Short-Term TSLA Price Movements - 2016

[Yggdrasill]

Anyone notice the powerpack specs?

One significant detail is the "210 kWh (AC)" wording. That means that this is the storage capacity for round-trip AC power, and that the actual DC battery capacity is closer to 220 kWh.

This also fits better with the Powerwall capacity, if we assume the Powerwall uses one pod, and the Powerpack uses 16 pods, the expected capacity for the Powerpack based on the Powerwall capacity would be 14 x 16 = 224 kWh.

Looking at the dimensions, a Powerwall is 1150mm x 755mm x 155mm, while a powerpack is 1014mm x 1254mm x 2192mm. The likely size of each pod would be around 700mm x 700mm x 110 mm. This means a volumetric energy density of around 260 Wh/l on the pod-level, where the old Powerwall had a volumetric energy density around 180 Wh/l on the module level. That's around a 50% improvement. (I'm assuming the new pod only contains one new larger-format module.)

I estimate at most around approximately 35% improvement as a result of a new pack layout and 21-70-cells. Which means that the chemistry must have been improved by *at least* 10%. More probably up to 20%.

That's pretty good. Improvements due to chemistry are basically money in the bank, as the cost per cell is pretty much constant.

EDIT: Actually, looking at the picture below, I think my pod dimensions are a bit off - I think the inverter is included in the pod itself, with only the temperature management outside the pod. The larger format battery module is likely around 700mm x 700mm x 110 mm, while the entire pod is likely around 900mm x 700mm x 110mm. It doesn't affect my calculations though, it just means I'm comparing module with module rather than module with pod.

 

[Sancho]

I love the way TM/EM is setting this up as a win-win for the utility companies. "Hey utilities, back our grand plan and you get more business, not less. The demand for electricity from electric cars will boost your profits." Musk indicated that he projects that 2/3rds of electricity will still be needed from utilities even after TM/SC has succeeded installing solar roofs around the country.

This strategy allows TM/SC to partner with utilities, which in fact is already happening.

Everybody wins here except the fossil fuel extractors and refiners, whose loss can't come quickly enough for the planet's environment and world peace.

Damn you EM, my head is hurting again tonight trying to wrap it around how big your vision is.

 

[JBRR]

Yeh maybe in the future when battery and solar drops to the cost of about 10%-20% as of today. But anyway, I'm not saying Powerwall+Solar Roof is dumb, just reminding people the cost for going 100% off-grid is stupidly high compared to drawing from the grid every now and then. Haven't done any calculations yet, but I think 2 Powerwall + Solar Roof might have you covered for 80% of your total need, but you would need 5 times larger or more to go 100%.

Or add a generator for those 5% outliers and save the 5x.

 

[Mario Kadastik]

I'm seriously surprised at the cost of solar installations you have in the US. Here in Estonia an 11kW system will set you back about 12k€ (ca 13.5k$). That includes the panels, inverter, installation, all the paperwork to become a microproducer i.e. have the ability to sell back to the grid and training. So it's a full package. That means the cost is ca 1.2$ / W. I've always been surprised how you can quote 3+$ / W pricing in the US.

Would be interesting what the full solution will be like here in EU as I need to change the roof soon and I'd love to go for a Tesla solar roof + powerwall (or two).

 

[AntronX]

Anyone notice the powerpack specs?
...This means a volumetric energy density of around 260 Wh/l on the pod-level, where the old Powerwall had a volumetric energy density around 180 Wh/l on the module level. That's around a 50% improvement. (I'm assuming the new pod only contains one new larger-format module.)

I estimate at most around approximately 35% improvement as a result of a new pack layout and 21-70-cells. Which means that the chemistry must have been improved by *at least* 10%. More probably up to 20%...

I have a sneaking suspicion that the old powerwall and powerpack were not really 6.4kWh and 100kWh, but more like 11.2 and 168kWh in reality. This is due to what looks like an old power wall was just one of 15 modules from the power pack. If you find a picture of one of the modules with top cover removed, you notice 2 battery modules that look identical to modules used in Model S. We know those modules are 5.6kWh each... Makes no sense for Tesla to bother with different chemistry cells for the energy storage products. Just cap the cycle depth of existing EV cells to just 57% and they will last over 5000 cycles easy.

So what they could have done now, is pack 25% more cells into power wall and 30% more cells into power pack, then market the new 2.0 products with their actual maximum 100% cycle depth capacity for extra media WOW, but later in the field we will find out that for longest battery cycle life they recommend only cycling 60 - 80% of marketed capacity. Sneaky Elon is sneaky.

 

[Yggdrasill]

I have a sneaking suspicion that the old powerwall and powerpack were not really 6.4kWh and 100kWh, but more like 11.2 and 168kWh in reality. This is due to what looks like an old power wall was just one of 15 modules from the power pack. If you find a picture of one of the modules with top cover removed, you notice 2 battery modules that look identical to modules used in Model S. We know those modules are 5.6kWh each... Makes no sense for Tesla to bother with different chemistry cells for the energy storage products. Just cap the cycle depth of existing EV cells to just 57% and they will last over 5000 cycles easy.

So what they could have done now, is pack 25% more cells into power wall and 30% more cells into power pack, then market the new 2.0 products with their actual maximum 100% cycle depth capacity for extra media WOW, but later in the field we will find out that for longest battery cycle life they recommend only cycling 60 - 80% of marketed capacity. Sneaky Elon is sneaky.

Tesla has confirmed that the Powerwall/Powerpack 1.0 used NMC, rather than NCA like in the vehicles.

And there's no way Tesla would screw their customers in that way.

 

[Matias]

Ok dealing with averages here so this is really rudimentary and back of the envelope.

Avg US electricty consumption ~ 30kwh daily (based on US EIA). This is covered by ~ 2 powerwalls = $11k + $1K (installation) = $12k

Assuming solar roofs can charge 2 powerwalls completely and battery life of 10 years. At steady state, your house can be completely powered by the sun for 10 years.

Avg. electricity rate in US = 12c/kwh. So for 10 years @ 30kwh/month, cost of electricty = $13,000.

Assuming the roofs can generate enough juice, it seems we have now reached a point where solar is at the tipping point with conventional. Further improvements will make solar better than current energy sources.

This seems to be an exciting time to be alive

You did not calculate the price of solarcells.

 

[Gerardf]

My 2 cents...

I saw a comment here questioning why Lyndon Rive had to kick of the presentation. Actually I feel the 'mistake' made is that they did not let Rive talk more. The event yesterday was mainly about a SCTY product (Solar Tiles) with the next-gen TE product announcement as an additional announcement. So Rive should at have been the one showing the first picture of the Solar Tiles, then handing over to Elon to tell the rest of the Solar Tiles and TE story.

The general public and market might now automatically see (and refer to them) the new Solar Tiles as Tesla products. Well, NO. Today these are SCTY products, and they will not be Tesla products until the merger is a fact. In theory the merger could break off and someone will then make a new bid on SCTY.. for sure that bid by other player will now be (much) higher than last time. This is part of the 'no-brainer' that will for sure be emphasized by Elon on November 1st.

With these new tiles solar SCTY (soon Tesla) created AND cornered a premium segment in solar 'panels'. The value of SCTY went up significantly yesterday. Models will need to be adjusted.

This also explains why Elon insisted so strongly that the time for the merger was now, and now in retrospect I agree. Delaying the introduction of the new tiles any longer was no option, it would have leaked. Having other parties going more aggressively after SCTY because of them would have complicated things immensely.

 

[Gerardf]

and 2 more ..

I see calculations (bordering bean-counting) here on economics disconnecting from the net, and if it all makes sense economically.
There are many markets where disconnection (or even PowerWall) simply makes no economic sense. In some markets cycling makes sense, in many it will not. In these markets however protection against power disruption will be the driver, the value of that is not easy to estimate. Just think about counties like India, with a 1B+ people, a strong tech industry depending on stable power, combined with very regular disruptions. With the new pricing their will be many such markets, way to many to serve with just one GF.

At these prices every single PP and PW that Tesla can produce will be sold easily. It will even leave a lot of room for the others to still sell their products. The term "Production constraint" is starting to sound very familiar for Tesla. No need for 'bean-counting' on cycling return, there is a big global market with many use cases that can not wait to order and install these TE storage products.

I think the market and analysts underestimates the 2017 TE revenues and profits by at least by one or two factors of 10. Models will need to be updated.

In fact I am about to change my opinion and now start to think that yesterday, combined by Nov. 1st, just might influence and give a nice push to the short-term SP. But the market can stay irrational much longer of course.

 

[EinSV]

There is a relatively straightforward scenario that fits with all the known evidence and could eliminate the need for a cap raise in Q4 2016 and Q1 of 2017 to fund the Model 3. Applying Occam's razor, this seems to be the most likely explanation.

• Elon said in August that 2016 TE production volumes would be heavily concentrated in November/December and is "really going to go ballistic" after that. Tesla to unveil new ‘Tesla Energy’ products by the end of the year, pushes volume deliveries in Q4
• We now have confirmation of several large TE projects, including one that must be completed by 12/31/2016.
  • This confirms that Tesla is on track for volume production to begin very soon, as Elon predicted in August.
• Reasonable (conservative) estimates of Phase I GF production made by others on this forum are 14-15 GWh/year.
• Little or none of that volume is needed in Q1 for Model 3 and assuming an October ("late") 2017 Model 3 launch, little will be needed in Q2 either.
• 15 GWh x $400/kWh for TE products from GF Phase I equals $6B/year or $1.5B per quarter (assumes some discounting for large projects).
• Tesla can sell a combination of Powerpack plus the Powerwall 2.0 (possibly with integrated inverter/charger that could provide additional margins.)
• Elon has made it clear on multiple occasions that demand is not an issue for TE.

Assuming 40 percent GMs on TE sales results in $1.2B in additional margins in the first half of 2017, or a little less if the full ramp is not achieved by January. SG&A for TE products -- especially Powerpack -- should be relatively low. <snip>

So the bottom line is that if TE ramps up quickly, which it appears on track to do, shorts are toast. Even if Model 3 is delayed, there will be substantial profits from TE that no one seems to expect and have not been priced into the stock. And if Model 3 is on time or early, the tsunami of hurt on shorts is going to be even worse.

Can't wait. I'm getting popcorn.

OK, a couple weeks ago, in the post above I laid out a case for Tesla not needing capital for Model 3 production based on a TE ramp plus cash flow from S/X sales, assuming TE ramped in a serious way in Q1, with a WAG average of $400/kWh for TE products.

We now have actual pricing for Powerwall, including inverter ($393/kWh). Assuming Powerpack pricing is a bit lower on average ($350/kWh) with an average inverter set-up, and guesstimating a 67/33 split of PP/PW results in an average price of $364/kWh or a bit lower than my earlier guesstimate. (For these purposes, I think it makes sense to just use an average per kWh price for PP/PW, including an average inverter set-up.)

$364/kWh x 1MkW/GW x 14GWh/year=$5.1 B/year in revenue, just for GF 1, Phase 1. Tesla's total revenue in Q3 was $2.3B (equivalent to $9.2B annualized) so this represents a revenue increase of more than 50% compared to Q3 2016, on top of any revenue increases from increasing S/X sales and margins.

The actual revenues from TE/GF 1, Phase 1 in 2017 will probably be a little less due to Q1 ramp. Prices will come down eventually but given likely demand I think it is not unreasonable to assume these prices stick at least through Q3 2017.

A WAG assumption of 30% GM results in $1.53B in GM/year, or $382 GM per quarter. Q1 2017 presumably will be lower due to the production ramp. I expect OpEx will be significantly lower for TE products than TA products.

Significantly, I don't think any of this has been factored into analysts' (or shorts'!!) projections.

And the nice thing is that even if the market chooses to downplay this now (tough to do, but nothing is impossible), I think it will be very hard for the market to ignore it much longer since it looks like the "really going to go ballistic" part should start by about January 2017.

Presumably, we will have more color on all this on Tuesday. In the meantime, the numbers above are based on quite a few guesses/assumptions, so anyone who has other thoughts/information feel free to weigh in.

Edit: Just saw @Gerardf's post. While it is worthwhile figuring out the economics of TE for individual customers, I agree with @Gerardf -- it is a foregone conclusion that Tesla will sell every single PP and PW it can possibly make in 2017. Definitely will be production constrained.

 

[Gerardf]

I love the way TM/EM is setting this up as a win-win for the utility companies. "Hey utilities, back our grand plan and you get more business, not less.

Agree.
And I will not be surprised if all the new PW's will have OTA-update capabilities, or already some build in features, to optionally enable the utilities to control the PW's remotely in such a way you can make your PW available to them for load balancing support. For a share of the cost savings as your compensation of course .

Would be one more "Tesla-Network" that the Tesla customers can choose to sign up for in order to make some money with the Tesla product they own.

 

[Matias]

I am pretty sure that Tesla will have two lines of solar products: one for new residential and commercial roofing and another for existing residential and commercial roofing and grid-scale solar. The second line of products can be used for installation on flat roofing.

I think Panasonic+Tesla joint venture will produce conventional solar panels.

 

[KLC13]

You did not calculate the price of solarcells.

I had assumed that solar roofs cost similar to put as regular shingles, so that doesnt affect the cost-benefit...

 

[NOLA_Mike]

Cloudy days don't make your production go to zero, and 2 powerwalls hold enough to serve a full day of usage at 30kWh. In reality, they only need to support the overnight. Day to day usage in typical homes doesn't change that much, maybe +/-20%. Most solar panel based systems installed today leave a lot of roof not generating power. Generate enough excess and it doesn't matter. Bumping up to 3 powerwalls is certainly enough storage for most typical homes to go off grid, especially in solar friendly areas.

Mostly agree.

I could see demand being a problem but then the homeowner might have to practice some common sense - like while off grid at night you can't run an electric dryer, electric oven, stove, electric heat or AC all at the same time. You'd likely need excess capacity of the panels and only charge the car in daytime when the panels are generating juice.

Interesting times, though. The thought of being able to go "off grid" is cool.

 

[hacer]

Here's the back-of-the-napkin math for going off-grid:

Typical house I've lived in averages about 30kWh of usage per day. Tesla recommends 2 Powerwall2.0's for that use case.

That's $12,000 installed. Tesla guarantees them for 10 years, so lets assume for a 20 year period you need to buy a second set (I suspect that the reality is you'll probably make it 20 years without replacing them, but lets err on the side of conservatism.

$24,000 in Powerwalls to support 20 years of power.

A fully off-grid house in the northeast with this usage profile needs somewhere between a 15 and 20kW nameplate capacity solar system.

Current solar costs are about $3/watt installed, so a 20kW system is $60,000, and they're typically targeting a 20 year lifespan.

$84,000 total cost for 20 years off-grid.

That's currently a fair bit more expensive than 20 years of my current hydro bill (which ends up around $37k).

Without knowing the new roofing product pricing, its hard to compare.

There are many rural locations where the land price is much lower because of the need to run utility 5-10 miles to service the property. This "expensive" off-grid solution could be a quite attractive idea for building a new home on such a property. The total costs could be comparable or even cheaper.

 

[NOLA_Mike]

I have an 8kw system that was installed 4+ years ago. I think we would all agree prices have come down since then no? I paid 27k!

For comparison purposes, I had a 9 kW solar system (32 @ 280W LG panels with enphase micro inverters) put in last year and paid $28k installed.

 

[Vitold]

Well, then I suggest you take it to the California Solar Statistics, according to their stats, an 8 kW solar system right now will cost the user $5.28/W on average, or $42k in total. Major flaw in their math.

And if you read my post, you would know I never said this Solar Roof+Powerwall was to help people 100% off the grid. But well, my mind may be easily read by others.

Reported System Costs ($/watt)

The current $/watt data available for California Solar Initiative projects present difficulties when comparing host customer-owned and third-party-owned systems (e.g. leases or power purchase agreements (PPAs)). The reported costs for host customer-owned systems are simple, as they reflect the purchase price inclusive of parts, labor, permitting fees, overhead, and profit. Third-party-owned systems, on the other hand, are reported in a variety of ways, and may also capture costs for additional services.

 

[Vitold]

Here's the back-of-the-napkin math for going off-grid where electricity is cheap:

Typical house I've lived in averages about 30kWh of usage per day. Tesla recommends 2 Powerwall2.0's for that use case.

That's $12,000 installed. Tesla guarantees them for 10 years, so lets assume for a 20 year period you need to buy a second set (I suspect that the reality is you'll probably make it 20 years without replacing them, but lets err on the side of conservatism.

$24,000 in Powerwalls to support 20 years of power.

A fully off-grid house in the northeast with this usage profile needs somewhere between a 15 and 20kW nameplate capacity solar system.

Current solar costs are about $3/watt installed, so a 20kW system is $60,000, and they're typically targeting a 20 year lifespan.

$84,000 total cost for 20 years off-grid.

That's currently a fair bit more expensive than 20 years of my current hydro bill (which ends up around $37k).

Without knowing the new roofing product pricing, its hard to compare.

I fixed it for you.

 

[winfield100]

I'm quite impressed and bullish on what Tesla just presented.

Would anyone knowledgeable mind chiming in on the expected efficiency hit, on average, due to being "stuck" with the orientation of the customer's roof, which may be at a suboptimal angle for most of the day?

as a general answer, most homes are fixed orientation, and not single or dual axis tracking
If you use the cosine of the incident angle of light you get a very close answer

40 years ago when PV was over $100/watt this was more important, now it is way under $1.00/watt

As a mildly extreme example the German Solar Decathlon house from 2009 had 5 of it's 6 sides covered in PV (only not the bottom) and generated ~200% of it's needs

 

[JRP3]

Interesting tweet from Elon for those of us in colder latitudes:

Solar glass tiles can also incorporate heating elements, like rear defroster on a car, to clear roof of snow and keep generating energy

Elon Musk on Twitter


I've always wondered why this idea wasn't implemented in other panels.