The title could also be asking: When will many bullish (NASDAQ:TSLA) analysts projecting millions of EV sales or even an "EV revolution" in the near term finally wake-up and care more about the (battery) supply side?
(For example, the German government is still talking about putting one million EVs on the road by 2020. Many observers express doubt they can realistically hit that target - even when including all hybrid cars.)
The bullish analyst scenario for the EV demand side goes something like this: EV battery prices will fall below the "tipping point" (around 150-200 $/kWh) over the coming years. Beyond this point, EVs will become cheaper than traditional gas or diesel cars (even assuming stable gas prices around the world). Many car buyers will choose EVs or at least PHEVs over "traditional" cars for the first time in their life.
So far so good (I talked about EV mass-market demand challenges in earlier entries already, let's for now assume the demand is there), but what about the supply side needed to produce all these EVs by 2020?
I think battery manufacturing (supply) challenges are still under-reported for TSLA and the EV industry in general - especially compared to possibly overblown short-term challenges these cars are facing (such as recent TSLA Model S fire incidents).*
Why is battery supply and pricing the elephant in the EV room?
Let's assume demand for long-range EVs (let's define this as 200+ miles of range on one charge) takes off in the near future as the battery $/kWh numbers keep falling. Some simple calculations in a scenario:
1. Let's assume 100 million new cars/year will be built by 2020. (This assumes demand in Africa and SE Asia continues to grow with net-new first-time car buyers and a rising middle class. A likely scenario looking at recent real-term purchasing power improvements in many countries. The current global number is around 85 million cars produced per year).
2. Let's assume 80 million of those cars will be new passenger cars (currently this number is at around 65 million cars/year).**
3. Let's assume the $/kWh numbers drop to the 150 range by around 2020 (in six years) and EV demand picks up in the mass-market.***
4. Let's assume 10% of cars sold will now be long-range EVs with 4000 battery cells each. Why 4000? That is using TSLA's current model of small 18650 cells and adding some energy density improvements/cell; for example, a higher Ah number above the current 3.xy, say 4.0 Ah.
5. Result of this scenario: 8 million cars x 4000 cells = 32 billion cells (or the equivalent in larger cell GWh) total will be needed for each car production year - this is only for EV car production; no smartphone or similar device "fighting" for battery supply added (and I assume we still use many of these devices by 2020, probably more so than today).
At the moment, the worldwide production capacity is at a fraction of the GWh number needed for EV batteries by 2020 (across the entire range of Li-Ion batteries), even if just a few of all cars are EVs or hybrids: I used 10% market share for EVs in the scenario above (8 million EVs/year) and didn't even include hybrid battery demand (!).
The "giga battery factory" TSLA is rumored to be planning could (for example) be capable to supply battery packs for 500k cars/year (especially for its mass-market Gen III car).
As you can see in the following video presentation, TSLA's CTO JB Straubel is already talking about 700k TSLA vehicles produced per year by 2019 in a presentation - I doubt Straubel was using exact (and likely) confidential numbers, these could be just rough sales estimates:
(See minute 21:00 and following in the YT video from September '13. Note that JB Straubel is assuming 5000 cells/car in the video, I'm only using 4000 cells/car average based on optimistic technology progress.)
If you don't time to watch the video, here are two important numbers:
40 GWh of capacity will be needed just for those 700k TSLA cars
25-30 GWh is the current global manufacturing limit (all plants)
The 25-30 GWh manufacturing limit includes (for example) the battery for the mobile phone in your pocket. I doubt mobile phones will be banned by 2020 so the world can manufacture more EV batteries ;-)
The YT video is not only an interesting technical overview on TSLA but also straight talk from a man who should know about battery supply - probably more so than any other auto executive apart from very few counterparts at Nissan/AESC, BYD or execs and researchers from dominant battery suppliers like LG Chem, Panasonic or Samsung SDI.
What do these battery supply needs and numbers mean for TSLA (and the rest of the EV makers, I'm using TSLA as a prominent example) ?
TSLA would need funding and break ground on the first battery factory by around 2015 and then plan for a second plant or a major expansion right after the first "giga factory" is finished by around 2017.****
You can do your own cap-ex calculations for these investments (I did some calculations and arrived at billions of $ in earlier Instablog posts).
And these are just the numbers for TSLA's planned car sales. Other manufacturers will have to build up massive battery supply chains as well - so far only Nissan as an EV pioneer has done that (three global battery plants; knowledge gained from a Nissan-NEC JV since 2007).
As all TSLA investors (and analysts) should know: TSLA already is (as of late 2013) one of the largest buyers of Li-Ion battery cells worldwide and it currently produces just 25k or so cars/year.
Admittedly with more cells used for each car - but this is a rounding error given the small production in 2013 and the difference is included in my calculations above (I assumed only 4000 cells/car in the future):
Currently, the largest Model S at 85 kWh still uses around 7000+ 18650-type cells from Panasonic Japan; the numbers will probably be quite similar for the largest Model X SUV due in late 2014. Another indicator of cell numbers will be the lower-range Mercedes B EV being built with TSLA technology - this car is due in the same timeframe.
To reiterate: I'm just assuming about 10% of all passenger cars equipped as long-range EVs in the 2020 scenario above; 90% of cars would still use older/other battery types [let's remember these battery types will compete on the supply chain for some of the same raw materials. So will consumer electronic makers such as AAPL. AAPL and peers are willing to pay (a lot) more per Wh in case supply is tight because of better margins.] or run on ICE, CNG or fuel cells:
Somehow the EV car industry has to supply these 32 billion cells (or the equivalent GWh with larger cells) efficiently and eco-friendly, including improving the raw material supply chain (replace Cobalt...).
You can call me a BEV "naysayer", but in my opinion, it will take 1 or 2 decades to solve this battery supply problem and ramp-up (unless a real breakthrough in battery technology comes along before then...).
In my opinion, the EV "revolution" will be a slow evolution. It will take longer than anticipated on a grand scale (again unless a battery tech breakthrough comes along...) and require more investments from battery suppliers and car makers than most investors anticipate today.
Without giant battery factories being built worldwide at a rapid pace, there won't be enough supply for even 10% of all new cars to be EVs.
The second result of my simple calculation and estimate is that while 10% of cars may be long-range BEVs one day, the rise in absolute production to 100 million new cars/year will probably mean there will be as many conventional cars on the road in 10+ years as today!
That being said I know that continued use of oil resources for personal transportation is also very problematic (hidden military costs/resource wars over oil, oil spills, shortages in conventional oil fields, problems with "fracking" and other exploitation of unconventional oil sources...).
Total pollution numbers may decrease a little since stricter standards for CO2 etc. may be passed. There may be more short-range EVs and hybrids (both mild hybrids and PHEVs) in the total "car population" - but there won't be a big difference in usage and lifecycle dirt assuming the production is rising to 100 million cars/year in total (unfortunately).
Summary On TSLA Factory Challenges:
I'm not arguing TSLA will be unable to build this plant once they raise the funds. But the plant's importance and challenges (also regarding the timeline) are underestimated in my opinion:
- This plant will require billions in investments from TSLA (even if done as a JV) because of the sheer size and risks involved for third-party battery suppliers.
- TSLA will especially need to choose the "correct" battery size and chemistry and ramp-up supply logistics - as well as get Gen III sales estimates right. Critical decisions given the plant's size.
- Quality control needs to be very good (Fisker car fires with A123 batteries and Mitsubishi using GS Yuasa batteries are a dire warning, these were self-ignitions, very different to Model S fire incidents). Panasonic has high quality standards today, this needs to be replicated for 2+ billions cells per year over time in the TSLA factory (about 4 times Panasonic's current volume per year!).
- Most bullish TSLA analysts are in my opinion underestimating the timeframe and investments needed to build his plant, some of them didn't even mention the plant in their reports (but on the other hand the same analysts model in hundreds of thousands of TSLA cars sold/year before 2020, which is of course impossible to achieve without the battery supply !)
Bullish TSLA analysts (and public policy makers around the world) modeling in millions of EV car sales by 2020 in their spreadsheets will need to start thinking harder about the supply side challenges. It's quite simple: no battery supply = no EV built or sold. So far, only a few people are discussing this, even after the latest 10-Q, for ex. here.
Needless to add that the TSLA plant decision could impact the debt structure and/or share count (follow-on offerings) of the company.
- TSLA plant construction will likely need to start in (early to mid) 2015 already since TSLA is still listing "2016-2017" for its Gen III car:
(see second-to-last slide in the TSLA PDF presentation for investors)
Because of all the plant challenges the Gen III car could face introduction delays. Alternative: A possible price hike for the Gen III car and continued third-party battery supply in smaller numbers to keep the "2016-2017" Gen III introduction date intact. (?)
If so, TSLA could push out the need for its plant by 1-2 years. The disadvantage would be a mix-and-match of cells for the Gen III between the first (signature?) series and later model years, this is very problematic to handle technically and also more expensive.*****
There is no easy solution in personal (car) mobility ahead of us and this challenge goes way beyond TSLA and other EV manufacturers.
Maybe we even need to think about creating a soft plateau for a global 100 million new cars/year (2020 scenario) using different incentives:
The hard answers might include taxing road use/more congestion charges, changes in behavior/new car demand (more rental cars and less car ownership, use more substitutes such as bikes, e-bikes with small battery needs), technology improvements ("automated driver-less taxis" on demand in 2020-2030 that you can order 24/7) and - most importantly in my opinion - better public/mass transport (more subways and trams in cities, more local trains to connect suburbs and high-speed trains or similar ground transport between urban areas).
I know most of these ideas are not popular except with maybe a few green activists (and I wouldn't include myself there). But Mega-cities such as London already have car congestion charges, others like Singapore and Shanghai are increasing taxes and road charges as well.
The "personal mobility freedom" mantra for cars (especially commuters with one person/car) may soon be near its peak in these dense urban areas - just look at the rush hour congestion in cities like Los Angeles.
Maybe the real "EV revolution" will happen on two wheels, not four? E-bike sales are climbing fast and only require much smaller batteries.
PS: I recently closed my TSLA short position and no longer hold a position at the time of writing (nor do I intend to open one again until Q4 2013 numbers are announced).
* I think the absolute (fire) incident numbers concerning the TSLA Model S and other EVs so far are too small to draw a good conclusion, especially in comparison to decades of statistics for ICE incidents. We will have to wait longer until more EVs from TSLA and other makers are on the road to compare incidents and miles driven (especially fire incidents). Meanwhile, TSLA is using ICE fire statistics to defend itself.
** Production Statistics for 2012 (also provides link to earlier years): www.oica.net/category/production-statist.../
*** 2020 estimates for larger cells with equally optimistic targets can for example be found below. This includes variables such as replacing Cobalt and other problematic (or very expensive) raw material inputs:
Cobalt is not only much more expensive than any other chemical input, but it also often comes from conflict-scarred regions of Africa. Eliminating the need for cobalt will dramatically lower the battery price.
By 2020, Navigant Research expects that prices for batteries will be below $200 per kWh. At that point, EVs will carry only a small premium over their gasoline counterparts, and battery-based energy storage will be almost as inexpensive as natural gas generation in a peaker plant. When those milestones are achieved, the battery market will become worth more than $75 billion in annual revenue (as opposed to about $12 billion today).
I'm using an optimistic $150/kWh (not $200) because the EV's total cost of ownership advantages may remain a hard sell for some first-time EV buyers (i.e. buyer's irrational aversion to spend a little more for the EV or battery lease vs. ICE cars at the time of purchase - even when saving more on charging vs. gas at the pump later on).
Regarding battery leases and EV battery re-sale values, the outlook doesn't look so bleak: Both EV pioneers Nissan (JV with Sumitomo) and TSLA are thinking about re-using the EV batteries beyond their useful lifecycle in cars, e.g. for domestic or industrial use as "buffers" or recycling the batteries. This increases the battery value for car owners. Other car makers and battery suppliers will likely follow suit and deploy similar usage strategies for spent EV batteries in the future.
**** These estimates of 2 years total lead and plant construction time (until the first batteries are rolling off the production line) are first taken from Nissan's recent battery plant construction in the U.S.:
Start Date: May 2010 (breaking ground on the plant):
End Date: October to December 2012 (first battery packs about to be ready for charging by the end of 2012):
While this Nissan plant could produce up to 200k packs, it's still running well below capacity as of late 2013.
Note that Nissan is using larger cells and is not using smaller 18650-type cells (standard size for many other uses such as laptops, however TSLA may change the size for the Gen III).
Second, it took Panasonic about 1.5 years to create new 18650-type cell manufacturing lines in Asia (several factories in China and Japan).
I'm therefore assuming 24 months needed for a new "giga factory" by TSLA (given the huge capacity and possible cylindrical size changes).
Also of note, Nissan LEAF batteries only hold 24 kWh at the moment (short-range EV). TSLA's batteries will likely require about double that in each battery pack to achieve 200+ miles. A rough estimate: 50 kWh.
***** Note: Whenever I wrote "TSLA plant" above it could either be a plant fully owned by TSLA or a joint-venture (JV) with a large battery supplier.
In both cases, TSLA would (at least in part) vertically integrate its battery supply (similar to Nissan-AESC) including later re-cycling or re-use and sales of older "spent" EV batteries for other customers.