Battery Electric Vehicle News / Enjoying the Transportation Revolution

Martyn1981

Couple of extracts from the latest Carbon Commentary newsletter:

1, Tesla batteries. Rumours abound that Musk will say that its supplier in China is able to make batteries at around $80/kWh, well below today’s estimated costs of around $120. A 250 mile/400 km battery pack of around 65 kWh will therefore cost about $5,200. As a simple comparison, a new petrol engine in a small car will probably cost the manufacturer $2,500-$3,000 without including all the other necessary parts. This means that battery prices claimed by Tesla bring the cost of electric cars into close alignment with internal combustion engines. Other news that leaked from Tesla includes evidence of putting vehicle-to-grid capacity in new models as well as an application from the UK arm of the company to become a regulated electricity supplier. These latest stories indicate a strong possibility that Tesla aims to become a participant in energy markets, using the storage capacity of its cars to trade electricity. (Thanks to Raymond Betz).

6, Electric trucks in the Nordic region. Oat milk supplier Oatly said it would acquire 4 Einride electric trucks, based on a DAF skeleton. These vehicles will be capable of driving 120 km (not that far in Sweden) with 25 tonnes of vegan milk. Einride claims that the electric trucks will be the first in Europe to operate with this weight of cargo. The grocery retailer Lidl aslo plans to use Einride vehicles to deliver to its stores in Sweden. Among the many interesting features of Einride trucks is that they can be operated remotely by a ‘driver’ in a control centre. In Norway, food wholesaler ASKO announced an intention to move to all-electric transportation by 2026. As a first step, it intends to buy up to 75 battery trucks from Scania. ASKO had also been trialling hydrogen fuel cell trucks but seems to have gone for a battery-only fleet.

michaels

zeupater said:

Martyn1981 said:

zeupater said:

ABrass said:

It'll be interesting to see if Xpeng can produce them in significant numbers. So far they seem to be a very small manufacturer.

The FSD hardware isn't that important, it's the software that is still lacking. For that Tesla, with nearly a million units in the road, most of which are able to contribute to training the software, is in a pretty good and unique position.

Hi

... and even more importantly, the level of 'exposure' & 'experience' the AI has had to adequately tune the internal decision making process. That's the really big advantage that Tesla has & it'll take plenty of time for any on the competition to play catch-up ... by the time they've built enough cars (a million ??) to have access to around a billion miles of autonomous driving data they may be in a similar position to where Tesla are today, by which time there'll almost certainly be millions more Teslas on the road adding billions of additional experiences to the autonomous driving 'learning' ....

... catch up can be pretty hard if you're still on the start line & the competition is already miles down the road & still accelerating as hard (or faster!) as (/than) you're likely to be able to ....

If, as widely reported, Tesla currently have a 10 year development & technology advantage over any serious competition why would we expect everyone to have caught up in 10 years time? ... wouldn't it be more likely that at the end of 10 years the very concepts of continuous improvement & growing economies of scale will have extended the gap further? ... for example, maybe in 10 years time the gap will still be 2 years, or 5 years, or 10 years ... or more! ... without a crystal ball who knows (?), but as long as Tesla continue to innovate in line with their historical trend and the competition continue to respond to the market & technology situation that existed at a point almost a decade ago things will certainly remain interesting!!

HTH - Z

Hi Z, just watched a vid on software, and whilst some of it was above my pay scale, I thought of you and this post, and that you might find it interesting as it looks at the different approaches old school car firms v's tech companies take. And the 'bod' is a software engineer, cutting his teeth on the F-18 (please tell me he's joking about the use of punch cards?)

Why Tesla's Lead In Software Is Huge & Growing

Also, having re-read your post, I note you use the term 'continuous improvement'. This wouldn't have caught my eye before, but having watched the Sandy Munro vids of the TMY stripdown, I learnt the difference between continual improvement (build a better car each time you launch a new model year, or new car), and continuous improvement (build a better car throughout the production run) which is something he tried to promote with car companies through his career.

Hi

Call it whatever is the 'buzzword' of the day ... 5S's, six sigma, CI, LIFE ... etc it's a pretty basic process & product concept which is central to the reality that many manufactured products from companies that have embraced the approach are generally better quality, more reliable and considerably cheaper (relatively) than their counterparts of (say) 30-40years ago, even if they are superficially the same ...

Anyway, regarding the competitor that recent remarks have raised, I'm starting to see plenty of  reports that have sparked concerns as to where some of the major stakeholders (investors / customers etc) consider the future path to be ... if a company with extraordinarily large piles of liquid cash reserves isn't eager to continue to throw money into a 'guaranteed gold mine' and is starting to share exposure with other parties,then questions open up related to their own confidence in the future of their own project!! ....

The Tesla AI approach is pretty typical and falls in line with such systems I've evaluated, implemented & worked with ... this is exactly the same as what I've seen in the past & is basically built around the processing concept of  "I know what to do" ..alt ..."this is similar but different to previous experience so I'll stop & flag (/return to manual operation mode) - should I be worried?, what should I do when I see this condition again?, does this apply to intermediate conditions between previous parameter understanding & new condition?", etc .. obviously the rate of development is highly related to the number of extraordinary conditions experienced & a significant proportion of those will be geographically, seasonally & weather interconnected, all of which are related to the 'real world' as opposed to whatever conditions a limited development team can build a limited number of simulations for the AI to encounter, so the 'real world' encounters combined with a rapid development/engineering/prototype/solution approach would likely be a far more logical & successful strategy .... whilst certain autonomy solutions have racked up ~20million miles in extremely controlled & geographically limited environments supplemented by simulated operation , others are likely gathering as much data related to 'new experiences' on different types of roads all over the world in all weather conditions whilst operating within the scope of local (/differing!) highway & operating regulatory requirements, not every decade, year or month, but  probably every week, before testing solutions through their own simulation models! ..... that's what having ~2000x more test beds which have been sold to paying customer which are being operated as data collection nodes by those customers for free provides - an economic advantage over the what is widely considered to be the main competition, who continue to operate their own fleet using employed 'backup' drivers ...

Effectively, to accelerate development one approach burns their primary capital resource, cash, whilst the other makes a return on capital or even a profit!!

HTH - Z

The interesting thing is that google know this - they know why their search engine is better than everyone else's now, it is not the web crawling and valuing of deep links, it is the training their search algorithm gets from observing what people click on that give them an unassailable lead.

Martyn1981

michaels said:

zeupater said:

Martyn1981 said:

zeupater said:

ABrass said:

It'll be interesting to see if Xpeng can produce them in significant numbers. So far they seem to be a very small manufacturer.

The FSD hardware isn't that important, it's the software that is still lacking. For that Tesla, with nearly a million units in the road, most of which are able to contribute to training the software, is in a pretty good and unique position.

Hi

... and even more importantly, the level of 'exposure' & 'experience' the AI has had to adequately tune the internal decision making process. That's the really big advantage that Tesla has & it'll take plenty of time for any on the competition to play catch-up ... by the time they've built enough cars (a million ??) to have access to around a billion miles of autonomous driving data they may be in a similar position to where Tesla are today, by which time there'll almost certainly be millions more Teslas on the road adding billions of additional experiences to the autonomous driving 'learning' ....

... catch up can be pretty hard if you're still on the start line & the competition is already miles down the road & still accelerating as hard (or faster!) as (/than) you're likely to be able to ....

If, as widely reported, Tesla currently have a 10 year development & technology advantage over any serious competition why would we expect everyone to have caught up in 10 years time? ... wouldn't it be more likely that at the end of 10 years the very concepts of continuous improvement & growing economies of scale will have extended the gap further? ... for example, maybe in 10 years time the gap will still be 2 years, or 5 years, or 10 years ... or more! ... without a crystal ball who knows (?), but as long as Tesla continue to innovate in line with their historical trend and the competition continue to respond to the market & technology situation that existed at a point almost a decade ago things will certainly remain interesting!!

HTH - Z

Hi Z, just watched a vid on software, and whilst some of it was above my pay scale, I thought of you and this post, and that you might find it interesting as it looks at the different approaches old school car firms v's tech companies take. And the 'bod' is a software engineer, cutting his teeth on the F-18 (please tell me he's joking about the use of punch cards?)

Why Tesla's Lead In Software Is Huge & Growing

Also, having re-read your post, I note you use the term 'continuous improvement'. This wouldn't have caught my eye before, but having watched the Sandy Munro vids of the TMY stripdown, I learnt the difference between continual improvement (build a better car each time you launch a new model year, or new car), and continuous improvement (build a better car throughout the production run) which is something he tried to promote with car companies through his career.

Hi

Call it whatever is the 'buzzword' of the day ... 5S's, six sigma, CI, LIFE ... etc it's a pretty basic process & product concept which is central to the reality that many manufactured products from companies that have embraced the approach are generally better quality, more reliable and considerably cheaper (relatively) than their counterparts of (say) 30-40years ago, even if they are superficially the same ...

Anyway, regarding the competitor that recent remarks have raised, I'm starting to see plenty of  reports that have sparked concerns as to where some of the major stakeholders (investors / customers etc) consider the future path to be ... if a company with extraordinarily large piles of liquid cash reserves isn't eager to continue to throw money into a 'guaranteed gold mine' and is starting to share exposure with other parties,then questions open up related to their own confidence in the future of their own project!! ....

The Tesla AI approach is pretty typical and falls in line with such systems I've evaluated, implemented & worked with ... this is exactly the same as what I've seen in the past & is basically built around the processing concept of  "I know what to do" ..alt ..."this is similar but different to previous experience so I'll stop & flag (/return to manual operation mode) - should I be worried?, what should I do when I see this condition again?, does this apply to intermediate conditions between previous parameter understanding & new condition?", etc .. obviously the rate of development is highly related to the number of extraordinary conditions experienced & a significant proportion of those will be geographically, seasonally & weather interconnected, all of which are related to the 'real world' as opposed to whatever conditions a limited development team can build a limited number of simulations for the AI to encounter, so the 'real world' encounters combined with a rapid development/engineering/prototype/solution approach would likely be a far more logical & successful strategy .... whilst certain autonomy solutions have racked up ~20million miles in extremely controlled & geographically limited environments supplemented by simulated operation , others are likely gathering as much data related to 'new experiences' on different types of roads all over the world in all weather conditions whilst operating within the scope of local (/differing!) highway & operating regulatory requirements, not every decade, year or month, but  probably every week, before testing solutions through their own simulation models! ..... that's what having ~2000x more test beds which have been sold to paying customer which are being operated as data collection nodes by those customers for free provides - an economic advantage over the what is widely considered to be the main competition, who continue to operate their own fleet using employed 'backup' drivers ...

Effectively, to accelerate development one approach burns their primary capital resource, cash, whilst the other makes a return on capital or even a profit!!

HTH - Z

The interesting thing is that google know this - they know why their search engine is better than everyone else's now, it is not the web crawling and valuing of deep links, it is the training their search algorithm gets from observing what people click on that give them an unassailable lead.

Yep, that's the silly bit - "Data is the new oil." Good explanation of all this from Hyperchange a couple of weeks ago, mentions that machine learning was not necessarily possible in 2009 when Waymo started. Also supports what Z said with suggestions of costs of around $1bn pa, to operate the small autonomous fleets, whereas Tesla is getting paid (selling cars).

Tesla has 3 Billion Miles of Data - Waymo Has 20 Million

Martyn1981

Well, it is a BEV!

World’s largest all-electric aircraft set for first flight

The world’s largest all-electric aircraft is about to take to the skies for the first time.

The Cessna Caravan, retrofitted with an electric engine, is expected to fly for 20-30 minutes over Washington State in the US on Thursday.

The plane can carry nine passengers but a test pilot will undertake the inaugural flight alone, cruising at a speed of 114mph (185km/h). Engine maker magniX hopes the aircraft could enter commercial service by the end of 2021 and have a range of 100 miles (160km).

Among the other companies developing electric aircraft are Zunum Aero, which is building a 27-seat plane with a 680-mile (1,100km) range and the engine maker Rolls Royce, whose Accel programme aims to produce the fastest all-electric plane to date. However, in April, Rolls Royce and Airbus cancelled their plans for a hybrid electric aircraft. The German company Lilium is working on a five-seater jet-powered electric air taxi.

michaels

I wonder how the economics of small electric planes works out?  I believe jet fuel is not subject to the massive tax that is on petrol (at least in Europe) so perhaps the fuel cost savings are much  smaller - especially given the extra weight of batteries as opposed to aviation fuel.  Of course is IDE planes faced the true environmental cost of the emissions then it might be a different calculation.

Martyn1981

michaels said:

I wonder how the economics of small electric planes works out?  I believe jet fuel is not subject to the massive tax that is on petrol (at least in Europe) so perhaps the fuel cost savings are much  smaller - especially given the extra weight of batteries as opposed to aviation fuel.  Of course is IDE planes faced the true environmental cost of the emissions then it might be a different calculation.

Roei Ganzarski, the CEO of magniX, said current aeroplanes were both expensive to operate and very polluting. “Electric airplanes will be 40%-70% lower cost to operate per flight hour,” he said. “That means operators will be able to fly more planes into smaller airports, meaning a shorter and door-to-door experience, with no harmful CO2 emissions.”

Just personal guesses, but potentially additional benefits to both the airports and operators if BEV flights allow for night flights. This would open up additional use of the existing infrastructure, and possibly reduce the need for airport expansion. The operators might see savings if they aren't competing (as much) for the airport slots if the working day lengthens?

Martyn1981

S3X just got cheaper:

Tesla cuts prices across lineup, Model 3 now starts at $37,990

Tesla has reduced the pricing of its electric cars across its entire lineup* in the middle of the night – with thousands of dollars price cuts on some models.

The global pandemic and the lockdowns put in place to slow down the propagation of the coronavirus have ravaged the economy of most countries and the auto industry has been particularly affected.

In response, Tesla, who has rarely suffered from demand problems, has decided to reduce prices across its entire lineup of electric vehicles.

*Not quite true:

Model Y, which Tesla just started delivering, is the only vehicle in the automaker’s lineup that hasn’t been affected by the price changes tonight.

The automaker still enjoys a backlog of orders on the smaller electric SUV since it has been taking reservations for a year and only started deliveries in March.

michaels

Martyn1981 said:

michaels said:

I wonder how the economics of small electric planes works out?  I believe jet fuel is not subject to the massive tax that is on petrol (at least in Europe) so perhaps the fuel cost savings are much  smaller - especially given the extra weight of batteries as opposed to aviation fuel.  Of course is IDE planes faced the true environmental cost of the emissions then it might be a different calculation.

Roei Ganzarski, the CEO of magniX, said current aeroplanes were both expensive to operate and very polluting. “Electric airplanes will be 40%-70% lower cost to operate per flight hour,” he said. “That means operators will be able to fly more planes into smaller airports, meaning a shorter and door-to-door experience, with no harmful CO2 emissions.”

Just personal guesses, but potentially additional benefits to both the airports and operators if BEV flights allow for night flights. This would open up additional use of the existing infrastructure, and possibly reduce the need for airport expansion. The operators might see savings if they aren't competing (as much) for the airport slots if the working day lengthens?

So we do not have the batteries yet for even shorthaul jets but probably do for 10 seater 1000km jets - could these work out as cheap to operate as a Boeing 737 especially as they should be APT tax free if the tax is really about the environment and could fly high frequency multiple routes from small local airfields, with a big fleet they could easily flex the destinations to match demand (Monday morning 5 of 6 planes fly from small London airport to Manchester and Edinburgh at 15 min intervals, Saturday morning 2 fly to Manchester and one each fly to Dublin and Plymouth, the schedule changing as the demand changes) .  I suspect if they were autonomous the answer is a definite yes, paying for one or two flight crew perhaps not.

ABrass

Cessna sized planes going electric is good news. Cheaper skydiving.

Edit: and cheaper flying lessons, but really who wants to land in a plane like a tourist?

Coastalwatch

Well, that's certainly one way of supporting a cleaner future post Covid. I wonder what our government will introduce to endorse all the recent words about their good intentions!

French drivers get €7,000 incentive to go electric

French President Emmanuel Macron is offering drivers a €7,000 (£6,226) incentive to make their next car an EV.

The boost for the electric vehicle sector is part of a wider package of €8 billion (£7bn) to help the nation’s automotive industry recover after the Covid-19 pandemic.

The aid is also expected to include incentives for people who want to replace their old vehicles and buy a lower-emission model.

https://www.energylivenews.com/2020/05/27/french-drivers-get-e7000-incentive-to-go-electric/