claim to have some actual expertise in the data storage industry. I confess to having
no expertise in the energy storage industry. But I see major parallels here, so I want to ask people who know this better than I do.
- Fundamentally, every other part of the automotive business scales just fine. Automakers generally are able to produce and sell product, earn profit, and although it's not the most sexy industry, it functions. Globally 70-80M autos are sold each year.
- Globally about 2M BEVs were sold in 2019, putting it as 2-3% of the market. To replace the entire auto market, that means it has to scale by a factor of 20-30, but potentially as much as 40x. (Note: for this analysis I take at face value that BEV will be more reliable than ICEV, and thus simply assume that the service life of a BEV is longer than ICEV to give the benefit of the doubt and say BEV doesn't have to sell the same numbers every year as ICEV to satisfy the market. I do NOT include societal changes from ride-sharing and/or autonomous in that analysis).
- Even at <25% of the BEV market overall, Tesla is an outlier because they produce high-value and high-range cars, i.e. they need a lot of batteries. But at many points over the last few years, they have complained of being battery-constrained in production. They're <0.5% of the global car market, and they're battery-constrained?
- 5 years ago, the BEV market's demand for Lithium Ion batteries was only a tiny portion of the global L-Ion battery production. By 2020 it's over 50% of the L-Ion market demand. By 2030 it's expected that global L-Ion battery demand will increase tenfold, and EV will be almost 90% of all of that, showing that it is basically the ENTIRE driver of that growth. (World Battery Production | Energy Central)
- Even then, a tenfold increase in battery production won't satisfy a 20-30x (or 40x) increase in BEV demand, at least by 2030.
- Mining is where lithium (and the additional metals needed for battery production) comes from. Mining, like NAND flash, is an incredibly capital-intensive industry and producers don't want to pull something out of the ground if it means that they are going to sell it for less than it costs to extract. In addition, the question comes up of how big the world's lithium (and of the other component metals) reserves are, and how hard / expensive it is to get to them?
Now, obviously there are a bunch of things that could affect this. Tesla of course aims for the top-end luxury vehicles with the most range (and largest battery packs). Many BEVs, particularly those in developing economies, might be much smaller commuter vehicles with limited range. But on the other end, we're talking about BEV trucks and semis, which will require HUGE packs. As mentioned, I deliberately avoided mention of autonomous ride-sharing as a new paradigm in transport. Maybe that's a game-changer. Heck, after 2020 maybe WFH becomes much more the default and total vehicle miles traveled goes drops significantly so people worry less about range anxiety and all BEVs become low-mileage vehicles and people rent for long distances.
But I see so many parallels to the SSD/HDD industry and the way it changed, and almost nobody is talking loudly about this in that industry. So I'm wondering why nobody talks about this?
So... From those of you in the know, am I missing something, or is this the dirty little secret that the BEV industry doesn't want people to know because it dampens the story?