Tesla Battery Investor Day

[dhanson865]

Late 2021... So Elon time strike again. Which comes out first Lucid Air or Plaid S?

You'll have to be more specific

* Lucid Air with no qualifiers is for the ~$80,000 version in 2022
* Lucid Air Touring is supposed to be late 2021
* Lucid Air Grand Touring is supposed to be mid 2021
* Lucid Air Dream Edition is supposed to be early 2021

I can't imagine you mean the "Lucid Air" since that is the slowest trim and wouldn't compete with Plaid Model S in any way.

So if the Dream Edition is what you meant that could come out before Plaid S. Visual reminder of which is which can be seen at Blog - Lucid Air Makes Debut, Deliveries Planned for Spring 2021

 

[HVM]

The first version that they mass-produce and deliver, of course. And no way it is "early 2021"...





(How to get rid of split ends: 8 drugstore products to try)

 

[Buckminster]

Notes from the annual shareholder meeting
(everything I thought was significant in one post, to avoid clutter)

- Shanghai over 1M annual production over time
- "I think the future looks very promising from an annual profitability standpoint"
- Tightening supply chain major factor in achieving consistently positive FCF
- probably 30-40% deliveries growth in 2020 vs. 2019 => 478k-515k deliveries
- Tesla manufacturing injury rate below average, improving quickly
- Autopilot was stuck in local maximum => slow improvement, rewrite necessary

- now labeling in video
- dramatically more advanced software on the car
- step change in capability​

- Texas even faster progress than Berlin

So many slides, I cannot keep up . . .

View attachment 591044
View attachment 591043
View attachment 591042
View attachment 591041

Notes from Battery day:
(everything I thought was significant in one post, to avoid clutter)

### General note: The presentation was extremely information dense, which in my opinion has caused the "nothing new" sentiment that some are expressing here, because there isn't one shiny thing as a takeaway. I hope this is useful in digesting the new info. ###


Presentation:

- Long Term: 10TWh global annual battery production needed for transition to sustainable transport, additional 10TWh/y for energy storage
- Current Tech: 135 Giga Nevada's, $2T investment, 2.8M people needed for this goal => drastic improvement in efficiency needed
- Plan to halve the cost per kWh (!)

- not dependent on a single innovation => no single point of failure
​

- Tabless cell => improved charge rate vs. cell diameter curve

- 5x shorter electrical path
- 20% higher power density due to tabless
- 16% higher range from form factor alone
- 14% $/kWh reduction from form factor+tabless​

- 4680 cell
- Kato Road pilot plant has 10GWh design capacity, to be reached in about a year
- Production plants to be ~200GWh/y
- Dry electrode process

- 10x footprint reduction
- 10x energy reduction
- "close to working" => does work, currrently poor yield​

- 20GWh/y per line, 7x increase per line
- "Tesla is aiming to be the best at manufacturing of any company on earth", manufacturing as long term competitive advantage
- Formation 25% of CAPEX

- 86% CAPEX reduction
- 75% footprint reduction​

- 10x production density increase *across plant*
- 75% CAPEX reduction *across plant*
- Tesla goals:

- 200Gwh/y in 2022, 100GWh internal
- 3TWh/y in 2030​

- Formation+dry electrode: 18% $/kWh reduction
- Raw Silicon anode

- Ion polymer coating, integrate with binder

- design for expansion, don't fight it​

- 20% range increase
- 5% $/kWh reduction​

- Zero Cobalt

- 15% $/kWh reduction on cathode level​

- 3 battery cathode tiers:

- LFP, NMx (33% Manganese, 66% Nickel), Nickel, depending on application​

- Cathode production *very* complicated, in part due to organic process/supply chain growth => potential

- metallic Nickel, no sulfate
- 66% less CAPEX
- 76% less process cost
- no wastewater
- simpler recycling
- 80% less miles travelled
- 12% $/kWh reduction​

- 33% reduction in lithium cost
- Tesla to use Lithium clays in Nevada (rights secured by Tesla), acid free saline extraction

- TWh scale supply secured
- (No Lithium coup in Bolivia?! I'm shocked I tell you, shocked!)​

- 100% cell recycling *today* by third parties, in house recycling starting to ramp up next quarter at Giga Nevada
- Front and rear of the car single-piece cast, connected by a structural battery pack

- no modules
- non cell portion of pack "has negative mass" because of mass savings in other parts of the vehicle
- pack is a "honeycomb stucture between face sheets" => exremely high stiffness, higher than normal car
- better volumetric efficiency => cells more in the center of the vehicle => less chance of cell puncture during side impact
- 10% mass reduction (of pack or vehicle?)
- 14% range improvement opportunity
- much simpler vehicle production​

- Grand Total (POTENTIAL, not currently realized):

- 54% range increase
- 56% less $/kWh (pack level)
- 69% less CAPEX per GWh (cell level)
- start seeing benefits in 12-18 months
- full potential probably achieved in about 3 years​

- Long term: 20M car sales per year
- 3 years from compelling $25k car *at a profit*
- PLAID:

- <2 sec 0-60, <9 sec quarter mile, >1100hp
- "best track time of any production car ever"
- preoders open now, available end of 2021
- 520+mi range, 200mph
- 140k​

- No mention of energy density on cell or pack level, only range increase potential, likely to keep OEMs guessing as to where the range improvements come from exactly (eg cell vs pack level)


Q&A:

- Cell production in Berlin confirmed
- HVAC a "pet project" of Elon
- Direct sales in Texas will "hopefully be cleared up in the future"
- Stationary storage is a 25 year asset or greater => essential for storage cost, environmental concerns
- "could overdo cell production and supply to others", but already going as fast as possible on cell production

- no direct intention of supplying other OEMs, will be done if Tesla has a production capacity surplus (ie if they can scale beyond their needs reasonably)​

- approach to potential Nickel shortage: powertrain efficiency improvements to make LFP viable => limit Nickel consumption
- "presenting a model to other maunfacturers on how to vertically integrate cell production" and battery architechture/chemistry across product stack
- V2G possible through software in Europe, need addtional hardware in the US (due to Plug differences), limited opportunity, want to keep storage and automotive mostly seperate, may still be done at some point, not a priority
- ~3 years until cost/feature parity between Tesla and ICE in the $25k market segment, already there in higher priced segments
- "massive problem[...] need everybody's help[...] It's everyone's planet" (from employee)

​

Someone help me out piecing some of this stuff together:
1. When visiting Giga Berlin, Elon mentioned that the Y produced there (starting next spring or early summer) will be a complete structural redesign, nothing like the Y made in Fremont
2. Elon showed today a new frame design that is centered around the new battery pack with the new, larger, 5x more energy dense cells
3. When asked, Elon confirmed that Tesla will be producing its own cells in Berlin
4. We know from the factory schematics that the casting section at Giga Berlin will include 8 Giga-presses.

Additionally: for those worried about Elon’s slip that not all processes described today are fully running, and that it will still take a while to achieve full optimisation benefits, here’s the thing: they don’t need all the optimisations to get some of the cost benefits! They said as much; there are vehicle frame optimisation benefits, battery cell size benefits, Si anode benefits, Ni-rich cathode benefits, dry electrode benefits, factory size benefits, etc., and most of those are independent of each other, meaning that some of those cost and/or energy efficiency benefits will manifest themselves right away and will be reflected in COGS soon enough.

Model 2 will have close to 300 miles range (per Elon), and it's new $50/kWh bty pack will be an LFP chemistry ("Medium" Range Tesla vehicles, per "Bty Day" presentation).

This is quite profound. Look for all 3/Y SR versions to also switch to Iron bty chemistry (LFP). The cost advantages are undeniable, and the with tabless dry process cells, their performance is as good as current NCA chemistry bty packs.

Did anybody else notice that onstage today Drew said that these new manufacturing process apply equally to Iron battery cells?

It's also important notice that Elon said Iron battery's have 50-60% of the range of nickel batteries. But the total stack of Tesla bty manufacturing innovations add 56% to the range of a vehicle (per the "Bty Day" slide).

I think this settles it: in the future, Tesla Std Rge (SR) vehicles will come with Iron (LFP) based chemistry, and will cost about $50/kWh.

This is amazing progress. First out in $25K Model 2 in 3 years? I think that'll be the Berlin car. Shanghai or a new plant in China would be the logical place to build the Model 1 "World Car".

Cheers!

Wow - what a day.. :-D Soo much information and ground breaking work.

If I were to pick one thing - the shorter electricial path on tabless has to be it.

This is the real bomb in my opinion. And should be way better than a 5x improvement.

A 2170 had a 1meter long anode/catode sandwich - making this tabless = 50cm travel turns into 7cm travel. = theoretical 7x improvement

The new 4680 battery - should have a 2meter (?) sandwich, tab in the middle = tab distance 100cm which turn into 8cm with tabless. = 12.5x improvement.

And the effect on heat management - Just WOW. = No more heat issues!

This was the part which I would have like to get more info on.

What effect does this have on longevity?
What effect does this have on SuperCharger rates?
What effect does this have on cooling needs? Less energy spent on heat management?
Can this be implemented on 2170 batteries from Panasonic at gigafactory?

All batteries used should be tabless asap. No need to stick with tabs.

On the million mile battery, my impression is the big limit on this is the cracking that takes place in parts of the battery. What I saw that addressed this was some part of the battery being layed out in an elastic adhesive. Shouldn't that help the cracking?

 

[Rothgarr]

Shows up for me. $140k.

It is a powertrain option on the Model S.

Order? Yes.

Take delivery of? End of 2021.

Which is underwhelming as hell to be honest.

At the time I checked (immediately after he said it) it wasn't showing up on the Model S order page. Only the Long Range and Performance variants were showing up. I did check again before going to bed and I saw it.

 

[dhanson865]

The first version that they mass-produce and deliver, of course. And no way it is "early 2021"...

* Lucid Air Grand Touring is supposed to be mid 2021
* Lucid Air Dream Edition is supposed to be early 2021

The "dream edition" is limited production so you could easily argue it isn't "mass produced". So even if it comes out in early 2021 it won't meet your criteria. The "grand touring" is the first one they are going to try and produce in larger numbers and that is slated for mid 2021.

I guess really you now need to define how many they have to make in a quarter to equal mass production.

 

[67King]

I have to say, I was blown away. I shouldn't be, but the stuff Tesla does is just amazing. Musk is this century's Henry Ford. He just keeps coming out with completely new things. I say that as a total gearhead engineer who moved to Michigan (from Atlanta) to do engine development for Ford. The complacency in the industry is remarkably strong. One of the reasons left was the sheer unwillingness to push the envelope. I've commented to various friends and acquaintances before that Tesla's lack of automotive experience is not a problem, but is perhaps the best asset it has. The work on the batteries is a HUGE example of that.

Now, as an investor, I'm in for the long haul. And I was joking with my father-in-law who is a huge Porsche snob, and has no qualms insisting that no Tesla can touch the Taycan, that I can about buy a Plaid with how much I've made on TSLA since I bought in in March. I fully expected the stock price to be way up today. And the exact opposite. Which gets into the above comment about not pushing the envelope, and why I think the stock market is such a horrible thing for innovation. So the stock falls because the new technology isn't going to be rolled out at SOB in CA today. I'm sitting here looking at how many problems their design solves - most notably resource constraints. But also addresses what so many seem to think can't be improved, the core battery, itself. More efficient, more energy dense, less expensive, less use of scarce and expensive elements (e.g. cobalt), design complexity, manufacturing complexity, and manufacturing constraints. Throw in superior thermal management, complexity reduction, and improved packaging while you are at it. I mean, up until now, Tesla's secret sauce has been power management, and not battery technology. Well, that just changed. So yeah, the price of the stock necessitates orders of magnitude more sales in the coming decade. Well, a big enabler is what was shown yesterday.

Anyway, I guess I'm just ranting. Maybe looking for folks who see things differently, or are of the same mind. Utterly amazing to me that something so seemingly revolutionary, at least to this dumb ME, is just blown off by the stock market.

Oh, and dhanson865, I've got the white M3P with aftermarket wheels usually around Farragut and Oak Ridge, since you are another local.

 

[bruce4000]

Clear that stock market is all about short term so should not be surprised that stock is down today. It was stated multiple times that their goal is to transition the world to sustainable energy, not just make money like most companies. So why did they outline a bunch of tech Thats not production ready yet? I think they want other autos to see how dead ice will be in 3 to 5 years and get busy (thereby achieve the transition world goal) Also was very technical which appeals to a EE like me but I’m sure went way over most analysts heads. Some of the battery dev may not pan out but things like using the battery as structure is straight forward (and suggested by Sandy Monroe a long time ago) and make ev cheaper than ice which is game changer. So as a long term investor, I’m holding my Tesla because given how far ahead in Ev’s Tesla is and will be, they will be the biggest auto company by units delivered and profits.

 

[Cal1]

Did they ever mention anything about a million mile battery?

 

[bpd1069]

Did they ever mention anything about a million mile battery?

The LFP chemistry is the 'Long Cycle Life' tier for the 25K car, the SR Model 3's and Grid Storage. They never said the phrase, but it's there in the presentation and slides.

 

[Doggydogworld]

My biggest takeway is DBE still doesn't work. It's too early to call it another Silevo, but it's humming a familiar tune. DBE is their main differentiator, they really need to figure out how to scale it.

If tabless is truly easier to manufacture it's a plus. Otherwise they might be better off with 4 tabs. Doubling cell diameter makes it easier to build packs, but it's no breakthrough.

"Structural pack" follows the cell-to-pack craze taking over China, but with a cylindrical twist. Gluing cells to the top and bottom layers of the pack to make a honeycomb is a clever idea. I assume the glue is highly conductive? Or they apply glue around the circumference and use a conductive gel in the center? Enquiring minds want to know!

Silicon anode and cobalt-free cathode seemed like throw-ins. I doubt either are anywhere near production.

They still haven't really warmed up to LFP, which IMHO will dominate the low end and mid-range. And short-haul trucking. And storage (I still can't figure out why it doesn't dominate that already).

10 GWh/year is much higher throughput than I expected for Kato Road, but it sounds like a ~1 GWh line they plan to run at 10x speed. Sounds like Alien Dreadnought / Granny and her walker / bullets from a machine gun / limited by air resistance / so fast you need a strobe light all over again. Except this time Lucy won't pull the football away....

Did they say Model S Plaid will use the 4680s? What about structural pack? That implies a complete redesign of the car, which is overdue, but it looks the same. I wasn't paying much attention by that point and probably missed something, though.

 

[mongo]

My biggest takeway is DBE still doesn't work. It's too early to call it another Silevo, but it's humming a familiar tune. DBE is their main differentiator, they really need to figure out how to scale it.

DBE is working and they have made 10s of thousand of cells. The yield is still low though.

If tabless is truly easier to manufacture it's a plus. Otherwise they might be better off with 4 tabs. Doubling cell diameter makes it easier to build packs, but it's no breakthrough.

Tabless is easier and gives huge benefits in heat generation/ removal, electrical performance, and cell uniformity. Doubling the existing welded on tabs is likely not worth the effort fro the minimal improvement.

"Structural pack" follows the cell-to-pack craze taking over China, but with a cylindrical twist. Gluing cells to the top and bottom layers of the pack to make a honeycomb is a clever idea. I assume the glue is highly conductive? Or they apply glue around the circumference and use a conductive gel in the center? Enquiring minds want to know!

Non electrically conductive adhesive.

Silicon anode and cobalt-free cathode seemed like throw-ins. I doubt either are anywhere near production.

Seems like the silicon goes hand in hand with the DBE, but it could also be a pre-coating step.

They still haven't really warmed up to LFP, which IMHO will dominate the low end and mid-range. And short-haul trucking. And storage (I still can't figure out why it doesn't dominate that already).

They called out using iron in the lower cost cathode, and China is switching to LFP for the SR 3s.

10 GWh/year is much higher throughput than I expected for Kato Road, but it sounds like a ~1 GWh line they plan to run at 10x speed. Sounds like Alien Dreadnought / Granny and her walker / bullets from a machine gun / limited by air resistance / so fast you need a strobe light all over again. Except this time Lucy won't pull the football away....

If the line can make 10GWh, how coul it have been a 1GWh line??? If it can do 10x speed, then that is the real performance.

Did they say Model S Plaid will use the 4680s? What about structural pack? That implies a complete redesign of the car, which is overdue, but it looks the same. I wasn't paying much attention by that point and probably missed something, though.

The protoype Plaid has wider wheel wells. Not much info given on it, but the implication is the new cells/pack will go into it.

 

[JRP3]

I wasn't paying much attention by that point and probably missed something, though.

I'd say you weren't paying attention at any point since you seem to have missed most of the important information.

 

[Ulmo]

I just watched the Tesla Battery Day presentation.

I can’t really add much to what they said, other than that they are doing the obvious things that they should do.

Tesla hiring competitors and also making the cells themselves: best of both worlds, internal improvement & competition and external competition & improvement, improving total market and making total market better for customers. Can also control balance of internal vs external when costs shift. Very agile approach, and very foundational secure approach. A healthier market makes more profit for everyone.

(As I’ve always said, the only good profitable market is one with good competition, not regulatory capture. Regulatory capture causes companies to decline and go out of business; healthy competition weeds out the sick and the rest of the companies do much better than in the sickness of decline.)


It looks like most of this stuff gets implemented in stages as one would always expect: iterate increasing scale from small development to eventual large scale implementation. Current factory construction projects are building whatever current best available tech they know how to build, and future factory construction projects build the best available to install tech. Since everything they talk about in the factory in the presentation is toward reducing factory footprint, upgrades should be possible in existing factories. As each tech is ironed out, they figure out how to make the machine that makes the machine that makes the machine and they make factories less expensive. As Elon said, discovering a better way to do something often takes time, but then in retrospect once you know it you know it (and it seems easier once you already have it, because it is at that point). They keep rolling those improvements into older factories when the cost-benefit analysis is advantageous, and that should be somewhat straightforward according to looking at the input output numbers for existing installations vs. upgrades. Having multiple factories is good for redundancy and space to upgrade some lines while others are up and running with the latest and/or best and/or whatever is needed. They probably get new factory lines at full scale before they slow down older lines for upgrades (whether upgrade or replacement is up to the numbers) as the normal case, but with more factories and development all of this becomes more agile so even that can be adjusted.

I don’t have any main takeaways, other than that they are doing the obvious things that they should do. Everything they said is what one would expect to hear in exactly the industries they are in. One thing that that means is I can finally say that Tesla has broken the mold of decay, destruction, and despair that the old Harvard General Ed teachers hammered starting in the 1940s and 1950s into an entire generation of degenerates that offshored all of USA’s industries and slowed down worldwide innovation, and we are finally waking back up industrially. It helps USA and it helps all the countries that are able to keep up and participate, and we are finally participant in improving the lot of humans again in positive value ways. As with all good quality work performed, we get more out of it than we put in. What this also means is that Tesla will succeed according to the imaginations and work of the younger people, often in excess of the lowered expectations of those older generations. None of this changes physics, but it lets us access and leverage physics to improve us more than the work we put into learning how and doing it.

 

[Fernand]

It's funny. When Musk goes gung-ho circus, they put him down. When he's in techie mode, they don't follow, and develop doubts. When he's super-optimistic, they jump and buy. They think being autonomy-capable means driving your kids to school in downtown SF. When he's realistic and modest, they say "FSD isn't on the shelf", no billion-mile $25k car, and go bear. But they never miss a chance to grumble.

 

[MC3OZ]

Did they ever mention anything about a million mile battery?

They mentioned a 25 year lifetime for energy storage.

The silicon anode surprised me, lower cost and higher energy density, silicon normally means some trade-off in terms of longevity. They have solved that to some extent, but we don't know if the silicon anode currently lasts 1 million miles.

I would not be surprised if energy storage and the million mile battery continue to use a carbon or carbon-silicon anode for sometime.

For DBE low yields in early production is not unusual, Tesla seem confident that they can improve yields.

 

[Buckminster]

It seems that Tesla was only able to announce a Plaid S, because it is a newer more expensive model and it doesn't Osbourne the other versions of the S. However, it would seem likely that if they have to modify the S to handle a different sized battery pack that they would update all the S versions to use the new battery. Otherwise, how will they fit an 80mm battery in a case size designed for 65mm?

So I'm thinking when Plaid is ready to ship, they'll announce newer versions of the LR+ and maybe Performance. Currently Performance is rated 348 miles and LR+ at 402 (15.5% more). If Plaid is good for 520 miles, then a new LR+ with the same pack might also be 15.5% more or 600 miles. They can't announce that now or many people would wait a year for it.

And there is a final problem, if they announce new versions of S with the new batteries, that will probably Osbourne the X models because people would expect those to get upgraded next. So, they'll have to upgrade S & X at the same time.

So, according to some EVTV documentation, a Model S battery module is 3.5 inches (89mm) tall after it is removed from the case, but I can't find any info on the inside height of the case. But that means it is at least possible to fit an 80mm battery inside the existing space, although that doesn't leave much room for the connections and other stuff.

 

[powertoold]

I don't think the new larger cells will provide any benefit on charging speed per kilowatt. Tesla showed a graph where there is a compromise between larger cells and slower charging speeds due to heat.

Essentially, the newer larger cells will be a little bit slower to SuperCharge but provide a wide array of other benefits.

I don't think Tesla sees supercharging speed the limiting factor for electric card adoption. They are probably more focused on supercharger locations and availability.

 

[mongo]

I don't think the new larger cells will provide any benefit on charging speed per kilowatt. Tesla showed a graph where there is a compromise between larger cells and slower charging speeds due to heat.

View attachment 591828

Essentially, the newer larger cells will be a little bit slower to SuperCharge but provide a wide array of other benefits.

I don't think Tesla sees supercharging speed the limiting factor for electric card adoption. They are probably more focused on supercharger locations and availability.

What what??
The bottom red line is the new tabless cell, little to no impact on supercharging speed due to diameter. The graph is normalized to the 21mm current diameter to show change in rate vs diameter. It does not mean the new cell and old cell would charge at the same rate if they were both 21mm.

 

[powertoold]

The bottom red line is the new tabless cell, little to no impact on supercharging speed due to diameter. The graph is normalized to the 21mm current diameter to show change in rate vs diameter. It does not mean the new cell and old cell would charge at the same rate if they were both 21mm.

The line is normalized relative to the existing 2170 cells. My interpretation is that the larger cells have a little bit slower supercharging times compared to the smaller 2170 cells (see the bottom line slightly slope upwards).

For example, it would take a little longer to charger X kWh of the 4680 vs the same X kWh in 2170 cells.

So, there's no supercharging time benefit with the larger cells (my interpretation).

 

[mongo]

The line is normalized relative to the existing 2170 cells. My interpretation is that the larger cells have a little bit slower supercharging times compared to the smaller 2170 cells (see the bottom line slightly slope upwards).

For example, it would take a little longer to charger X kWh of the 4680 vs the same X kWh in 2170 cells.

So, there's no supercharging time benefit with the larger cells (my interpretation).

Maybe I'm arguing the wrong thing.

For the tabless, a larger diameter does not improve superchare rate (on a kW/kWh basis). However, there is no way a 21mm tabless charges at the same rate as a 21mm tabbed. The shorter electrical and thermal path offers too much benefit for that to be the case. (Over the full cycle, if not any specific SOC).