2017 Investor Roundtable:General Discussion

[ValueAnalyst]

I don't know what the ASP will be yet but your estimate looks reasonable to me. I can't predict long-term market share yet. Your GM assumption looks reasonable to me.

Glad we agree on the two most important assumptions.

If you are going to model this, I would recommend that you think through the economics of the charging network if you haven't already. According to one source, diesel fuel for Semis constitutes 39% of the cost versus only about 11% for the cab. The Real Cost of Trucking - Per Mile Operating Cost of a Commercial Truck - TruckersReport.com. In other words, on average fuel costs the operator 3-4 X the cost of the cab on a total cost to own basis. The economics are therefore very different than the car business (even leaving aside the transportation service component) and much more of the cost to the operator and potential profits are currently in the diesel component, which will be replaced by battery plus charging network, some of which could be provided directly or indirectly by Tesla and some of which could be provided by fleet operators.

I agree that Tesla Semi will be a game-changer in part because fueling is such a big part of incremental cost for trucking.

We have been discussing this a bit on the Semi thread but I believe Tesla may set up the Semi charging network as a profit center that replaces current diesel fueling operations, although at much lower costs. Tesla could also build some of the cost of the charging network into the cab price or package deal, which would effectively raise the ASP. My current thinking is that they'll rapidly build the charging network and include part of the cost in a package deal to the owner, with a low charge per kWh for use of the network. This could increase the ASP further. However it is structured, the charging network plus part of the battery cost replaces diesel, at what should be a significant cost savings to the operator.

As a rule of thumb, I don't see Tesla charging for electricity in any meaningful way. The company is driven by its mission statement, and charging for electricity does not align with its mission statement vs. making it essentially free or at-cost.

Even though I agree with you that making the charging stations a profit center would be a good financial move, and even accelerate its mission statement in the longer term, on the face of it, it would be a bad PR move, I think...

Of course, there is also the matter of valuing the autonomy component and how to model that ....

I'm modeling Autopilot/FSD as an add-on to all vehicles for the next few years with gradually declining ASP, but I'm assuming it will be included as standard in all vehicles starting in 2023/24, for the reasons I explained here.

 

[DurandalAI]

 

[Reciprocity]

I have written a lot about FSD in the past and I finally found a video from a company called Drive.ai that shows what I mean by High-Def 3D maps with landmarks and paths. If you watch the screen closely you can see objects on the sides of the road and a pre set path that the car follows throughout the demo:

The size of the tile is not something that I could guess at, but most tiles would be cached locally and should be fairly small because they are mostly 3d point clouds, which are like 3d models with no textures. It's the high quality textures that make 3d models huge, the point clouds are just millions of points with X,Y,Z positioning which can easily be compressed.

Where EAP uses road markings to stay in the lane, this system uses pre-determined paths and uses the vision/radar to avoid anything that would be along that path, such as a UFO landing on the road. These paths would be updated as roads change by people like us driving on them. No need to send out a car with Lidar to make the point clouds. The car uses GPS to find the right tiles to load, then uses points in the tiles to put the car in the 3d space down to 5-10cm accuracy. It does this by looking for landmarks stored in the tiles and positioning based on the orientation to those landmarks. The system constantly updates its position, but does not need GPS again as long as it finds the landmarks in the tiles as it travels. A very good GPS is important, but the system could work without GPS at all as long as it knows what tile you are in and general orientation in that tile. The system can get that information from telemetry from the car and a known position, like your garage.

Why no Lidar? Radar + Vision can replicate what lidar does, I mean Vision is based on light and Radar is used to get ranges. The vision system sees a stop sign and the radar gets the distance. this positions you in the 3d space. This is constantly updated as you move, which dials in your location to 5-10cm. Lidar has some issues with things like fog and rain, that can be solved with multiple lidar at different angels and a ton of computing power, or you can just use Radar + Vision. The other issue I have with combining Lidar with Radar is which system do you trust when they disagree. They are so similar in what they offer on the positive side and a lidar can backup a radar and vice verse, but they cannot be used together. I think Tesla has made the right decision to use machine learning with Vision and not rely as much on radar. Radar is only used for getting distances and not running into things (combined with vision).

 

[doctoxics]

I'm trying to put together a list of upcoming tsla stuff, anything missing in the next 6-ish months?

Tesla semi.

 

[EinSV]

<snip> As a rule of thumb, I don't see Tesla charging for electricity in any meaningful way. The company is driven by its mission statement, and charging for electricity does not align with its mission statement vs. making it essentially free or at-cost.

Even though I agree with you that making the charging stations a profit center would be a good financial move, and even accelerate its mission statement in the longer term, on the face of it, it would be a bad PR move, I think... <snip>.

I completely disagree that making a profit off of a Semi charging network is contrary to Tesla's mission, any more than making a profit off electric vehicles, solar power or storage is. The charging network is a a likely expensive component critical to the success of Tesla Semi, with a significant cash outlay likely needing to be incurred upfront since a strong charging network will be necessary to persuade long-haul truckers to buy the Semi,

In fact, I would argue that if the charging network were to be run as a non-profit or money losing anchor around Tesla Semi's neck, it would slow the transition to sustainable transportation in the trucking business, not accelerate it, and therefore would be contrary to Tesla's mission. So we can agree to disagree about that.

I also don't believe it is a bad PR move. Semi operators should be delighted to save more than 50% on fuel costs. They are businessmen/women and I think they would respect rather than begrudge Tesla earning a profit on charging.

 

[ValueAnalyst]

Tesla does not need to compete on price when there is no competition. IF by some odd chance there is a large volume producer of Self Driving EVs with an awesome world wide charging network. I have zero doubt that every single EAP/FSD feature will eventually be enabled at the prices as we see them today. Certainly they could be 20-30% cheaper on Model 3? Maybe, maybe not. If they truly have a value of $75,000 per year as you claim, then they can charge an arm and a leg for the features. Now having said that, the hardware will eventually become so cheap that they could lower the price, but they wont until they need to compete. Tesla is not a charity, they need profits to fund future expansion.

For the record, I estimate the value of FSD at $50,000 per year over five years, and that's the total value of the feature. The total benefit will have to be divvied up among Tesla, the owner, and the rider (via lower cost rides) for a sustainable network that benefits everyone. I assume ~1/3rd for each for now.

FSD: How Do I Estimate The Value Of Full Self-Driving? - ValueAnalyst | Seeking Alpha

I used to think it was a terrible idea to include the HW2 in every car, but now I have seen the light. At some point in the vehicles life, all those features will be enabled. If not the first owner, then the second or some fleet operator. The car will have to be totaled or a one owner car that doesn't want the feature and owns the car forever. Otherwise, I see every single one be activated at some point. Its like Tesla is going to be shipping $5k in banked profit with ever car. What other car has that?

Agreed. Also, don't forget that including HW2 in every car since October 2016 has enabled Tesla to collect billions of miles of FSD testing data even when the driver doesn't have EAP on.

Elon has indicated before that 10 billion miles of data would be needed for FSD and to get the regulatory process going. I estimate that Tesla will reach that point by the end of 2017, just in time for the demo ride from LA to NY.

 

[Reciprocity]

For the record, I estimate the value of FSD at $50,000 per year over five years, and that's the total value of the feature. The total benefit will have to be divvied up among Tesla, the owner, and the rider (via lower cost rides) for a sustainable network that benefits everyone. I assume ~1/3rd for each for now.

FSD: How Do I Estimate The Value Of Full Self-Driving? - ValueAnalyst | Seeking Alpha

Disagree, more like $50k over 5 years, but still very valuable.

Agreed. Also, don't forget that including HW2 in every car since October 2016 has enabled Tesla to collect billions of miles of FSD testing data even when the driver doesn't have EAP on.

Elon has indicated before that 10 billion miles of data would be needed for FSD and to get the regulatory process going. I estimate that Tesla will reach that point by the end of 2017, just in time for the demo ride from LA to NY.

The data is valuable, but $2k per car is a high price to pay for that data. They could do what others are doing and buy the data, but I think the real value is what they are going to do with the data, I dont think its going to be to sell it. Which is of course beat others to FSD and have that drive sales. Has anything thought of what fleet sales would be like with FSD? Could you see a special stripped own Model 3 for fleets?

 

[ValueAnalyst]

I completely disagree that making a profit off of a Semi charging network is contrary to Tesla's mission, any more than making a profit off electric vehicles, solar power or storage is. The charging network is a a likely expensive component critical to the success of Tesla Semi, with a significant cash outlay likely needing to be incurred upfront since a strong charging network will be necessary to persuade long-haul truckers to buy the Semi,

In fact, I would argue that if the charging network were to be run as a non-profit or money losing anchor around Tesla Semi's neck, it would slow the transition to sustainable transportation in the trucking business, not accelerate it, and therefore would be contrary to Tesla's mission. So we can agree to disagree about that.

I also don't believe it is a bad PR move. Semi operators should be delighted to save more than 50% on fuel costs. They are businessmen/women and I think they would respect rather than begrudge Tesla earning a profit on charging.

Alright let me say a couple more things and we can agree to disagree, and revisit once we have more data/guidance from management.

I think you may be overestimating the cost of a Semi charging network. Tesla really needs to cover just the major highways/routes that are heavily used by truckers, which it kinda already has. Semi charging network buildout will be easier than building Superchargers in cities, but the latter is for Model 3/Y.

Secondly, I think Tesla would rather build the cost of charging network to the ASP, which is a one-time cash layout for customers, whereas paying more than at-cost every trip may be seen as less favorable. With solar panels at chargers being rolled out, the incremental cost of electricity to Tesla will really be nominal, and I can see them taking advantage of every opportunity to build customer loyalty.

Let me put it this way: what is different about Semi charger network that Tesla would make it a profit center, if it did not pursue that route with Superchargers so far?

As I indicated before, I agree with you that making the charging network would actually accelerate Tesla's mission, but so would charging $5,000 more for Model 3 since demand is multiple times of supply, but they don't do that. They haven't increased Model S or Model X ASP's either when demand was so much more than supply in previous years. That's just not the Tesla way...

 

[EinSV]

I completely disagree that making a profit off of a Semi charging network is contrary to Tesla's mission, any more than making a profit off electric vehicles, solar power or storage is. The charging network is a a likely expensive component critical to the success of Tesla Semi, with a significant cash outlay likely needing to be incurred upfront since a strong charging network will be necessary to persuade long-haul truckers to buy the Semi,

In fact, I would argue that if the charging network were to be run as a non-profit or money losing anchor around Tesla Semi's neck, it would slow the transition to sustainable transportation in the trucking business, not accelerate it, and therefore would be contrary to Tesla's mission. So we can agree to disagree about that.

I also don't believe it is a bad PR move. Semi operators should be delighted to save more than 50% on fuel costs. They are businessmen/women and I think they would respect rather than begrudge Tesla earning a profit on charging.

Incidentally, @ValueAnalyst, insideEVs estimated that the battery size for a Semi would be 1200kWh for a 600 mile range. This sounds reasonable. to me, although it is still speculative. Tesla Semi Truck Battery is How Big?

When you are counting up GFs, assuming 150GWh per GF (which is what I assume), this is one GF per 100K Semis, plus a little storage. You can go through a LOT of GFs in the Semi business, even if you assume GFs will produce more than 150GWh of batteries.

For example, to match the 415K truck output of Daimler (current market leader), you would need about 4 GFs at current expected capacity, while producing 30GWh of storage per GF.

 

[Reciprocity]

Alright let me say a couple more things and we can agree to disagree, and revisit once we have more data/guidance from management.

I think you may be overestimating the cost of a Semi charging network. Tesla really needs to cover just the major highways/routes that are heavily used by truckers, which it kinda already has. Semi charging network buildout will be easier than building Superchargers in cities, but the latter is for Model 3/Y.

Secondly, I think Tesla would rather build the cost of charging network to the ASP, which is a one-time cash layout for customers, whereas paying more than at-cost every trip may be seen as less favorable. With solar panels at chargers being rolled out, the incremental cost of electricity to Tesla will really be nominal, and I can see them taking advantage of every opportunity to build customer loyalty.

Let me put it this way: what is different about Semi charger network that Tesla would make it a profit center, if it did not pursue that route with Superchargers so far?

As I indicated before, I agree with you that making the charging network would actually accelerate Tesla's mission, but so would charging $5,000 more for Model 3 since demand is multiple times of supply, but they don't do that. They haven't increased Model S or Model X ASP's either when demand was so much more than supply in previous years. That's just not the Tesla way...

I dont see why Tesla would do anything different then they do today with Supercharging. They will charge cost for charging, meaning not take a profit, but not lose money. I would assume they are sugaring the price a bit to cover the build out and expansion of the network. The cost savings from fuel would still be very good.

There is still so much unknown about the Semi. I have ready so much how its not possible to do without a 1200KWh battery pack, but that size of pack would cost nearly $200k retail and that does include a truck. The other issue is the weight of the batteries would cut down on the loads it could carry, because there is an 80,000lbs limit. But you hear JB say that the Semi is basically a scaled up model S, meaning that if the truck is 14000lbs, then its like 3 X S100D. Thats only 300KWh of battery, maybe add another 100KWh for the load. And Elon says its a bunch of Model 3 motors, which makes sense as they are about 300HP each and 6 of the, one for each wheel in a big rig, would be 1800HP and tons of torque which would put it on par for HP and exceeded dramatically the most powerful Semis. Is it really that simple? because they could charge 4 packs today with current Superchargers, they would just need 4 plugs. To much is uknown, I never cared in my life about Semis and now I find myself fascinated and cant wait for this Sept. reveal.

 

[DurandalAI]

VA values the value of FSD at $50k/yr. I don't think that's incorrect for a Semi Truck. If you look at the infographic that I posted earlier, the cost of the driver is 26% of the $180,000 cost. That's $46,800/yr. So it all depends on how much it's used of course, but the advantage of course for FSD is FSD doesn't get tired and can drive without rest stops. So for a truck, I would easily say FSD is worth $50k/yr. Even before FSD is utilized fully, it will reduce accident rates, which will reduce insurance costs.

 

[ValueAnalyst]

Incidentally, @ValueAnalyst, insideEVs estimated that the battery size for a Semi would be 1200kWh for a 600 mile range. This sounds reasonable. to me, although it is still speculative. Tesla Semi Truck Battery is How Big?

When you are counting up GFs, assuming 150GWh per GF (which is what I assume), this is one GF per 100K Semis, plus a little storage. You can go through a LOT of GFs in the Semi business, even if you assume GFs will produce more than 150GWh of batteries.

For example, to match the 415K truck output of Daimler (current market leader), you would need about 4 GFs at current expected capacity, while producing 30GWh of storage per GF.

Agreed.

This is why I switched my DCF from units by model to incorporate kWh and GF guidance. Tesla Semi will surely consume a lot of GF capacity.

 

[EinSV]

Alright let me say a couple more things and we can agree to disagree, and revisit once we have more data/guidance from management.

I think you may be overestimating the cost of a Semi charging network. Tesla really needs to cover just the major highways/routes that are heavily used by truckers, which it kinda already has. Semi charging network buildout will be easier than building Superchargers in cities, but the latter is for Model 3/Y.

Secondly, I think Tesla would rather build the cost of charging network to the ASP, which is a one-time cash layout for customers, whereas paying more than at-cost every trip may be seen as less favorable. With solar panels at chargers being rolled out, the incremental cost of electricity to Tesla will really be nominal, and I can see them taking advantage of every opportunity to build customer loyalty.

Let me put it this way: what is different about Semi charger network that Tesla would make it a profit center, if it did not pursue that route with Superchargers so far?

As I indicated before, I agree with you that making the charging network would actually accelerate Tesla's mission, but so would charging $5,000 more for Model 3 since demand is multiple times of supply, but they don't do that. They haven't increased Model S or Model X ASP's either when demand was so much more than supply in previous years. That's just not the Tesla way...

Just a couple quick points and then I'll move on.

At least five major differences between the Semi charging network and auto charging network that immediately come to mind that favor making the Semi charging network a profit center:

• According to @jhm, an average semi uses 40-50X more fuel per year than an average car. Tesla Semi
  
  
• You mostly charge your Tesla car at home with an occasional visit to the Supercharger. In contrast, charging from the Tesla Semi charging network will be a much higher percentage of charging.
  
  
• Unlike with cars, it will likely be necessary to build a substantial charging network upfront.
  • While you may be right that Tesla could initially get by with a skeletal network but that would slow adoption and impede the business so would not be optimal.
  • This could work in the very short term while they are ramping the business.

• Customers are business people, not consumers.

• The economics should be compelling enough that the enticement of "free" or at cost Supercharging is unnecessary.

While Tesla could and very likely will build in some of the costs into the upfront cost of the vehicle, given the massive energy use by Semis, including all of the costs upfront would encourage overuse/misuse of the charging system and be extremely inefficient. For example, fleet owners may be able to do some of the charging at their own centers at lower cost, and this should be encouraged rather than discouraged. We have already witnessed problems with a few car owners abusing Superchargers and given the much higher energy use by Semis encouraging the same behavior would be problematic.

 

[ValueAnalyst]

I never cared in my life about Semis and now I find myself fascinated and cant wait for this Sept. reveal.

I also notice a lot more Semi's around me these days It's weird how quickly Tesla can change the way I see the world around me.

 

[ValueAnalyst]

Just a couple quick points and then I'll move on.

At least five major differences between the Semi charging network and auto charging network that immediately come to mind that favor making the Semi charging networks a profit center:

• According to @jhm, an average semi uses 40-50X more fuel per year than an average car. Tesla Semi
  
  
• You mostly charge your Tesla car at home with an occasional visit to the Supercharger. In contrast, charging from the Tesla Semi charging network will be a much higher percentage of charging.
  
  
• Unlike with cars, it will likely be necessary to build a substantial charging network upfront.
  • While you may be right that Tesla could initially get by with a skeletal network but that would slow adoption and impede the business so would not be optimal.
  • This could work in the very short term while they are ramping the business.

• Customers are business people, not consumers.

• The economics should be compelling enough that the enticement of "free" or at cost Supercharging is unnecessary.

While Tesla could and very likely will build in some of the costs into the upfront cost of the vehicle, given the massive energy use by Semis, including all of the costs upfront would encourage overuse/misuse of the charging system and be extremely inefficient. For example, fleet owners may be able to do some of the charging at their own centers at lower cost, and this should be encouraged rather than discouraged. We have already witnessed problems with a few car owners abusing Superchargers and given the much higher energy use by Semis encouraging the same behavior would be problematic.

I think you make some good points; especially the one with "charging from the Tesla Semi charging network will be a much higher percentage of charging."

I'll keep any revenue/profit from Semi charging network out of my DCF for now, but I'll work in a placeholder and keep an eye on any developments.

Thank you for improving my understanding of the future.

 

[SwTslaGrl]

Imho, Tesla works with fleet owner on the Semi. They decide to use Tesla Semi for specific routes. The Semi charging network is only needed for these routes, not random build-outs.

 

[EinSV]

I think you make some good points; especially the one with "charging from the Tesla Semi charging network will be a much higher percentage of charging."

I'll keep any revenue/profit from Semi charging network out of my DCF for now, but I'll work in a placeholder and keep an eye on any developments.

Thank you for improving my understanding of the future.

That makes sense to me -- this is all speculative and there are a lot of variables already. We'll know more soon enough.

 

[Hogfighter]

Can we please have a ValueAnalyst thread elsewhere, and return this thread to its original intent?

 

[D-egg-O]

Charging infrastructure should only be required for long haul and some regional haul applications.

I see Tesla working with fleets to install superchargers at their hubs. Utilities will get involved with the fleets and negotiate discounted electricity rates.

 

[D-egg-O]

Utilities will get involved with the fleets and negotiate discounted electricity rates.

This was discussed at the EV Roadmap conference.

Roadmap Conference