Shorting Oil, Hedging Tesla

[TheTalkingMule]

Finally noticed on my Fidelity app that Morgan Stanley upgraded Chevron from $104 to $107 yesterday, hence the 7% share price pop. Kinda strange to upgrade an oil company in the middle of a pandemic 10 days before earnings are reported.

Overall this early oil & gas consolidation seems well received by the investment community. Should have a considerable near term impact on valuation of the oil majors. This phase will likely be seen as a positive for the lines of CVX and XOM(who will probably now follow suit with purchases). "They're getting these great assets for pennies on the dollar!". IMO we'll quickly find out these assets are worthless.....and the debt remains.

Hoping for a CVX near term macro oil market dip to offload my Aug20/Jan21 puts then will wait on the "vaccine skyrocket" to buy January 2022 puts by the boatload.

So far only up a hair on these various puts(even with a couple home runs), but getting better at it. When Chevron popped to $100+ i should've offloaded them all for a 2 week 60% gain rather than just selling half. Greedy!

 

[jhm]

https://twitter.com/jpr007/status/1286311128092434432?s=20
Nice thread on why the economics of battery electric trains will beat that of hydrogen fuel cells. I strongly suspect the same for trucking too.

 

[ammulder]

https://twitter.com/jpr007/status/1286311128092434432?s=20
Nice thread on why the economics of battery electric trains will beat that of hydrogen fuel cells. I strongly suspect the same for trucking too.

That's a great thread, but it's not clear to me that an overall lower cost of battery truck vs. fuel cell truck makes for an easy decision for long-distance trucking. As long as drivers are involved, I think a charge stop of an hour or more is unlikely to be attractive. Boredom aside, my understanding is that drivers are paid by the mile, so a lengthy charge stop is time for which they're unable to be paid -- in other words, it comes directly out of their potential income. I'm not sure where Tesla stands on the charge speed with the megacharger, but it may be that the more it's improved, the more attractive battery trucks get for long-distance routes.

Any autonomy you put in could change that picture, though. Even if you just get a second autonomous truck to convoy with a first driverful one, increasing the driver's rate modestly could make it worth the charge stops. Not to mention if the truck can just drive itself full stop.

 

[jhm]

That's a great thread, but it's not clear to me that an overall lower cost of battery truck vs. fuel cell truck makes for an easy decision for long-distance trucking. As long as drivers are involved, I think a charge stop of an hour or more is unlikely to be attractive. Boredom aside, my understanding is that drivers are paid by the mile, so a lengthy charge stop is time for which they're unable to be paid -- in other words, it comes directly out of their potential income. I'm not sure where Tesla stands on the charge speed with the megacharger, but it may be that the more it's improved, the more attractive battery trucks get for long-distance routes.

Any autonomy you put in could change that picture, though. Even if you just get a second autonomous truck to convoy with a first driverful one, increasing the driver's rate modestly could make it worth the charge stops. Not to mention if the truck can just drive itself full stop.

It's good not to get caught up in potential edge cases where fueling time is economically decisive. Rather I think it is best to think about lower hanging fruit first. (Diesel trucks can handle the extreme cases for a decade or more.) Tesla will put out a semi with 600 mile range or better. This is more than adequate to handle the bulk of use cases with a single charge per day, and so does not impact driver time at all. Just plug it in at the end of your shift. So for the bulk of truck applications, I don't believe that truck driver time costs will be appreciably different.

Ten years from not batteries will be cheaper, have longer range and have longer life, and EV trucking infrastructure will be much more built up. Also logistics companies will have more innovation to optimize the logistics of EV trucking. And yes self-driving tech may be in play too. So the most extreme use cases will look very different 10 years from now. For example, suppose you want to do coast-to-coast transit in minimum time. Even without full autonomy you could have the following system. Have local trucker pick up load and take it to platooning hub. The hub has a stand of charged tractors and human pilots. The trailer is autonomously transfered to a tractor. A platoon of tractors follow pilot to the next hub. Trailers are transfered again. Each trailer proceeds through the network of hub to its destination hub. At this point a local trucker drives it to ultimate destination. So in the whole relay, the load only waits a little time to potentially transfer to a different pilot. Charging time for the tractors happens before a trailer arrives at a hub. The labor cost is also kept to a minimum as platooning enables one pilot driver to move 2 to 5 trucks at once. Eventually full autonomy will make even the pilots unnecessary, but the same sort of relay approach can remove charging time from the transit time of the load.

But stepping back from questions of driver labor cost and charging times, it takes about 3 times as much electricity to power an hydrogen vehicle from electrolyzed hydrogen than it does to power a battery electric vehicle. Nikola is banking on the idea that they can buy 3 times as much power when it is cheaper than an electric truck gets charged. I love the idea of electrolyzer operating 50% of the time when electricity is cheapest. But I think Nikola underestimates the potential of battery electric trucks also to find clever ways to charge when power is cheap. For example having a stand of tractors at a hub means that the stand can charge opportunistically as power is cheapest too. But there are many other clever ways to charge cheap, and needing only a third as much electricity as an electrolyzer makes it not hard to do better. That's the fundamental advantage the battery electric train has as well.

 

[Dr. J]

That's a great thread, but it's not clear to me that an overall lower cost of battery truck vs. fuel cell truck makes for an easy decision for long-distance trucking. As long as drivers are involved, I think a charge stop of an hour or more is unlikely to be attractive. Boredom aside, my understanding is that drivers are paid by the mile, so a lengthy charge stop is time for which they're unable to be paid -- in other words, it comes directly out of their potential income. I'm not sure where Tesla stands on the charge speed with the megacharger, but it may be that the more it's improved, the more attractive battery trucks get for long-distance routes.

Any autonomy you put in could change that picture, though. Even if you just get a second autonomous truck to convoy with a first driverful one, increasing the driver's rate modestly could make it worth the charge stops. Not to mention if the truck can just drive itself full stop.

I find this thread on the other forum instructive as to how the Tesla semi and Megachargers would work with long-haul trucking regulations.

Freightliner eCascadia - impressions from a professional driver - Tesla Owners Online

 

[ammulder]

It's good not to get caught up in potential edge cases where fueling time is economically decisive. Rather I think it is best to think about lower hanging fruit first.

That's fair, so long as we're agreeing with the Nikola position that BEV is better for shorter routes and it's either diesel or hydrogen for longer ones.

But stepping back from questions of driver labor cost and charging times, it takes about 3 times as much electricity to power an hydrogen vehicle from electrolyzed hydrogen than it does to power a battery electric vehicle. Nikola is banking on the idea that they can buy 3 times as much power when it is cheaper than an electric truck gets charged. I love the idea of electrolyzer operating 50% of the time when electricity is cheapest. But I think Nikola underestimates the potential of battery electric trucks also to find clever ways to charge when power is cheap. For example having a stand of tractors at a hub means that the stand can charge opportunistically as power is cheapest too. But there are many other clever ways to charge cheap, and needing only a third as much electricity as an electrolyzer makes it not hard to do better. That's the fundamental advantage the battery electric train has as well.

I'm not sure I agree with this. I don't know a lot about hydrogen filling stations, but in my mind I imagine that it's a lot easier to store 20 or 50 truckloads of hydrogen than it is to store 20 or 50 truckloads of electricity. That is, if you buy all your power overnight and convert it to hydrogen, you can have a full day's worth of hydrogen sitting around in a big tank waiting for trucks to arrive. But if you buy all your power overnight (or generate with solar or whatever) and store it for a day's worth of BEVs, you'd need a titanic amount of battery at the station. Is that cost-effective?

I wonder how many solar-panel-square-feet * hours-of-sunshine are needed to complete one semi charge? Is generating and storing solar even in the picture here? Or are we just talking about buying at low rates from the utility?

Also, Trevor claimed in one interview (perhaps the Tesla Daily one) that by putting the hydrogen stations on certain highways he'd have access to lower-cost electricity in some other way -- from a federal backbone or something. I clearly have no idea what he was talking about, but is there any truth to the claim that by selecting your station locations correctly you can access cheaper power even within the same general area?

 

[SmartElectric]

The round trip efficiency of converting "excess" electricity to hydrogen (via electrolysis) and then reconverting that to running a vehicle is profoundly worse than just powering the vehicle from electricity alone, that is accepted scientific fact.

See "power to gas":
Power-To-Gas - an overview | ScienceDirect Topics
There are large sums of money being spent to determine the suitability of this.

Whereas, today, if you have "excess" electricity (during the day when solar is good or evening when wind is good), the obvious consumer of that electricity is EV's via smart charging enabled via grid management, not converting it inefficiently to another form (H2).

 

[jhm]

That's fair, so long as we're agreeing with the Nikola position that BEV is better for shorter routes and it's either diesel or hydrogen for longer ones.



I'm not sure I agree with this. I don't know a lot about hydrogen filling stations, but in my mind I imagine that it's a lot easier to store 20 or 50 truckloads of hydrogen than it is to store 20 or 50 truckloads of electricity. That is, if you buy all your power overnight and convert it to hydrogen, you can have a full day's worth of hydrogen sitting around in a big tank waiting for trucks to arrive. But if you buy all your power overnight (or generate with solar or whatever) and store it for a day's worth of BEVs, you'd need a titanic amount of battery at the station. Is that cost-effective?

I wonder how many solar-panel-square-feet * hours-of-sunshine are needed to complete one semi charge? Is generating and storing solar even in the picture here? Or are we just talking about buying at low rates from the utility?

Also, Trevor claimed in one interview (perhaps the Tesla Daily one) that by putting the hydrogen stations on certain highways he'd have access to lower-cost electricity in some other way -- from a federal backbone or something. I clearly have no idea what he was talking about, but is there any truth to the claim that by selecting your station locations correctly you can access cheaper power even within the same general area?

Okay, I see where you are coming from. Let me spell out some detail to clarify.

Nikola claims that their semis will get about 6 miles per kg of hydrogen. Electrolyzers are at best about 75% efficient, so it takes over 45 kWh to produce 1 kg of hydrogen plus extra to compress and distribute the gas. So the Nikola truck will consume over 7.5kWh per mile. Tesla claims their semi will consume less than 2 kWh per mile. Thus, the Nikola will consume more than 3.75 times as much electricity as the Tesla.

Now to power a truck for 300 miles per day. The Tesla needs about 600kWh/day, and nearly 2400kWh/day for the Nikola. If we are to generated this from solar at average 25% capacity factor on an average day, one Tesla needs 100kW of solar while the Nikola needs about 400MW of solar. At commercial scale we are talking about $900/kW, so $90k capex for one Tesla truck and $360k for Nikola. Plus Nikola will need 400kW of electrolyzer capacity to pair with the solar. The have an order with Nel that brings this down to an amazing $300/kW, but compressors, tanks and distribution pumps are extra. Tesla will also need 100kW of charger capacity to charge while the sun shines. (A fleet of 6 trucks can switch off charging at 600kW for 1 hour per day. So stationary storage is not needed for a sufficiently large fleet. One truck stores one truckload of electricity.) Lets assume the compressors, tanks and distribution hardware roughly nets out. So Tesla is looking at about $90k capex to generate energy for 1 truck while Nikola is up to $480k capex to generate hydrogen for one truck.

Now Nikola can reduce this substantially by simply buying cheap power from the grid rather than than using only self-generated solar. For example, rather than spending $120k on 400kW of electrolyzer (at 25% capacity factor), they could get away with just $60k for 200kW of electrolyzers (run at 50% capacity factor). The sweet spot for Nikola will places like the midwest where wind power is abundant. Wind is a little more pricey per Watt, but in the right places can pack a substantial capacity factor. So there are places where Nikola could get away with $240k energy capex per truck rather than the $480k above.

I am confident that Nikola will focus first on geographies where infrastructural capex will be cheapest. But if they are to grow, they will have to expand geographically into places where the spend per truck could be 2 or 3 times higher than their most economical regions. This geographic variation in capex will make it challenging to open up to truck routes that span a whole country. To make matters worse, Nikola want to package the cost of fueling with the truck lease. If I understand their intent, they won't be charging anything for fuel, and so they can't just charge more where the fuel is more expensive. Rather, Nikola will have to eat the cost differential. My hunch is that they will find renewables and electrolysis too expensive in certain geographies, and out of necessity they will buy hydrogen derived from fossil fuel or consume cheap, mostly dirty power. So this cause all sorts of problems with environmental claim. For example, a Tesla running on coal power is marginally better for the environment, but imagine how bad consuming near 4 times as much coal power to get as far can be. This is a pretty nasty worst case.

I do believe that Nikola is sincere in its desire to source mostly renewable energy, as is Tesla. But to survive economically, Nikola will need to cut some corners, same as Tesla. In the long-run solar, wind, batteries and electrolyzers all get cheaper and more widespread. But it remain a stretch for Nikola to need to source nearly 4 times as much power per truck as Tesla. Energy capex per truck will remain in a neighborhood of 4 times as much too. I honestly do want Nikola to do well, but I would not want to risk my own capital into what I believe will be a massively capital intensive enterprise.

 

[adiggs]

That's fair, so long as we're agreeing with the Nikola position that BEV is better for shorter routes and it's either diesel or hydrogen for longer ones.



I'm not sure I agree with this. I don't know a lot about hydrogen filling stations, but in my mind I imagine that it's a lot easier to store 20 or 50 truckloads of hydrogen than it is to store 20 or 50 truckloads of electricity. That is, if you buy all your power overnight and convert it to hydrogen, you can have a full day's worth of hydrogen sitting around in a big tank waiting for trucks to arrive. But if you buy all your power overnight (or generate with solar or whatever) and store it for a day's worth of BEVs, you'd need a titanic amount of battery at the station. Is that cost-effective?

I wonder how many solar-panel-square-feet * hours-of-sunshine are needed to complete one semi charge? Is generating and storing solar even in the picture here? Or are we just talking about buying at low rates from the utility?

Also, Trevor claimed in one interview (perhaps the Tesla Daily one) that by putting the hydrogen stations on certain highways he'd have access to lower-cost electricity in some other way -- from a federal backbone or something. I clearly have no idea what he was talking about, but is there any truth to the claim that by selecting your station locations correctly you can access cheaper power even within the same general area?

If Nikola is doing stuff where they claim they'll have access to cheap electricity in order to accomplish something (electrolysis to make hydrogen), then Tesla will have access to the same sources of cheap electricity.

And Tesla will be providing 90% efficient use of that electricity where Nikola will be providing 30% efficient use of that electricity (round #s for simplicity).


In this case, the only vector I see where Nikola can win is a theoretical combination of filling speed (has to be >3x to match recharge) and quantity of energy that can be carried along (longer range, fewer recharging stops) that also needs to be >3x just to break even. Hydrogen starts a long ways back in this 'race'.

Doesn't mean it can't catch up.

EDIT: My post and @jhm crossed (I was in a meeting and came back to this ). Similar ideas about the efficiency, but when in doubt - go with his over mine (I'm more directional - he's also directional while also being way more precise / accurate)


The vector around hydrogen being used as a fuel in random vehicles on the road that I am most personally concerned with, is that hydrogen is about the most flammable and powerful explosive stuff around. We've got lots of equipment, processes, and knowledge on how to handle hydrogen in static situations (such as fueling stations!). In mobile applications though, I don't see it. It's probably ok in newer vehicles that are well maintained and haven't had time to break down.

As vehicles get older though? I guess this could be good for manufacturers - planned and rapid obsolescence.

We're accustomed to driving around in gas cars that catch on fire and burn down to the frame on the side of the road. It happens often enough that it mostly doesn't make the news, even in an environment where 'if it bleeds, it leads'.

Hydrogen will be different. On the plus side, the safety systems are robust and among other things, they are designed to bleed off hydrogen "up", frequently in the form of a VERY hot jet of fire (which is mostly a good choice, but if the vehicles has been turned over or on its side...). I figure the typical car-car collision will end this way, given that the hydrogen tank is breached at all.

But rarely (much more rarely than with gasoline), a bad situation will come together and the 5kg hydrogen bomb will go off. This isn't a "gee that was pretty" or fireworks type of explosion. Think shattered glass a km away, and maybe a small hole in the ground, with bits and pieces of the car flying all over. Or if you're a terrorist, you've got something way better than fertilizer to work with (just got to get around the safety systems so you can get the bottle of hydrogen to blow up rather than burn off).


So electrolysis to create hydrogen with spare / cheap electricity, to store and run the other direction later when electricity is needed (as bad as the round trip efficiency is), as well as other industrial uses for hydrogen - I'm all there. Hydrogen in mobile / on-the-road applications, I'd rather not.

 

[jhm]

And Tesla will be providing 90% efficient use of that electricity where Nikola will be providing 30% efficient use of that electricity (round #s for simplicity).


In this case, the only vector I see where Nikola can win is a theoretical combination of filling speed (has to be >3x to match recharge) and quantity of energy that can be carried along (longer range, fewer recharging stops) that also needs to be >3x just to break even. Hydrogen starts a long ways back in this 'race'.

Interesting, I had 3X stuck in my mind too. I think that is the best case for hydrogen, if everything past the fuel cell is just as efficient as the EV. So I think Nikola gets to 3.75X because is is also an inefficient EV in addition to all the hydrogen inefficiencies. We've seen this with other comparisons between Teslas and other EVs. Tesla is pretty aggressive in seeking out every efficiency gain it can.

 

[engine ear]

Okay, I see where you are coming from. Let me spell out some detail to clarify.
Now to power a truck for 300 miles per day. The Tesla needs about 600kWh/day, and nearly 2400kWh/day for the Nikola. If we are to generated this from solar at average 25% capacity factor on an average day, one Tesla needs 100kW of solar while the Nikola needs about 400MW of solar. At commercial scale we are talking about $900/kW, so $90k capex for one Tesla truck and $360k for Nikola. Plus Nikola will need 400kW of electrolyzer capacity to pair with the solar. The have an order with Nel that brings this down to an amazing $300/kW, but compressors, tanks and distribution pumps are extra. Tesla will also need 100kW of charger capacity to charge while the sun shines. (A fleet of 6 trucks can switch off charging at 600kW for 1 hour per day. So stationary storage is not needed for a sufficiently large fleet. One truck stores one truckload of electricity.) Lets assume the compressors, tanks and distribution hardware roughly nets out. So Tesla is looking at about $90k capex to generate energy for 1 truck while Nikola is up to $480k capex to generate hydrogen for one truck.

I think the economics of the two different approaches is very interesting. To play devil's advocate, I think the analysis to support nikola's case is a bit different.

Lets say you are a company that develops wind farms in kansas. There are already many wind farms in kansas, and when the wind is strong, the wholesale price of power approaches 0. When wind is light, the wholesale cost of electricity goes up well above average.

A company like nikola can strike a deal with the wind farm owner to buy cheap electricity when the wind is strong and to not buy power when the wind is light. There is no additional capital expended to supply power to nikola, nikola is basically buying surplus power utilizing assets that are already in place. Nikola can then store that surplus energy as hydrogen and supply it as fuel to trucks.

My only issue is if surplus power is so cheap during windy days, and if hydrolyzers and fuel cells are so economical, why doesn't nikola use its hydrogen to power the grid on low wind days instead of all the complexity of building a fuel cell truck?

Tesla on the other hand can't easily shift energy between high wind days and low wind days. They will have to pay the average power cost. Since they are not using surplus capacity, there is a capital cost associated with the power they will consume. Tesla also will need a backup (fossil fuel?) power source for those low wind days and that also has capital costs.

 

[jhm]

I think the economics of the two different approaches is very interesting. To play devil's advocate, I think the analysis to support nikola's case is a bit different.

Lets say you are a company that develops wind farms in kansas. There are already many wind farms in kansas, and when the wind is strong, the wholesale price of power approaches 0. When wind is light, the wholesale cost of electricity goes up well above average.

A company like nikola can strike a deal with the wind farm owner to buy cheap electricity when the wind is strong and to not buy power when the wind is light. There is no additional capital expended to supply power to nikola, nikola is basically buying surplus power utilizing assets that are already in place. Nikola can then store that surplus energy as hydrogen and supply it as fuel to trucks.

My only issue is if surplus power is so cheap during windy days, and if hydrolyzers and fuel cells are so economical, why doesn't nikola use its hydrogen to power the grid on low wind days instead of all the complexity of building a fuel cell truck?

Tesla on the other hand can't easily shift energy between high wind days and low wind days. They will have to pay the average power cost. Since they are not using surplus capacity, there is a capital cost associated with the power they will consume. Tesla also will need a backup (fossil fuel?) power source for those low wind days and that also has capital costs.

If you have been following my posts here on hydrogen, you know that I am a big proponents of using electrolyzers to balance the grid. I believe they will prove to be the lowest cost solution for long-term storage, addressing seasonal and regional imbalances.

It is clear that their are sweet spots for Nikola. Specifically places that already have a surplus of wind should be looking to add electrolyzer capacity. Finding a customer for this hydrogen is a separate issue. If I we to invest in electrolyzers, I'd want to sell to buyers willing to pay the highest price. The are in fact many industrial uses for hydrogen.

Nikola would be able to make a much stronger case for their truck lease business model, if they were to cultivate a robust secondary market for their surplus hydrogen. Suppose for example that they scaled up hydrogen production about 50% higher than what their leased trucks would on average consume. Thus, they'd need a secondary market to sell about a third of the hydrogen they produce. Nikola investors would be keen to know who these customers are and how much they are willing to pay for a kg of compressed hydrogen. Knowing this local price for hydrogen would inform investors whether the truck leases are profitably priced or not. These leases will be priced under the assumption that it costs $2.40/kg H2 to produce. Suppose the local price that Nikola could get for a third of their supply is say $3.00/kg. This would give Nikola investors confidence in building out more capacity. Nikola could generate a 20% gross margin on fuel all day long. Suppose OTOH that this local secondary price is just $2.00/kg. Now Nikola is losing $0.40/kg, a -20% gross margin, on supplying the secondary market. This would be an area where Nikola should make a much smaller investment. In fact, they would do better to buy hydrogen at $2.00/kg than to try to produce it at $2.40/kg.

I would point out that Tesla faces similar challenges. Smartly Musk has always aimed at having about a third of battery production for non-automotive markets. A robust stationary battery market assures that as Tesla invests in its battery supply chain it will always find a buyer at profitable prices regardless of the automotive business. This is very smart. Investors are very confident that Tesla is not over investing in battery production capacity. Also we know that Tesla is not paying too much for batteries it sells in its cars.

Similarly Tesla has the ambition to power Superchargers and factories with solar power it produces. But in many places they consume power from the grid. Tesla Energy does produce more power than Tesla consumes, but that is done largely in separate locations. Where Tesla consumes power, it seeks out the lowest cost. Where Tesla produces power, it seeks out the highest price. And on balance they are a net producer. This is the most economical way for Tesla to be a profitable net producer of energy.

So Tesla is very savvy about secondary markets for both batteries and energy. They exploit local efficiencies to minimize the overcommitment of capex and to optimize cash from operations. Nikola will need to step up its game to perform better for its investors. If they can show that there is a profitable secondary market for their surplus hydrogen, I would be much more confident that they are not wasting capital. Maybe these are in the works. We shall see.

 

[Dave EV]

It's good not to get caught up in potential edge cases where fueling time is economically decisive. Rather I think it is best to think about lower hanging fruit first. (Diesel trucks can handle the extreme cases for a decade or more.) Tesla will put out a semi with 600 mile range or better. This is more than adequate to handle the bulk of use cases with a single charge per day, and so does not impact driver time at all. Just plug it in at the end of your shift. So for the bulk of truck applications, I don't believe that truck driver time costs will be appreciably different.

Ten years from not batteries will be cheaper, have longer range and have longer life, and EV trucking infrastructure will be much more built up. Also logistics companies will have more innovation to optimize the logistics of EV trucking. And yes self-driving tech may be in play too. So the most extreme use cases will look very different 10 years from now.

This is the key here. Just look at what EVs have done since 2010 when the Roadster and the Nissan LEAF were released.

Tesla now sells a full sized sedan with nearly double the range of the Roadster for far less money $75k vs $110k. The original Roadster had a 53 kWh pack, the Model S/X has 100 kWh now.

Nissan sells a 40 kWh LEAF for the same price as it originally sold a 24 kWh LEAF and a 60 kWh LEAF for slightly more.

By the time the Tesla Semi takes care of all the short-medium haul routes with the 300-500 mi range battery pack (5-10 years), they will have a Semi that goes twice as far for about the same cost. They will have batteries within 10 years that enable the Semi to go 500-1000 mi on a charge and with that kind of range and Megachargers placed in the right spots, there really will not be any significant routes that can't be handled by EV Semis which will only further surpress oil demand.

 

[TheTalkingMule]

Another Permian bankruptcy.

Rosehill Resources Inc. and Rosehill Operating Company, LLC File Voluntary Bankruptcy Petitions

Is this why the dollar has been sinking? Continued erosion of US oil production from 13Mb/d to something like 9Mb/d?

 

[ammulder]

@jhm, that was a great post with great numbers and detail, but here's where you lost me:

Tesla will also need 100kW of charger capacity to charge while the sun shines. (A fleet of 6 trucks can switch off charging at 600kW for 1 hour per day. So stationary storage is not needed for a sufficiently large fleet. One truck stores one truckload of electricity.)

Let's say we're trying to handle 6 truck fill-ups through a given Megacharger per day:

Case 1: Let's say Tesla has solar panels running all day. If a truck needs 600 kWh for the next 300 mile leg of the journey, and 100 kW of solar provides that at 25% over 24 hours (is that how the capacity factor works? I'm not even sure)... then you need I guess 600 kW of solar per Megacharger to handle the 6 daily fillups. There's no on-site battery storage. A truck arrives just after sunset. Now what? Does Tesla buy power in real-time that's maybe generated in the nighttime hours by a fossil-fuel plant? Because they figure they put 3600 kWh of solar power into the grid that day, so who cares if these particular 600 kWh worth of electrons came from natural gas instead, they've still covered their usage with equivalent renewable generation? That would suggest that Tesla doesn't just lay solar on/around their charger and wire it up directly to the charger, but instead puts a large grid-attached solar array somewhere in the vicinity to just offset their usage?

Case 2: instead of solar, let's say Tesla is just plain buying power from the grid. I know I'm simplifying but let's say it's half-price overnight, because everyone has their lights off and AC use is lower in the cooler nighttime hours and whatever. The Nikola station next-door can buy all their power overnight on the cheap and electrolyze into Hydrogen in a tank and then fill trucks with that Hydrogen during the day. Whereas Tesla has to pay double the price for daytime electricity rates to fill any trucks that arrive during the day... unless they have some meaningful fraction of 3600 kWh of battery on-site to buffer the time between purchasing and using the electricity. Do they buy the batteries or pay more for the electricity?

(FWIW I'm not arguing that the Nikola model is better, just trying to understand the practical differences.)

 

[jhm]

@jhm, that was a great post with great numbers and detail, but here's where you lost me:



Let's say we're trying to handle 6 truck fill-ups through a given Megacharger per day:

Case 1: Let's say Tesla has solar panels running all day. If a truck needs 600 kWh for the next 300 mile leg of the journey, and 100 kW of solar provides that at 25% over 24 hours (is that how the capacity factor works? I'm not even sure)... then you need I guess 600 kW of solar per Megacharger to handle the 6 daily fillups. There's no on-site battery storage. A truck arrives just after sunset. Now what? Does Tesla buy power in real-time that's maybe generated in the nighttime hours by a fossil-fuel plant? Because they figure they put 3600 kWh of solar power into the grid that day, so who cares if these particular 600 kWh worth of electrons came from natural gas instead, they've still covered their usage with equivalent renewable generation? That would suggest that Tesla doesn't just lay solar on/around their charger and wire it up directly to the charger, but instead puts a large grid-attached solar array somewhere in the vicinity to just offset their usage?

Case 2: instead of solar, let's say Tesla is just plain buying power from the grid. I know I'm simplifying but let's say it's half-price overnight, because everyone has their lights off and AC use is lower in the cooler nighttime hours and whatever. The Nikola station next-door can buy all their power overnight on the cheap and electrolyze into Hydrogen in a tank and then fill trucks with that Hydrogen during the day. Whereas Tesla has to pay double the price for daytime electricity rates to fill any trucks that arrive during the day... unless they have some meaningful fraction of 3600 kWh of battery on-site to buffer the time between purchasing and using the electricity. Do they buy the batteries or pay more for the electricity?

(FWIW I'm not arguing that the Nikola model is better, just trying to understand the practical differences.)

Generating from solar only is an extreme case, but it is fun to think through. So yes, you get need 600kW of solar. Now charge the tractors one at a time for 1 hour each at 600kW. This mean that one truck can depart full each hour of a 6 hour day. How many trucks need to be waiting to be charged? Only one. So while one is charging, the other 5 are working.

Now this requires very tight logistics. One truck has to arrive every hour and one depart. So suppose you want to buffer this with a battery. Great. What size? How about 1.2MWh about two charges? How much does this cost. Well, maybe you could buy some PowerPacks at say $150/kWh, so $180k. Or you could just buy an extra LR Semi for $180k. Why not add a seventh tractor to the fleet. You can always have an extra tractors waiting to charge in case an hourly arrival is a bit late. Plus, you don't have to spend any time transferring the charge from a stationary battery to a Semi. The extra Semi is already charged up and ready to go.

Once you realize that a Semi is just a 1.2MWh battery on wheels, you see that a slightly larger fleet is a better option than stationary storage. You simply have to structure your logistics so that whenever the sun is shining, you've got some Semis ready to take that charge. If you can do that, you also can charge whenever grid power is cheap too.

 

[ammulder]

Generating from solar only is an extreme case, but it is fun to think through. So yes, you get need 600kW of solar. Now charge the tractors one at a time for 1 hour each at 600kW. This mean that one truck can depart full each hour of a 6 hour day. How many trucks need to be waiting to be charged? Only one. So while one is charging, the other 5 are working.

Now this requires very tight logistics. One truck has to arrive every hour and one depart. So suppose you want to buffer this with a battery. Great. What size? How about 1.2MWh about two charges? How much does this cost. Well, maybe you could buy some PowerPacks at say $150/kWh, so $180k. Or you could just buy an extra LR Semi for $180k. Why not add a seventh tractor to the fleet. You can always have an extra tractors waiting to charge in case an hourly arrival is a bit late. Plus, you don't have to spend any time transferring the charge from a stationary battery to a Semi. The extra Semi is already charged up and ready to go.

Once you realize that a Semi is just a 1.2MWh battery on wheels, you see that a slightly larger fleet is a better option than stationary storage. You simply have to structure your logistics so that whenever the sun is shining, you've got some Semis ready to take that charge. If you can do that, you also can charge whenever grid power is cheap too.

OK, this is what I misunderstood. You're suggesting swapping cabs? Like, instead of driving up and charging your cab, you drive up with a load, unhook it, go over to a freshly charged cab, hook it up to the load, and drive off?

That might be cool for those super-fixed routes that Nikola likes to talk about. Factory to warehouse over and over again, or whatever. If we're talking cross-country, I guess e.g. FedEx/UPS and Walmart and various food distributors probably have pretty regular routes.

Still, is it at all realistic for the general case of trucking? First, all the owner-operators are out of luck. Second, it has all the robotaxi issues of "what if the last guy left his lunch all over the cab?" But perhaps most importantly, what proportion of routes are that regular such that a single trucking company could afford to leave spare cabs all over the place? I was once involved in a marketplace where truckers would register their location/rate/availability and load-owners would browse for available truckers, pick one, and do the deal. So I came away with the impression that many loads are more ad-hoc, go to the best bidder, whatever. (That was just my impression, and it may be wrong.) I'm sure there's a big market of fixed routes, but is it big enough to eliminate all the rest of the market right out of the gate? Or force all the companies to lease the trucks with the terms that Tesla can force them to switch cabs at every charge stop?

I wonder to what extent trucking companies and drivers are willing to change their process to accommodate better economics, or whether Tesla has to fit into the existing model to be attractive.

 

[Dr. J]

I wonder to what extent trucking companies and drivers are willing to change their process to accommodate better economics, or whether Tesla has to fit into the existing model to be attractive.

You can look at who has already placed orders. From September 2019:

Customers that have acknowledged ordering Tesla Semis for their fleets include: UPS, 125 Semis; PepsiCo, 100; Los Angeles-based TCI Transportation, 50; Bee’ah, a United Arab Emirates-based refuse-hauling company, 50; Sysco, 50; Wal-Mart, 45; Anheuser-Busch, 40; Canadian grocery chain operator Loblaw Cos., 25; DHL, 10; Florida-based City Furniture, 5; and Michigan-based grocery chain Meijer, 4.

Ryder Systems and J.B. Hunt also have placed orders but have only said they are for “multiple” trucks. One published report has J.B. Hunt’s initial order at 40 Semis.

Here’s Everything We Know About the Tesla Semi

 

[TheTalkingMule]

Reality starting to set in.

End game for oil? OPEC prepares for an age of dwindling demand

It has prompted some officials in the Organization of the Petroleum Exporting Countries, oil’s most powerful proponent since it was founded 60 years ago, to ask whether this year’s dramatic demand destruction heralds a permanent shift and how best to manage supplies if the age of oil is drawing to a close.

This article was quickly pulled from the app, of course.

 

[jhm]

OK, this is what I misunderstood. You're suggesting swapping cabs? Like, instead of driving up and charging your cab, you drive up with a load, unhook it, go over to a freshly charged cab, hook it up to the load, and drive off?

That might be cool for those super-fixed routes that Nikola likes to talk about. Factory to warehouse over and over again, or whatever. If we're talking cross-country, I guess e.g. FedEx/UPS and Walmart and various food distributors probably have pretty regular routes.

Still, is it at all realistic for the general case of trucking? First, all the owner-operators are out of luck. Second, it has all the robotaxi issues of "what if the last guy left his lunch all over the cab?" But perhaps most importantly, what proportion of routes are that regular such that a single trucking company could afford to leave spare cabs all over the place? I was once involved in a marketplace where truckers would register their location/rate/availability and load-owners would browse for available truckers, pick one, and do the deal. So I came away with the impression that many loads are more ad-hoc, go to the best bidder, whatever. (That was just my impression, and it may be wrong.) I'm sure there's a big market of fixed routes, but is it big enough to eliminate all the rest of the market right out of the gate? Or force all the companies to lease the trucks with the terms that Tesla can force them to switch cabs at every charge stop?

I wonder to what extent trucking companies and drivers are willing to change their process to accommodate better economics, or whether Tesla has to fit into the existing model to be attractive.

I think we need to focus on the sweet spots for this new technology. It can take many years for Tesla Semi to become 10% of the total trucking fleet. So that means Tesla can focus the 20% use cases where it can deliver the best financial returns.

I think independent trucking is dead in the long-run. Simply put, autonomous trucks make truck drivers obsolete. Right now you need truck drivers for 100% of freight miles. With autonomous platooning this could be reduced to 90% or less. Longer term this shrinks down to a niche smaller than 10%.

Thus, developing an EV trucking business model around the needs of independent truckers is self-defeating. That business model will shrink into a tiny niche along with the human drivers.

Rather I think it best for Tesla to focus on optimizing large fleet operations. This needs to be a highly integrated logistics operation. So things like generating power and charging the fleet is completely integrated with loading up trailers and warehousing. Increasingly every step along the way is automated. So things like swapping a trailer from one tractor to another or charging a tractor will be automated. At distribution hub, there will be a stock of trailers ready to be loaded, there will be a stand of tractors charged up and ready to go, the tractors will have autonomy to pick up a trailer as needs, and if necessary, there will be a waiting room of drivers ready to hop into a cab as it exits the hub. The companies that master these logistics first and develop a substantial network of hubs will be able to undercut most competitors in trucking.

Remember that WalMart and other big box stores were able to take huge market share away from traditional retailers largely by virtue of integrated logistical efficiencies. Gaining any edge that BEV and autonomous trucking can offer has huge stakes for who wins and loses market share among producers and retailers. So I think Tesla wants to be focused on the cutting edge of integrated logistics.