Apple Car Challenges: Sobering Thoughts
mondaynote.com
mondaynote.com
For reference, the battery pack is 24kWh, so it could theoretically run the heater while stuck on a road for 24 hours straight. A Model S with a 85kWh battery pack could keep the heat on for over 3 days.
I live in Montana, so such trips are common, not exceptional. Having a pass close in front of you (or while you're on it) is not terribly exceptional either.
Two drivers leave on this trip you described with the same fuel in the tank (250 miles worth), one in a Tesla and one in a Toyota. Both travel 100 miles before getting stuck. Both have to wait 8 hours in the cold on a closed road. Both have 80 miles to go.
The difference is, the Tesla has 112 miles of range left at that point, and the Toyota only has 86. It consumed 2.56 gallons of gas idling to run the heat, a much larger percentage of its overall fuel than the Tesla.
Assuming neither driver planned to run out of fuel on this trip (since 360 miles is beyond the range of either vehicle), both are within range of either a gas station or a charging station. They're all along the length of I90 and I15 in Montana. The Tesla has a larger margin of error for detours and closed roads on the way.
Small towns (which for Montana I'll define as towns with a population of around 1,000) with gas pumps are plentiful. Small towns with hotels, a bit less plentiful. Small towns with superchargers are nonexistent.
> You wouldn't have left if your destination wasn't within 80 miles of I90 or I15,
Bozeman to Great Falls, 180 miles. A supercharger on one end of the trip, a hotel on the far end, and lots of nasty winter driving in-between. Also known as Tuesday for many folks I know.
Limiting yourself to 100 miles from a Supercharger knocks out a lot of space in Montana. Most of the interesting parts of Montana are well over 10 miles from a supercharger. Yellowstone, Glacier National Park, the reservations, the feed lines into southern Idaho, Wyoming and North Dakota... No southern Idaho.
There's a joke that says that on the east coast, 100 miles is a vacation, but in Montana, 100 miles is a grocery trip. That applies to a lot of the non-coastal US as well. Have a look at https://www.teslamotors.com/supercharger?redirect=no, and look at how many huge open spaces exist.
An EV uses 20+ kilowatts to move at highway speeds. It uses 1 kilowatt to run climate control.
Sitting in traffic for an hour is equivalent to driving for 3 minutes, in terms of how much battery you use.
This past weekend the high temperatures in large parts of the midwest were below the cutoff point.
Actually, a Tesla would keep you warm while waiting for longer than your gasoline car.
A 2 liter engine, idling at 700 rpm, will consume about .32 gallons per hour. A 15 gallon tank then gives you 46 hours of heat.
A 85 kilowatt hour battery pack can run a 1000 watt cabin heater for 85 hours. That's almost twice as long.
If you travelled 50 miles from home before getting stuck, the gasoline car will have 40 hours of heating time left, and the EV will have 68 hours of heating time left.
It still can (and does) go in all electric mode, even up to 100 mph.
The BWI w/ range extender still only goes 150 miles between gas ups. That's pretty annoying. It's great for days when you just need a bit extra to get through. But for road trips, it's nearly a deal breaker. Though I guess you could just suck it and stop every 2 hours.
A small battery EV with a ICE range extender is simply a better sweet spot and takes better advantage of both mode's unique features and avoids a whole bunch of mechanical parts running through the car.
Also worth noting that the i3 gas tank was electronically limited in capacity (along with other tweaks) to meet California regulations for getting a rebate.
That is beyond anxiety, it means planning your life around your car range, instead of the other way around. And with the current availability of charging stations, it also determine the place you are allowed to go, that's a very big drawback for people that use their car mainly on the weekend for leisure.
I'd say at rush hour 2/3 of those cars are being used. I'd imagine Uber would just buy thousands of cars. Why not? What's to stop them? I'd just scrap my car and have an Uber car parked somewhere close to my house. I don't care, just so long as I can get to work. I imagine that utilities will soon provide automatic Uber car docks and cars will cluster around them at night. There must be a lot of Tesco car parks that are 99% empty at the moment (at night) rent this space to Uber and in the day, the car park would empty as those Uber cars would be driving around.
Also, I don't mind car sharing so long as I don't have to wait. Uber could easily plan trips that pick up and drop off without having to wait like you do for more traditional public transport such as trains and buses...
And fun fact: I'm 26 and I don't have a driver's license. So do many of my peers. I totally rely on public transit (save for the few occasions where I need to move furniture etc.). Public transit service is exceptionally good and reliable in most of Germany's big towns.
This is interestingly close to the Kozmo/PeaPod/whatever on-demand grocery delivery problem: as long as you only have to service high population densities, you can make a profit. As soon as you try to handle suburbs, the costs of providing service increase at a rate larger than the square of the radius you are trying to serve - and nobody wants to pay that, and the urban core won't subsidize the suburban outliers until you've already dominated the market.
It seems to me like people who're buying a car in that price range would either own a house or at least have access to a garage.
It's fairly common in London to not have a garage, and even if you do then there is unlikely to be any power there. Most likely is on-street parking, and that's assuming you're not in a flat.
Beside the cost, European cities are keener on getting rid of cars entirely, so they are reducing the number of parking spaces as years goes by, so you may be right after all, the zero emission solution for people living where I live is to not have a car at all ( which is, to be fair, not at all unrealistic. There is a strong possibility that the technologies to solve my few "not conveniently covered by public transport" will come from some sort of rental/uber before electric car is actually practical where I live )
Even just thinking of my brother in law and his wife both make $150k+ live in a condo in sf and even with their garage, they would not be able to install power at their parking spots.
I do see a few of them around here and there but it's a logistic nightmare to even consider a power-cable-charged electric car in a city like this.
Also, my garage has plenty of cars in that price range (and far above), but there's no outlets for people to charge cars.
So, an interesting number would be the % of luxury car purchases that are electric over time. I would imagine this is probably trending up, and that's just due to a single model from a single manufacturer.
I don't think he has underestimated the challenge.
I'm a supply chain guy. I work for a company with one of the top peer-rated supply chains in the world and Apple is still so far ahead of us that it isn't even funny. Apple could, with minimal comparative effort (as in, easier than any other company, not easily), translate their manufacturing supply chain knowledge towards automotive manufacturing and optimize it to the point of absurd superiority over every other automaker out there. That alone could change the profitability of the company significantly toward self-sustainability within a matter of a 3-4 years. Throw in a well funded software and hardware research division, poach the top specialized auto manufacturing talent, and you have a recipe for the next Toyota or (forgive the faux pas) VW.
But if you start from scratch, you are going to be 10 years away from making a dent in the market, regardless of how much expertise you have.
I was told by a GM employee in their tech center in Warren, they have COBOL running in VMs on modern servers whose terminal screens get scraped into excel spread sheets. This is one of many examples I could give you. I could literally spend hours telling you the horror stories. It's government levels of inefficiencies and leadership has little or no visibility into the system.
Everyone in Metro-Detroit knows of the problems in the auto industry. The leadership lacks vision and workers find their niche and retire hoping to do the same things with maybe a promotion to management. There is not a constant drive, at least in the software side of things, to constantly improve and learn like in there is in Silicon Valley.
The above is applicable to GM and Ford. Chrysler, on the other hand, is a whole other level of screwed. There's a reason they've been bailed out 2 times by the government and now bought/merged by a foreign car company, AGAIN.
Apple, Google, Tesla and Uber will crush all but a few existing automakers, around the globe, and leadership has just now come around to acknowledging the threat. The regular Metro-Detroiter has no idea.
It's a good thing that Apple is a software company.
> Apple, Google, Tesla and Uber will crush all but a few existing automakers, around the globe, and leadership has just now come around to acknowledging the threat. The regular Metro-Detroiter has no idea.
Of those four, only Tesla has actually produced cars that can be sold to the general public. Tesla is a 13 year old company. It has well known reliability problems, and it is supported by a luxury-only price point and market. It's market share is sub 1%.
That is exactly my point. Apple could, if they wanted, start from scratch and it will take 10 years absolute minimum before they are even remotely relevant. Or they could buy an existing car manufacturer with actual market share, and in the course of 2-3 years make it extremely profitable, and in 3-5 years make it into something Apple would be willing to put their logo on. Starting from scratch might be reasonable when you are talking software, but building an automotive company takes decades.
It doesn't even have to be Fiat Chrysler or domestic for that matter...any existing automaker with more than 20 years experience could be on the table when you have $200B in cash on hand. They could buy BMW for ~$50B or Daimler for ~$70B. Hell, they could buy Toyota at ~$175B.
Like you said, Apple has $200 billion. That can buy some speed but Apple doesn't have to rush.
Tesla did. From Tesla's founding to the first Model S being delivered took 9 years and they did it with under a billion dollars. They also spent time proving electric cars were viable with the Roadster and secure more funding. Apple doesn't need to take this first step.
So a crappy guess would be 5-7 years to make a car. This is not worth the risk of spending $40-70 billion, especially when you can get away with spending $2 billion and get better results than Tesla.
My main point was the auto companies are so bogged down with legacy (legacy software, legacy manufacturing, legacy thinking, legacy costs, legacy habits, legacy dealerships) that starting from scratch is the better option, even if it does take 10 years.
> and in the course of 2-3 years make it extremely profitable, and in 3-5 years make it into something Apple would be willing to put their logo on.
What your suggesting is, essentially, to unravel 50-60 years of legacy. People have tried and failed. This isn't just a company like GM, a massive institution in its own right, it's all the 3rd party suppliers. I think you're underestimating the work required, even if you assume no worker revolts.
I also don't think Apple wants to manage or take on the liability of car divisions from around the world.
Tesla has taken 13 years to ramp up production to 55k per year, and that was even after lucking out with a down-on-their-luck Toyota manufacturing division complete with experienced management and auto workers getting dropped from Toyota's production portfolio. You hold them up as a standard for progress, but they are literally a toy car company. Ford started producing cars in that quantity in 1911. 20 years from now, Tesla will still be paying high priced consultants with global manufacturing experience for their experience building cars at scale. Tesla will no doubt succeed if they can keep up that pace, but they have an extremely long road ahead before they can even compete with Nissan or Hyundai in the US in terms of production capacity, reliability, and production costs.
That legacy culture that you deride has another name: knowledge and experience. The same Tech-industry "we can do it better than you with no domain expertise" arrogance is exactly the reason why Silicon Valley hasn't made any dents in Health Care, Finance, or Energy, despite the untold billions that their VCs have dumped into those exact industries.
In the 80's people were still designing cars on paper on top of desks the size of a kitchen table. Imagine pulling out an eraser or creating a fresh copy manually if your drawing was too messy. It was so ... damn ... slow. People got paid to drive from one side of Metro-Detroit to another delivering a large cylinder of drawings.
> to bring Toyota's production quality up to the standards of GM in the 1970's
Now cross an OCEAN and do that.
Things get exponentially faster and everyone around the world is now talking to each other instantly. Its hard not to learn from other car manufacturers.
> even after lucking out with a down-on-their-luck Toyota manufacturing division complete with experienced management and auto workers getting dropped from Toyota's production portfolio
This is a failing on Toyota's part and a successful use of this talent on Tesla's. It's a sign of things to come. Not unlike Apple's hiring of A123 System's battery engineers from Detroit or Uber's hiring of Carnegie University PHD students out from under a GM funded research project.
> You hold them up as a standard for progress
Yep. ICE powertrains are complicated, expensive, more prone to wear, need more maintenance and are part of an epidemic global problem. The only sticking point of Tesla's plan is the expensive battery. They're working on that too.
And they are a standard. GM just copied them. From the huge battery at the bottom of the car to Model 3 sticker price on the Bolt. Tesla is now setting the standard.
> but they are literally a toy car company
We all gotta start somewhere. In one sentence you talk about how it took one company 20 years, another 40 years. In comparison what they've done in 13 is pretty spectacular.
> Tesla has taken 13 years to ramp up production to 55k per year
Do you mean to start a company, build the Roadster to prove it could be done, attract capital then eventually sell 55k cars? Because that's way more impressive than buying a factory in 2010 and hitting that goal 5 years later, which is still really impressive.
And why are you only shitting on them? In only 3 years of producing the Model S, its the 2nd best selling large luxury car. They did this with 0 ads and 0 dealerships with 0 dealership ads. Everyone in this country knows GM, Ford, Toyota and Mercedes. They get slapped in the face constantly with TV ads. I wouldn't be surprised if less than 10% of the country knew about Tesla. They only have room for growth.
> That legacy culture that you deride has another name: knowledge and experience.
Good point. I mean bad legacy. To be successful you have to understand what to ditch (ICE powertrains and SOAP protocols), have the guts to ditch it, and know what to hold onto (rubber wheels and robotics). Automakers are not ditching bad legacy fast enough. As of 5 years ago GM was running COBOL and probably still is. I'll bet you $100 Tesla isn't.
> is exactly the reason why Silicon Valley hasn't made any dents in Health Care, Finance, or Energy, despite the untold billions that their VCs have dumped into those exact industries.
All 3 of those industries are heavily regulated and mostly hard problems. Until recently, Silicon Valley has mostly been focused on low hanging fruit. Starting a website is easier than starting a bank due to the shear amount of compliance.
The reason the VCs are dumping tons of money into these spaces is due to their difficulty. Facebook needs money when they buy data centers a few years in, finance needs money right away.
> but they have an extremely long road ahead before they can even compete
Big things like starting a car company take time. Now add Google, Uber and Apple into this mix. There all in the same place but without the COBOL. In fact innovation and ditching bad old ideas are in their blood and Detroit is still running SOAP and COBOL.
Other sources claim that Tesla is making 25% profit per Model S, and similar for the Model X. Is this article cherry-picking numbers?
However, industry-wide figures for power density show a clear upward trend. It's not exponential like Moore's law, but in just under a decade or so, electric cars are going to have range parity with gasoline ones.
Also, the graph is misleading. If you combine Tesla EV and other EVs, there is a very promising upward curve in the lower right-hand corner.
Only if investment and breakthroughs continue. Moore's law worked out well because people kept buying computers constantly since they were useful even while much slower. When it comes to electric cars though, there is little incentive for consumers to buy them at the moment so there is a significant risk of investment petering out.
Investment and breakthroughs in lithium ion batteries were not and are not only driven by electric vehicles. Far, far from it, as a matter of fact.
I think it goes without saying that they're not going to use the same engineering techniques they use for their other software to develop the software for the car. It's not about culture. It's about adopting very stringent procedures similar to NASA, aeronautics firms, and other industries that need to build software with extremely low rates of error. This has nothing to do with culture. These processes exist because culture fails when such low error rates and resiliency is required. Airplanes have been running incredibly complicated software without issue for decades. For that matter, cars too. Making this a culture issue is frankly ignorant of the whole software development process.
As for his past sales figures and argument that people aren't buying electric cars, he misses the point that other than Tesla there are few practical electric cars if any (Leaf's range is too short, Prius isn't an electric car, it's a hybrid, etc.) and the Teslas are priced for the upper class only.
Not to mention a ton of software.
Those are several products in one. Wether or not they actually build and design everything themselves is irrelevant. Not all car companies makes everything themselves, most share engines and other parts.
It's a massive undertaking. Musk was able to pull off Tesla because he saw a frontier that needed to be saturated, and has a history of trailblazing through new industries. It's not that Apple can't do it. They'll just have a hard time doing so.
While Apple may be supply chain wizards in electronics they would need to replicate it many times over to build cars and even then create a whole new distribution and support system.
I think their best bet is to be the smarts of all cars, why restrict themselves to just their own cars. Just launching one car would cost them a sizable chunk of their money trying to launch more than one would be many times more difficult, just ask Tesla. Tesla isn't out of the woods because unlike Apple they don't have the reserves to spend; assuming Apple stock holders would go along with the idea of spending tens of billions to start yet another car company.
No, I think Apple is better at being another supplier and getting everyone into their game. That gives them more flexibility and more penetration for a lot less. They are likely also the only company that could get their logos on the screens of all cars
Systems engineering was basically created by the auto and aerospace industries for this reason. It coordinates all these disparate disciplines into a single, cohesive effort. What I have seen of systems engineering[1] in SV does not leave me with the confidence that most of those companies will actually succeed beyond some science experiments. At least, not the companies that haven't started out as fundamentally car companies like Tesla did.
[1] Real systems engineering, not glorified, overworked sysadmins
> PC guys are not going to just figure this out. They’re not going to just walk in.
EDIT: Found what I was thinking of. It was Ed Colligan, Ex-Palm CEO on 16 Nov 2006.
>>We've learned and struggled for a few years here figuring out how to make a decent phone. PC guys are not going to just figure this out. They're not going to just walk in.
I still have issues with the environmental impact with electric fuel cell manufacturing and recycling. It's horrible on the environment and the recycling issues (hell, we still have problems just recycling plastics) cause me to doubt electric as a long-term viable solution.
Just in case:
http://www.theguardian.com/vital-signs/2015/jun/10/tesla-bat...
In a 2013 report, the U.S. Environmental Protection Agency’s Design for the Environment program concluded that batteries using nickel and cobalt, like lithium-ion batteries, have the “highest potential for environmental impacts”. It cited negative consequences like mining, global warming, environmental pollution and human health impacts.
http://pubs.acs.org/doi/pdf/10.1021/es903729a
The main finding of this study is that the impact of a Li-ion battery used in BEVs for transport service is relatively small. In contrast, it is the operation phase that remains the dominant contributor to the environmental burden caused by transport service as long as the electricity for the BEV is not produced by renewable hydropower
> In a 2013 report, the U.S. Environmental Protection Agency’s Design for the Environment program concluded that batteries using nickel and cobalt, like lithium-ion batteries, have the “highest potential for environmental impacts”
The highest potential for environmental impact among current Li-ion battery chemistries. This chemistry is currently best-in-class for energy storage, but it is an area of very active research. For example, lithium-air batteries hold promise for more energy storage with a carbon cathode.
Of course grid power is also moving to renewables.