Things that are actually risk factors for Tesla

With all the crazy reasons the shorts have to short the stock and talk negatively about Tesla, let’s look at the actual risk factors that face Tesla over the next 12 months

1. Lack of demand. This would probably manifest itself and some form of economic recession. The good news for Tesla is that the wealthy tend to prosper no matter what, and that is their target market. 

The other possibility is if Tesla can’t get to the point where they can offer the $35,000 standard range model. A good chunk of the pent up demand will evaporate if those reservation holders can’t buy the $35,000 car. Given Tesla’s work in Q3 2018 to reduce the cost of the car, it is likely they’ll be able to continue that work 10-15% cost reductions per quarter) to allow them to offer the base price car. This has already manifested itself through the Medium Range vehicle now offered because of work done to reduce costs. 

The other ways that a lack of demand could be a problem don’t seem likely – specifically other car manufacturers figuring out how to build compelling electric cars in large enough volumes to meet demand, and choosing to sell them nationwide. Companies like GM and VW promise they’ll introduce a ton of EVs and plug-in cars, but will probably only sell them in CA and OR and anywhere else required by law. And the US is last to get the cars because Europe and China currently have stronger requirements for EV sales. Global manufacturers will design, build, and ship cars for those markets first and the US is an afterthought. 

2. Product mix. This quarter will be the highest average selling price the model 3 will ever have. Because they aren’t selling the cheaper models. The only models sold are the $75,000 performance, $55,000 LR AWD, and $50,000 LR RWD. The average sale price was north of $55,000, which won’t be the case probably ever again (I think it’ll eventually settle around $47,000 – lots of base models with modest option mix and many AWD/LR models and few performance units). 

This means that the free cash flow and the profit Tesla earned this quarter might be only about 15% less than what they will ultimately earn at a full run rate of 7,000+ Model 3 units a week at an average sales price of $47,000 and 25%+ gross margin. So expect somewhat flat free cash flow and profits for the next 6 quarters (which is fine if they bank it and pay off their debt and their stock price rises to above $360 so most of their outstanding convertible debt becomes shares rather than debt). 

3. Paying for future expansion while paying off debt. Specifically the factory for the Model Y vehicle – how much does it cost and where does it get built? Fremont/NUMMI is full and can’t accommodate another model line (if it could there wouldn’t be a tent), so it has to be assembled somewhere else. If the factory costs $3B most of that capex will come in the second half of 2019 and first half of 2020. While Elon wants it to be an alien dreadnought it is more reasonable to expect that there will be incremental improvements – things that have already been discussed are wiring harnesses and wiring bundles that are easy to manipulate by robots so allow more mechanization of the product line. 

4. Federal investigation. Apparently there may be a federal investigation of Tesla going on. Not quite sure what they’re investigating (the SEC has settled the civil charges over the “finding secured” debacle) so we’ll see but unless there is outright fraud going on I don’t think we’ll see charges. If no banking executives were charged criminally after the 2008 crash I find it hard to think that Elon or anyone else will he charged for being too optimistic on their projected growth or anticipated production rates. 

Tesla and Hype

There has been much made recently about Elon Musk going on the offensive (and sometimes being offensive) on Twitter in the last week or so.

Disclaimer: I’m an EV fan, drive a Chevy Volt, and own $TSLA and some Lithium mining stocks.

Drawing the line in the sand

Recently, Elon drew the line in the sand, effectively saying Tesla wont raise another dime from Wall Street (bonds, stock issuance) during the Q1 conference call by cutting off a “boring” question from a Wall Street analyst, and taking a question from a YouTube analyst. He has stated that in Q3 and Q4 of 2018 Tesla will be profitable, and he has pushed his company to start efforts to conserve cash and increase margins.

Good. This is probably something long overdue at Tesla and will help their gross margins on their products and increase profitability (or reduce losses). Tesla right now is roughly 400M a quarter in the hole from a balance sheet perspective (the operating profits from selling cars and other stuff don’t pay for all the overhead of running the company), so they’d need to generate that much to at least break even (not including any increases in debt service coming up in future years).

But the PR and news coverage…

The PR coverage over the last few weeks has been bad for Tesla. Between the Consumer Reports non-recommendation of the Model 3 (which was reversed yesterday after they updated the car’s performance via OTA software updates), and Elon has been going to war with the media and random people on social media, and having his mom defend him on Twitter (was not expecting that). It can be difficult for a non-enthusiast to think Tesla will be able to deliver, or even be a solvent company, by the end of the year.

Timing is everything

I learned something about myself and probably others this weekend about the news and PR and current events. It was announced that ICE had “lost” 1,475 children, and that minor children were being separated from their parents at the border. I drew the conclusion that ICE was taking those kids away and lost them. But that was not the case, in fact they are two separate things – ICE places children with family/extended family/suitable adults and is then no longer legally responsible, but does follow up with the kids occasionally to make sure they’re OK. Following up is good. Separately, ICE separates minor children from their parents at the border. Bad. But I had erroneously linked them in my head until I read a twitter thread about it. And it corrected my thinking on it.

Elon essentially did the same thing – by saying things like if Tesla made the $35,000 version of the car now, they’d bleed money and die, and then turning around and announcing the availability and prices for the AWD and Performance models of the Model 3, he linked those two items in people’s minds. So the media and people are having a collective freak out over the fact that Tesla isn’t going to be able to deliver a $35,000 model at all. Although having done this once before (the 40kWh version of the Model S) doesn’t help Tesla’s case.

While his timing was bad (layoff the twitter, hire more PR people, let them deal with this rough patch), none of the bad press has any material impact on whether or not Tesla will succeed in its goals. It might have an impact on the stock price, or if he has to go back to Wall Street to raise money, but that isn’t in their plans right now. It certainly wont deter the true believers who handed Tesla $1,000 more than two years ago.

Some napkin math

Tesla’s stated goal is that they will have 20% gross margins (profit on each car, minus the costs of the parts, assembling the car, set-asides for warranty work, etc.) on the car. So on a $35,000 car, that is $7,000 of gross margin and $28,000 in parts and assembly. But on a $60,000 AWD car, the gross margin is $12,000 and the parts and assembly cost $48,000. (these aren’t the actual numbers, typically the gross margins on the cheapest car are smaller than the goal while the gross margins on the highest-end cars are more than the stated goal)

The next battle, once Tesla (hopefully) gets to 5,000 cars a week at the end of Q2 is all about gross margins, getting the cost of manufacturing the car (assembly) down enough to make it a cash flow generator for the company. The first option I can think of is to work to trim the hours per car (through automation or otherwise). Right now the plant runs 24/7 – three 8 hour shifts per day everyday of the week. The first step is to reduce that to six days a week, and maybe down to five, and still produce 5,000 cars per week. Then move from 3 to 2 shifts. That trims the hours worked from 168, to 120 for M-F, and then down to 80 hours for two shifts. This is a reduction of 14% for each day they trim, and a total of 55% for 2x8x5, but is probably an even larger impact on labor costs due to overtime, shift premiums, etc. that Tesla is probably incurring with the 24/7 operations. This is easily six months of effort (Q3, Q4), if not longer but the gains are nearly immediate and flow to the bottom line quickly.

Back to the top, if Tesla needs 400M a quarter, at an average selling price of $45,000 and a gross margin of 20% ($9,000) that comes out to needing about 45,000 cars per quarter, or “only” about 3,500 to 4,000 units per week. If they want to be profitable (generate money for the Semi, Roadster 2.0, etc.) anything beyond that they’ll want more cars – like maybe 5,000 a week, 12 weeks a quarter for 60,000 cars.

It seems doable. I’m optimistic about their chances (if I wasn’t I would have sold my stock). Don’t believe the (negative) hype, but you should be skeptical of Tesla’s constantly moving goals, which look more like timelines for software projects that I’ve been a part of, rather than a mass-produced physical product.

Oddly Specific SMS Spam/Phishing Attempt

So this isn’t good – got a spam SMS at one in the morning. It’s a bad sign when the SMS is specific enough that it uses your name, your wife’s name and its sent to your phone number only. The action to take was a link hidden behind URL shortening. It said “there is a new request from {wife’s name} 5 mins ago trying to connect right now” from a number 706-752-xxxx.

I’m trying to think who would have all those pieces of information (my name, my phone number, wife’s full name) and how they could be compromised. Was it a company like Amazon or some other online retailer? Could it be my employer or health care provider? Financial?

In some ways its kind of scary to think how many people have access to this information. I know better than to click on these kind of links, but it is incredibly annoying to sit and wait for the other shoe to drop (the company involved announcing they’ve been hacked) and then work on remediating whatever needs to be addressed. In the meantime, maybe I need to change my passwords again for my email and online shopping websites.

Tesla’s First Lithium Agreement

Last week, Tesla announced they have made a deal with two companies for a mine in northern Mexico to mine Lithium products (Lithium Hydroxide, 29% Li by weight, and Lithium Carbonate, 18% Li by weight). The companies involved still need to get funding to build the mine and commence operations, but it should be a bit easier for investors know they have a deal lined up with Tesla Motors to buy all that lithium that will be mined.

The deal starts with 35,000 tonnes of Lithium initially, and may scale to 50,000 tonnes of Lithium as the production at the plant scales up. So how many kWh does that translate to?

Lithium Ion batteries vary in their amount of their elemental lithium (vs. lithium hydroxide or lithium carbonate) based on the type of chemistry and other materials present. The NCA cells found in the Model S use about 300 g of elemental lithium per kWh. This translates roughly into either 1.67 kg of Lithium Carbonate (LI2CO3) or 1 kg of Lithium Hydroxide (LiOH).

Let’s assume that its a 50/50 mix of both Lithium products, which yields a requirement of 1.33 kg of Lithium products per kWh of battery produced. At 35,000 tonnes of Lithium, that is 26.3 GWh of manufactured cells, assuming that there is no waste product from the lithium coming into the factory (this is not likely the case, so the 26.3 GWh will be the upper bound for our 50/50 assumption). The maximum initial production would be 35 GWh if 100% of the supplied lithium product was Lithium Hydroxide, and the minimum initial production would be 21 GWh if 100% was Lithium Carbonate.

At the contract maximum of 50,000 tonnes of Lithium product, you’re looking at a minimum of 29.9 GWh, a midpoint of 37.6 GWh, and a maximum of 50 GWh of battery cells for the Gigafactory.

So given the inputs and assumptions, we can estimate 26.3 GWh of battery cells to start with, and 37.6 GWh of cells at the top end of the range. If 75% of the cells manufactured go into cars, and the other 25% of the cells go into Tesla Energy products like the Powerpack and Powerwall, it would mean a production of between to 350,000 and 500,000 55 kWh battery packs for the Tesla Model 3 (55 kWh is my estimate for the average pack size sold).

Tesla’s initial stated capacity for the Gigafactory is 35 GWh of cells (and 50 GWh for battery packs – meaning 15 GWh of finished cells are delivered to the plant). However recently Elon has stated that the manufacturing capacity may be even higher than that, as they seek to more efficiently use floor space in the manufacturing process. Tesla may be able to build the same sized building, but rather than get 35 GWh of cell manufacturing capacity, they could get 50 GWh or more. Its currently unknown how much additional capacity they might be able to get, we won’t get our first look until the first phase opens in the first half of 2016 for Tesla Energy cell manufacturing.

(part of this post cribbed from my own Ars Technica comment on this same story)

As a postscript, I’m disappointed the company I had money invested in, Western Lithium, was not Tesla’s first choice for Lithium for the Gigafactory, despite that the mine for the Lithium was in the same state as the Gigafactory and less than 200 miles away. It might be that Tesla needs more than one lithium supplier, and there is still a chance for Western Lithium. And Tesla is not the only game in town either – other battery factories like LG Chem also need Lithium.

Tesla will miss their Model 3 $35,000 target price, but not for *that* reason

Tesla will likely miss their $35,000 target price tag for the Model 3 due in 2017 or so. But not because the batteries cost too much, or production costs are too high.

Rather, demand will be what keeps prices higher.

It is my opinion that the demand for a 200+ mile range, well-designed, luxury Tesla EV will be so huge that Tesla simply raise the price to match demand with what they can supply. It might end up happening in a similar way to how Tesla cancelled the 40kWh version of the Model S. It might be bad PR, they might shrug their shoulders, but its smart business.

The initial roll out in 2017 or 2018 won’t feature vesicles at the $35,000 base price tag – in the pattern of the Model S and Model X launches, we’ll see the highest margin units go out first – signature, largest battery pack, AWD, supercharging, for around $50,000. It will be 6 months or more until they can start offering the lower priced cars (probably around 40K) that are the standard battery pack and trim levels.

Maybe after a year or so, they’ll get around to making the lowest margin units. But even then I don’t believe that Tesla will sell their base model for $35,000. Its likely that they wont get down to that point until about 2020 – after initial demand has been satiated and the Gigafactory is running full steam producing battery packs below $200/kWh.

Chevy’s 200 Mile EV

I’m quite happy to hear a mainstream automaker like GM will release a 200 mile EV that will be available for general sale (not limited to CARB states like CA, OR, NY) and it will be affordable. Its just their timing is bad. With gas around $2.75/gal around the country, people aren’t worried about switching to EVs. Its not until its back above $3.50 do people start to flinch and $4/gal is when they pitch a fit. We’ll see what gas prices are in early 2017 when the vehicle is released, and how that affects consumer behavior. I just don’t foresee a return to $4+/gal gasoline anytime soon.

I’ve been a proponent of raising the gas tax, and indexing it to inflation. I’d like to see the national gas tax raised 6c/gal, and then indexed to inflation. Boosting the current federal tax by 33% would allow the country to repair its deficient highway system, with the side effect of construction jobs.

Some notable pundits have suggested a “gas floor tax” of $3/gal, but that really wouldn’t work since oil companies, refiners and station owners would just keep the price at $3/gal and keep the profits for themselves, rather than sell it below $3/gal and turn the money over to the feds to bolster the highway fund. If you wanted to keep the price of gasoline elevated, you would need to create a supplemental gas tax that would index it to the price of a barrel of oil – an extra 25c the following month when the average oil price is below $60, 50c below $50, 75c below $40.

Ok, enough about infrastructure and gas taxes.

The other issue I have with this EV is that it’s built on the new Chevy Sonic platform. Which is pretty small. It’s almost as if GM is making a mistake putting 200 miles of range into a car that no one would be happy driving 200 miles/3 hours in. There is some amount of psychology built into that 200 mile number – people want to see a range number starting with a “2” before they feel comfortable buying an electric car, regardless of how far they drive it.

The concept has a little crossover style to it, which means it’s likely the batteries are going in the floor like Tesla and the car is sitting up on top of the batteries, 4-5″ higher than normal. This is a good design in that you still get trunk space and seating for 5. It might be that this car has a much narrower appeal in the US (where gas is relatively cheap and people like bigger cars), and higher appeal in Europe and Asia which is OK with smaller cars and higher fuel prices.

Another factor is fast-charging. Tesla has their supercharger network they’re constructing, but its proprietary (for now). The 90kW CCS charging standard needs to take off before people start feeling comfortable buying EVs. The recharge time on a 90kW station from a near-empty 200 mile EV battery is rather quick – the first 75 miles will recharge in under 15 minutes, allowing people to get around and finish their errands.

Finally, price. $30,000 after the federal rebate is still too much for a Sonic-sized car. Even factoring in a $3,000 premium for it being a crossover and $7,000 in first five years gasoline vs. electricity savings. The base model at $30,000 is going to need to be very well equipped (LT model) if it wants to attract buyers. I don’t think we’ll see a lot of price cuts until the $7,500 tax credit starts to expire for GM and Nissan, and they have to reduce the price of their smaller EVs to around $25K before people get interested.

Southern Nevada Transportation: 2015 Edition

This week, the southern Nevada RTC released their transportation projects list, one of which was a subway underneath the Las Vegas Strip. I’ll touch on a few of the options they’ve mentioned.

Taxi stands: Taxi staging areas at casinos and the two airport terminals will need to be expanded as we build more casinos and more convention space. Also, with Uber coming to the state, how people bypass the standard taxi line and take an Uber instead needs to be figured out. You cant have Ubers in the taxi line, and most hotel porte-cochères don’t have lots of space for pickup and drop off needed to accommodate a massive amount of Uber vehicles, nor do they have waiting space for people in Uber cars to park and wait for their passengers to walk out of the casino and to the car (I’ve been yelled at before by hotel security waiting to pick up people in a drop-off area). You can’t walk out to Las Vegas Blvd and hail an Uber, that will be a disaster for traffic, nevermind that it’s illegal to hail a cab on the strip – you have to go to a casino taxi stand. So where do the Uber cars go? The parking garages at the casinos? The passenger pickup lanes at the airport? I can’t think of a good answer to this.

Pedestrian Overpasses: More of these are needed along the stip in the areas away from the major intersections. This one is a big “duh”, the most difficult is the massive amount of right-of-way needed on the sidewalks to land these immense structures and allow for elevators to be ADA compliant.

Las Vegas Monorail: The RTC suggests linking the recently expanded Mandalay Bay Convention Center with the monorail, and adding a stop at the Sands Convention & Expo Center. First and foremost, the monorail should go to the airport (notwithstanding the subway idea below). Thats job one, and if the taxi industry was defeated by Uber, then surely the Las Vegas Monorail group can stand up to the taxi lobby this time. Next, good luck trying to get Shelly to buy into the idea of putting a monorail stop near the Sands Expo – he rejected it the first time and I’m sure he hasn’t changed his mind. Until there is a new CEO of Venetian/Sands, there won’t be a monorail stop there. I love the idea of the monorail being upgraded to 1) support more passengers at peak times, 2) connect to the airport, and 3) connect all three major convention spaces in town. I just don’t think it will happen for reasons that aren’t related to funding and business plans.

Las Vegas Strip Subway: I like this idea. There are two major geotechnical issues with building a strip subway – the water table below the strip is shallow, so there will have to be a lot of effort into making sure that the tunnels don’t leak, and don’t fill up with water during construction. Geotechnical work will be the single largest factor in the pre-construction phase of this project. However, a few casinos have built underground structures (the Palazzo parking garage, for example) so its not unreasonable to think this can be done. The second is caliche, which is a concrete-like substance found in the soil in the Las Vegas area. Its expensive to go through, and I don’t know how deep it runs – it could be a foot or six feet thick. If you build the subway at a sufficient depth (below the water, sewer, gas and power lines underground) then you can get under the caliche and be ok, only needing to penetrate it at access points and passenger stops. Phase one should go up and down the strip from the old Saraha (now SLS) to the Mandalay Bay, which could do a better job connecting all three major convention spaces because the Rivera will become an extension of the LVCC, and a people mover system on LVCC property could move people from the strip to the main hall located a half mile away. Phase 2 would link it to the airport east from Mandalay Bay underneath the two north/south runways at McCarran Airport to terminals 1 and 3. Phase 3 extends north to the Arts District and Downtown Las Vegas.

I’ve taken light rail from SeaTac Airport to Downtown Seattle and I loved it. It was way better and cheaper than dealing with a taxi. Then I took the monorail from Downtown Seattle to the Space Needle. That was neat. I only had to rent a car to go north to the Boeing factory in Everett.

Express Exit Ramps and HOV/HOT Lanes: NO, BAD RTC! NO! The HOV lanes that are in place now don’t get used much (they already have pretty limited hours of operation for HOV use only), and building more ramps between highways or side streets isn’t going to get people in Vegas to carpool. This isn’t a business town where everyone works 8-5, this is a service industry town where everyone works different hours, can be let off their shift early if it’s dead or be asked to stay late if it’s busy. People just can’t carpool here without substantial risk. What they can do however, is setup a park-and-ride system that allows strip casino workers from all over town to park in their neighborhood in a parking lot, and then take a bus for the area they work in (downtown/north strip, mid-strip, south strip/airport) and be taken to work on mass transit, and be returned to their car when their shift is over (the busses would run 24/7).

US 95 (I-11) Interchange at Maryland Parkway: This is great idea, the downtown Las Vegas area needs more access from the highway, the one-and-a-half exists from the US 95 and one from I-15 isn’t enough to handle the large spikes in traffic from AAA baseball games and nights at the Smith Center, plus all the normal traffic from the very large outlet mall downtown.

Electric Jets?

I’ve been thinking a lot lately about electric jets. With all the purported battery breakthroughs, and a discussion of what aviation might be like in 2030, 2040 or 2050.

For the battery baseline, we’ll go with Solid State Lithium metal batteries. These batteries are completely solid, they don’t have a liquid electrolyte that can oxidize and catch fire. They are currently under development, and are expected to go to smaller scale commercial production in the 2016-18 timeframe. Dyson has invested in one of the leaders in the space, Sakti3, with the hope that Sakti3 can produce cells above and beyond current Li-Ion cells. The experimental values for the weight and volume that have been developed so far are about 800Wh/kg and 1500Wh/l, compared to 250Wh/kg and 700Wh/l for the top of the line batteries today. I would expect these batteries to be widely available and inexpensive by the mid-2020s.

I first looked at a Boeing 737, and whether you could build a similar size and weight, but replace the fuel tanks with batteries, and the engines with ducted fans, and see what that would give you.

Boeing 737-800 weight of max fuel: 24,025 kg
Boeing 737-800 volume of max fuel: 29,660 liters

Replace 24,025 kg of fuel with 24,025 kg of batteries: 12,813 liters of batteries (43% of original fuel volume)
Energy storage: 19.2 MWh

Replace 29,660 liters of fuel with 29,660 liters of batteries: 48,198 kg of batteries (200% of original fuel weight)
Energy storage: 38.5 MWh

You can only realistically use the smallest value of the two, since both volume and weight are limiting factors, its about which one you hit first. In this case, its weight limited. So our new aircraft has 19.2MWh of batteries weighing 24,025kg and taking up 12,813 liters of volume (less than half of the volume the fuel used up, so there may be opportunities to redesign the aircraft and reduce its overall weight). The power these batteries could generate based on the cell weight is a maximum power of 12MW, estimating 500W/kg of cells. This estimate is in line with what today’s batteries can produce on a sustained basis.

So what kind of demands are going to be put on the battery to propel the aircraft from ground to flight? Initial take-off thrust required will be high. The Airbus E-Fan demonstrator had hub-motors in the landing gear to help get the aircraft up to speed. This reduced the load on the engines to propel the aircraft up to take-off speed. If the aircraft is running maximum throttle, then the two engines are producing about 242kN of total force. Based on the E-Fan’s substitution of ducted fans, they have 30kW = 0.75kN, or 40kW = 1kN, which means the engines would need about 10MW of power over the course of about 45 minutes to get up to cruising altitude (using 7,500 kWh of energy). Cruise thrust is about 40% of max thrust (depending on altitude, air density, etc.), so energy usage per hour of flight is 4,000 kWh, and energy usage during descent is 25% for about 30 minutes (1,250 kWh). To have a three hour flight (plus 45 minutes of reserves per FAA FAR 121), you’d need about 18,750 kWh of energy, just under our estimated capacity of 19.2 MWh.

Turns out no matter how you run the numbers on any sized airplane, you really only get about 2-3 hours of operating the aircraft, which is bad for aircraft that tend to fly 2-5 hours (think B737 and A320). But, its good for regional jets where the longest flights are only 2-3 hour. It would seem logical to have Regional Jets be the first type of aircraft using batteries for propulsion. In order to have electric aircraft fly longer routes, you’d need to improve engine efficiency (use less kW to generate 1kN in thrust), increase battery weight and volume characteristics (store more energy per unit mass or per unit volume), or figure out a way to put a highly efficient generator on the plane along with a fuel source (e.g. a 5MW turbine and natural gas to provide power during take-off and in case of emergencies).

There are other side effects to running an aircraft on electricity – you’ll end up redesigning the aircraft since you can ditch the fuel tanks; you’ll be flying slower, probably M0.7 instead of M0.78 or M0.8 that current jets fly at, which means that three hour flight wont go as far as it used to; better protection against lightning; more efficient interior use of energy; and more. The ranges from various airport hubs (700mi) show that it’s range wont be a big deal (map from Great Circle Mapper).


Apple & Cars

So the latest rumor this week is that Apple is going to develop a car. They’re hiring automotive designers and engineers. Yes, it would be totally awesome if Apple came out with a car, and it kicked GM/Ford/Chrysler’s asses the same way the iPhone kicked Microsoft and Blackberry’s asses.

But can Apple fix any of the issues that currently face electric vehicles? Or will they just be a slightly different $100,000 Tesla, splitting the market that is not really that big in the first place?


As I’ve discussed before, the batteries in the new iPhones rival the batteries in Tesla’s Model S in some aspects, but fall behind in others. The six critical battery parameters are:

  • Cycle life: number of full battery charge/discharge cycles to 80% of its original capacity
  • Volumetric Energy Density: number of watt-hours of energy the battery can store per unit volume, usually measured in watt-hours per liter (Wh/l)
  • Gravimetric Energy Density: number of watt-hours of energy the battery can store per unit of mass, usually measured in watt-hours per kilogram (Wh/kg)
  • Power: the ability of the battery to generate or accept power, measured using rate-capacity defined as the C-rate – 1C is charging or discharging the battery in one hour, 0.5C is two hours, 2C is 30 minutes, and 10C is 6 minutes
  • Safety: how much torture can the battery withstand before it becomes a danger to the people around it
  • Cost: the price per usable kWh of battery capacity for the vehicle

Assuming the 1,000 cycle life promise Apple made when it went to sealed batteries is still true, that would provide for a long lifetime (for a 200 mile EV, 1,000 cycles to 80% yields about 180,000 miles on the pack before it only gets about 160 miles per charge).

The iPhone 6 and 6 Plus battery’s energy densities are quite good – 250Wh/kg and 575 Wh/l. The battery cells in the Tesla Model S are around 250Wh/kg and 700Wh/l. This means Apple’s equivalent batteries would weight the same, but take up 22% more space – this is a difficult thing to overcome, so Apple would need to be very creative on how they can come up with more space to store the battery pack relative to Tesla’s battery pack.

The power output of the current iPhone batteries is unknown, rate capacity generally isn’t an issue for batteries in small consumer electronics. The iPhone and iPad batteries can usually recharge in about 1 to 2 hours, which indicates a C-rate of 1C. Batteries for EVs generally need a C-rate of 2C to support fast chargings and highway speeds in all conditions (rain, snow, headwinds, etc.).

Apple’s batteries are generally safe. The lithium polymer cells are a lot safer than the NCA chemistry used in the Tesla Model S.

Finally cost, Apple and Tesla produce roughly on the same scale now (see below) but Tesla has a much more aggressive ramp planned for battery production than Apple does. And the lead time on building new battery manufacturing capacity is pretty long.

Quantities, Oh God The Mass Quantities, of Batteries

Next I wanted to figure out how many kWh of batteries Apple sold in 2014. This is pretty difficult because Apple’s phone models have different cell sizes: 5/5S/5C varied between 5.45 and 5.96Wh, the 6 has 6.91Wh, the 6 Plus has 11.1Wh. So beyond that, the mix of how many phones sold is unknown, so thats another estimation we have to factor in.

Lets assume that for the first three quarters of 2014 (no iPhone 6/6 Plus), the average battery size per phone sold was 5.7Wh, and in the final quarter the average battery size was 6.5Wh. In the first three quarters they sold 118M iPhones, and in the insane fourth quarter they sold about 75M iPhones (mix of 5-series and 6-series phones). This results in 672 MWh of batteries sold in the first three quarters and 487 MWh of cells sold in the final quarter, for an iPhone total of 1,159 MWh of cells, or just over a gigawatt-hour of energy storage devices.

The iPad sold 63.35M units. We can judge from the average selling price of around $420, that a lot more iPad minis are being sold than traditional, larger iPads. If we assume that the mix is 4 mini iPads to 1 large iPad (either last gen or current gen), then the average battery capacity was 25Wh, which is a total of 1,583 MWh of batteries.

This brings us to an approximate total of 2.75 GWh of battery cells produced by Apple for just the iPad and iPhone line. This doesn’t include the batteries used in the iPod or in Mac laptops. Estimating the mixes and volumes of laptops and iPods is beyond my expertise at this moment.

Meanwhile, Tesla sold 31,600 or so cars. If the average unit battery capacity was 75kWh (3 85kWh units for every 2 60kWh units sold), that would yield about 2,370 MWh, or 2.37 gigawatt-hours. For comparison, the Gigafactory will be able to produce 35 GWh of batteries.

It is safe to say that Apple uses more batteries than Tesla in 2014. However, that may change in 2015, as Tesla will try to grow their overall production by 70%, increasing their total annual usage to about 4 gigawatt-hours. Apple, with iPad sales flattening or even declining, likely will not see a 45% increase in battery cell usage to keep up with Tesla.

(the logistics and supply chain people at Apple really do the Lord’s work, hats off to them)

Design & Engineering

I have no doubt Apple’s design team would have a field day with an Apple-mobile. I just hope its as practical as it is beautiful. One of the recent thoughts that has caught my attention is that the value in the car itself is changing. Thirty years ago, 0% of the value of a car came from the software. As the cars got better, engine computers became more advanced, and the infotainment systems in cars became more prevalent, the value of software has increased, from 10% to 40% over the next 10 years as cars learn how to drive themselves, manage their internal components, and become more “smart” in general.

This puts companies like Apple and Google ahead of the game, with their fleets of software engineers and development know-how. Ford, GM, and everyone else has to play catch-up. Can they offer sufficient amounts of money and incentives to lure developers away from places like Apple and Google, where they could invent and develop things to change the world, to Ford, where they will make another difficult-to-use in-car infotainment system.

One interesting aspect would be Apple deciding to take advantage of Tesla’s offer to release all their patents. They can use the same skateboard battery module design and powertrain to underpin the car, with a new design and Apple flair to the rest of the car.

Actually Manufacturing the Car

Tesla’s most recent quarterly conference call brought out the bears – they’re burning cash like crazy on capital expenditures in order to ramp up for an annual run rate of 2,000 cars a week (100,000 per year) as well as building the Gigafactory that could make cells for 500,000 cars a year in 2020, plus batteries for renewable energy storage.

However, all this spending – $5 billion on a battery factory and $2B or so more on its factories in California, is just petty cash for Apple. Apple currently has a $177B cash pile, of which $150B is net of debt. Apple could easily invest $5B in the facilities to build the batteries and the cars – its not a matter of whether they have the cash, its if its the right way to spend that money.

More Importantly, Supporting the Car

The genius bar is usually pretty good about customer service (I haven’t been in a while, knock on wood), even if the lines are horribly long. But how does that translate to getting your Apple EV fixed? Most Apple Stores are in malls, not a place you can drive your car into to get fixed. So what does Apple do? If they go with automotive franchises, they lose their exacting control over the process. Beyond that, they run into the same problem as Tesla with franchises – it’ll be multiple brands under one roof since they will be a small-time player to begin with, and its always more profitable for the dealer to sell a higher maintenance gasoline car compared to a low-maintenance Apple EV because dealers make their money on service, not on new car sales.

It would make a lot of sense for Apple to partner with Tesla on the supercharger network, and infuse a boatload of cash to expand it to support the number of Apple EVs made. Here there are a lot of brand synergies between Apple and Tesla.

But What’s the Sustainable Competitive Advantage?

Apple would only be thinking about becoming a car manufacturer (because eventually it will be more than one car – it’ll be a line of cars) unless it thought it could bring something to the table that all the other companies out there (Ford, GM, Toyota) can’t, and that it would have a long term sustainable advantage. They aren’t trying to be like Elon Musk, who just wants to advance EVs and save the planet from carbon poisoning.

Design? Apple has impeccable design under Jony Ive. The Model S has great design, but lacks luxury in many ways that show its newness to the car industry (the seats, the small visor), and those are being fixed, but it will take a while. Apple will likely have some of these issues out of the gate too, but they would likely be fixed within the first few iterations.

Batteries? Could Apple be working on engineering and developing its own batteries? Not likely. As I illustrated above, Apple ships a tremendous amount of batteries every year. Is it enough to rely on the battery industry at large to continue to innovate in the battery space? Maybe not, but battery research is remarkably difficult – the annual improvement rate is only 7-8% and big breakthroughs are very rare, even if the scientific papers stack up to the ceiling. If Apple has something up it sleve to differentiate itself like working, mass-producible solid-state batteries that offer 700Wh/kg and 1300Wh/l, it would be a coup in the portable consumer electronics and EV worlds – phones as thin as 15 playing cards, cars that can go 400 miles without recharging. But this is very unlikely (I really hope I’m wrong but I doubt it).

Integration? This is always where Apple shines. Apple isn’t generally the first to move (they weren’t first with contactless payments) but they are usually the first to get it right from top to bottom, in a way that the user can understand. The difficulty here is that cars are a mature industry, very mature. Its easy to say that just about every company could do in-car computers better, even Tesla. Apple will show everyone how its done. But after that, and people understand the new paradigms for how people interact with cars, then what? This knowledge and innovation diffuses throughout the industry and becomes general knowledge in the same way physical keyboards went away and capacitive touch screens became the norm.

Self Driving? The individual automakers aren’t doing all the heavy lifting individually, automotive suppliers like Bosch and startups like Mobile Eye are the ones coming up with the hardware and software to solve pieces of the autonomous driving puzzle. Apple could either redo that work or simply integrate parts from suppliers into a self driving system like Tesla is. It’s nothing terribly novel or unique.


The problem to be solved with Electric Vehicles is batteries – weight, volume, range, cycle life and safety. All five dimensions need to be improved, plus the cost will need to come down dramatically before the general public adopts EVs over gasoline cars (especially in the current gas price climate).

What isn’t a problem is design or features. Sure, design can be improved and refined, but a better designed car won’t bring out customers in droves. An electric car fits very nicely with Apple’s sustainability goals – working to have a cleaner environment, but there won’t be that much of a market given the current limitations on batteries. This is the problem Fisker had – brilliant design but they didn’t solve the battery problem in a new or novel way – and now they’re out of business.

Its difficult from the outside at this early stage to determine why Apple would want to develop a car, along with the immense investment that would need to accompany development and production if it had honest aspirations of being a worldwide automotive manufacturer. For Apple to enter the market, there needs to be some long-term competitive advantage here. I just don’t see it right now – just designing a better looking or more user friendly EV doesn’t solve the major pain points consumers have right now.

The problems with EVs are battery range, recharging time, and battery weight and volume. And Apple isn’t more or less likely to be the company with a group of electrochemists that discover a breakthrough than any other company, large or small, doing battery research today. It is for primarily that reason that I think Apple would be a fool to enter the automotive space, specifically EVs, in the short term. As cars transform from machery we operate to automated consumer electronics on wheels, there is a space for Apple and others who want to move in that line of products, but that transition is 10-15 years away.

Broadband & Title II

The core problem at the center of today’s broadband debate: laying down fiber, or even hybrid fiber/coax or 75Mbps DSL, to everyone’s house is extremely capital intensive. It leads to mono/duopoly conditions because the more competitors in the market equals more ways you split revenue pie equals marginal investment is less attractive.

So what do we do? We can heavily regulate these monopolies/duopoly (Title II) but that doesn’t fix the lack of competition. Prices will still be high because you only have one or two companies providing “Broadband” (25/3) in a given geographical area. There is no incentive to price compete when there will never be new entrants & no substitute products (wireless broadband wont cut it). So we’ll be stuck with high priced broadband that doesn’t block Netflix, that’s OK for now, but doesn’t address long term competitive issues.

We could let local PUC/Corporation Commissions regulate price for broadband, same way they do investor-owned utilities. But lets step back and look at this from a least-regulation position. Mono/duopolies shouldn’t be allowed to do whatever they want. So lets try and get rid of the fact they’re monopolies in the first place.

How do we increase competition? Government can’t force laborers to work cheaper to install the fiber, nor can they make the materials cheaper, so capital costs remain high. If we force companies like Comcast or Verizon (I-ISPs) to open up their lines to competitors (C-ISPs), it will never be an even field. Comcast or Verizon will always slant things to benefit them, in pricing or response time to fix issues, and run C-ISPs out of business. You think Comcast treats their own customers bad now? Imagine being a customer of a C-ISP using Comcast’s pipes.

We need a company to lay fiber that will resell transport but not compete with its customers (the C-ISPs). A municipal agency (govt) that owns the fiber and maintains it would be a good option. Better than your city/county maintaining roads because its a dedicated agency and the money coming in from the C-ISPs stays in the agency for upgrades and improvements (the digital highway trust fund cannot be raided to pay for other things). Limited scope by govt charter, regulated by the state to prevent scope creep. One mission, one goal of building 21st century roads. The agency has a fee structure for (flat + per Mbit + per Mbit/sec) companies who want to sell access. Companies are free to add value (alarm systems, television, phone, etc.). The net neutrality issues go away because of the high degree of competition – if company A slows down Netflix, switch to B, C, D or E.

If we put this plan in motion, existing ISPs try to run other companies out of business. We need regulatory help to prevent that. The optics get difficult, “XXXX is responsible for your higher bills!” they’ll say, but ultimately its to prevent anti-competitive behavior by the large incumbents. Without it, incumbents will wage a price war longer than the government and the competitive ISPs can sustain, which would spell doom for the entire program.