On August 26th I published an article that discussed Tesla Motors (TSLA) crushing battery supply constraints in general terms. Since that article generated several hundred angry comments from Tesla fans, I decided to do a deeper dive on existing global manufacturing capacity for the 18650 cells Tesla needs and the principal competitive uses for those cells. The detail in this article is courtesy of my friends at Avicenne Energy, a leading consultant to the lithium-ion battery industry and the sponsor of Batteries 2013, the second largest advanced battery conference in the world.
The venerable 18650 cell Tesla uses in its Model S was introduced in the early '90s and quickly became the industry standard in portable electronics. It was battery industry's answer to the floppy disk. Over the last 20 years the energy storage capacity of a typical 18650 cell increased from 1.1 Ah to a current average of roughly 2.5 Ah (4.5% per year) while cell costs fell from $10 to roughly $2 (8% per year). Manufacturing technology for 18650 cells has been fully optimized over the last two decades and the industry has finally reached a point where raw materials account for over 75% of cell costs, so future savings are highly unlikely.
Since 18650 cells were the de facto industry standard for portable electronics for two decades, cell manufacturers built an immense capacity base. According to Avicenne, the total global manufacturing capacity for 18650 cells is roughly 2.6 billion cells per year, or about 23.4 million kWh of energy storage capacity. The primary users of 18650 cells include:
Communications and computing
Other portable electronics
The 8.4 million kWh difference between global manufacturing capacity and non-automotive demand is the maximum cell supply available to Tesla unless it wants to get into bidding wars with competitive users that are better positioned to pass higher battery costs through to their customers.
The surplus manufacturing capacity is theoretically enough to support up to 100,000 Model S class EVs per year, but a significant share of that surplus capacity belongs to companies that:
- Can't manufacture 18650 cells that meet Tesla's quality standards;
- Can't manufacture high-energy cells that meet Tesla's performance demands; or
- Are otherwise unsuitable cell suppliers for the fledgling automaker.
Currently Tesla sources 100% of its battery cells from Panasonic. It is reportedly in the midst of negotiations with Samsung for cell supplies that will presumably be used in the Model X. Once the surplus capacity of these first tier manufacturers is absorbed, Tesla will find itself facing a difficult choice of either lowering its quality standards or reengineering its battery packs from the ground up to accommodate available substitutes.
This graph from Avicenne tracks global supply, non-automotive demand and surplus manufacturing capacity for 18650 cells over the last six years.
If you focus on the graph for a moment you'll notice two intriguing trends.
First, non-automotive demand for 18650 cells peaked in 2011 as manufacturers of portable electronics shifted to flatter and thinner designs that can't use 18650 cells. To put it bluntly, the battery industry's answer to the floppy disk is rapidly becoming obsolete for upscale users. While the trend to thinner portable electronics will make larger volumes of 18650 cells available in the future, there's no question that 18650 cell architecture, like the floppy disk, is a dead-end technology whose time in the sun has come and gone.
Second, global manufacturing capacity for 18650 cells flat lined in 2011 and 2012 as factories that were planned in mid-decade came online and lithium-ion battery manufacturers came to understand that their future will follow a different path that concentrates on rapidly expanding markets for:
- Prismatic cells - 6.5 million kWh of high-value demand;
- Laminated cells - 8.0 million kWh of high-value demand; and
- Large format automotive grade cells - 3.5 million kWh of high-value demand.
The world's battery manufacturers will no doubt be delighted to sell floppy disks to Tesla for as long as it wants to buy them, but they're not going to build new floppy disk factories. Even if new factories were built, they couldn't offer the same manufacturing cost profile of existing facilities that have already been fully depreciated or written down for accounting purposes.
I can't begin to guess the percentage of the 8.4 million kWh in surplus 18650 cell manufacturing capacity that Tesla will find suitable for its products, but I am certain that it's nowhere near enough for 100,000 cars a year. There's also no way for me to predict the rate of migration away from 18650 cells by other battery users, but even if everybody abandons the 18650 cell format global production capacity isn't likely to expand beyond current levels of 23.4 million kWh per year.
I'm comfortable with the suggestion that Tesla will be able to build 20,000 to 30,000 Model S vehicles per year using available supplies from Panasonic. I'm also relatively comfortable that it will be able to source enough cells from Samsung to support the proposed Model X that's supposed to launch at the end of next year.
I don't see any way for Tesla to source adequate battery supplies for its planned Gen3 vehicle without switching from a floppy disk battery to a more modern architecture and competing head-to-head with other battery users who will only need to pass 10 to 50 watt-hours of battery cost through to consumers instead of passing 40,000 to 50,000 watt-hours of battery cost through to consumers.
Tesla's current market price of roughly $170 per share is based on the widely held assumptions that:
- It will be able to sell every Model S it can make;
- It will be able to sell every Model X it can make; and
- It will be able to build and sell several hundred thousand affordable Gen3 vehicles per year.
While the battery supply chain for the Model S is secure and the supply chain for the Model X is close to secure, there is no chance that the Gen3 vehicle can launch unless Tesla decides to buy modern batteries at competitive prices or somebody builds a "giga-factory … a truly gargantuan [floppy disk] factory of mind-boggling size" to serve Tesla's future needs.
I believed Tesla's stock was dangerously overvalued when it was trading in the $35 price range with a book value of $1 per share. I believe it is psychotically overvalued in the $170 price range with a book value of $5 per share.