BMW i3 Production Process

Many are wondering whether Apple’s collaboration with BMW will give clues as to the future shape of the Apple Car.  Some have even gone on to speculate that Apple may use the BMW i3 as a design or basis for its own electric car.

This morning, technology analyst and avid Apple-watcher, Horace Dediu ( retweeted a 22 minute clip of the BMW i3 production process.  His accompanying tweet cryptically read ‘Watching how BMW makes the i3, it’s obvious why Apple had a chat.’

From this first clip alone (it is part 1 of 4 clips) we see two very important features of the BMW i3 production:

i) High levels of automation – very few people are involved in a predominantly automated process

ii) High levels of automated carbon fibre production – This is important because carbon fibre is traditionally a cost choke point due to its difficulty of manufacture in large quantities as this quote from Wikipedia suggests (CFRP stands for Carbon-fiber-reinforced polymer):

‘CFRPs can be expensive to produce but are commonly used wherever high strength-to-weight ratio and rigidity are required…’

It is clear from the video clip that BMW is able to manufacture carbon fibre at scale and that the i3 has a substantial amount of carbon fibre in it.

In an earlier article I speculated that the Apple Car will need to be extremely light, but also extremely strong.  Such a material would cause a ‘virtuous cycle‘ for an electric vehicle of being a) reduced weight since batteries weigh a lot b) better performance through better power to weight ratio c) cheaper as batteries are presently one of the most expensive components of an electric car.

Specifications of the Apple Car

In this piece I drill deeper into speculating what the Apple Car may be like, contemplating its likely specifications and performance characteristics, based upon existing cars.

Following on from my piece that sought to describe the physical parameters of the Apple Car, in this piece I go one step further (too far?) and attempt to apply performance characteristics to the Apple Car. Using specifications from existing and upcoming micro-cars (REFERENCE LINK), I attempt to extrapolate the likely possible specifications for a future ‘disruptive’ micro-car[1], scheduled for 2019-21 release.[2] The existing micro-cars that I referred to, and their specifications can be found on the next blog post here.

For the purposes of our exercise, we anticipate that the future ‘disruptive’ vehicle will have the following characteristics:

Passengers: 1 adult (with some type of convoying technology required to link other cars of the same type, either ‘in-line’ or side-by-side.) In Australia research suggests that over 90% of trips only carry the driver.[3] But note, that percentage would count a trip to drop off the kids at school as 2 trips, with one of those trips, the return trip, likely to be only 1 passenger.]

Dimensions: Not much bigger than an electric wheelchair – perhaps slightly longer and wider for safety reasons and cargo capacity i.e. Length: Approx 1.5 to 1.6m; Width: Approx 1m: Height, Approx 1.35 to 1.6m (similar to a Mini, 1.4m, or ‘Smart Fortwo’, 1.56m)

Weight: Less than one quarter the weight of a conventional family sedan, or 300-450kg; Less is more due, to the weight of batteries. I anticipate it to use super-strong lightweight materials like carbon-fibre, perhaps custom-made for the ‘Apple Car’ similar to Gorilla Glass or the gold alloy used in the Apple Watch. Note, the Morgan EV3 is said to be less than 500kg and will be larger than this vehicle. I therefore anticipate it should be capable of reaching 2/3rds to 80% of its weight. However, it is also likely to have more ‘mod cons’ than the Morgan EV3 (e.g. a ‘hardtop’ roof; air conditioning; entertainment system; ‘smart’ technologies/sensors etc, which might take the weight from say, 400kg to 500kg.)

Engine: 30kW to 55kW (I anticipate it to be similar to the electric Smart Fortwo, or slightly less to give it similar performance but with lower weight.)   Weight calculation: [Est. 400kg + 100kg (large male) = ] 500kg vs [880kg +100kg (large male)] = 980kg. Consequently, I anticipate a 30kW engine could have the same performance specifications as the electric Smart Fortwo. Elsewhere I suggest that those performance characteristics are all that are needed.

Battery capacity: Approximately the same as for the Smart Fortwo i.e. 17.6 kW·h lithium-ion battery by Deutsche ACCUmotive[44]

Range: Approximately 200-300km. This should account for more than 95% of trips.[4] 145 km (90 mi) range is available from the electric Smart Fortwo. Note, the range could be much higher considering the anticipated reduced weight of the proposed Apple Car. Consequently, it may be possible to have a smaller battery, reducing weight considerably. I think the weight/battery/performance/range equation will be a very well optimized balance.

Top Speed: Not capable of doing much more than maximum speed limit in most Western Countries’ i.e. 125 km/h (78 mph). This is the top speed of the electric Smart Fortwo. This speed was chosen because Apple has a strong tradition of not competing in ‘specification wars’, eschewing adding specifications for the sake of them, and instead aiming for qualitative benchmarks. For example, its iPod was not the smallest music player, nor the music player with necessarily the largest memory. Instead it went for ease-of-use. Likewise, the Apple Car will not be built for the purposes of drag-racing conventional motor cars. It just needs to get the passenger/driver from A to B.

Price: Comfortably below multi-passenger micro-cars, with multiple Apple Cars being about the same as a mid-luxury family sedan (e.g. Honda Accord) i.e. Sub US$13,000. Preferably under US$7-10K. Note, because it’s only a single passenger vehicle it may need to be substantially cheaper than most of the two seaters to provide a convincing ‘value proposition’. This is also why the ‘convoying’/platooning capability described in the earlier article is so important. There may also be economic pressures for this vehicle to be a subscription vehicle or some other business model of usage/ownership. (See other article on ‘Thinking behind Apple Car speculation’). Most micro-cars are sub US$15,000. It may be possible to achieve price ranges below US$10K with sufficient economies of scale e.g. Dediu’s suggested ‘1 million car’ mark for an Apple Car to be ‘meaningful’.

Smart Technologies: Pontooning/convoying’ technology will be important to allow for the Apple Car to disrupt the family car. An example of this concept is given for the EO Smart Connecting Car 2.

The EO Smart Connecting Car 2 imagined in 'convoying' mode

The EO Smart Connecting Car 2 imagined in ‘convoying’ mode



[1] Due to the highly speculative nature of this article, I am attempting to cover my bases here. Perhaps if Apple doesn’t make this, someone else will???

[2] According to the Wall Street Journal (WSJ) the Apple Car is scheduled to be released in 2019. Dediu notes this usually means the product would be available to the public one year later (2020) at the earliest. More recently, Tim Cook, when asked about the Apple Car did not deny the rumour, but instead implied it was a lot further away than people were expecting, saying: ““Do you remember when you were a kid, and Christmas Eve, it was so exciting, you weren’t sure what was going to be downstairs? Well, it’s going to be Christmas Eve for a while.” Source:


[4] and

What will the Apple Car look like?

This article provides a playful look at what the Apple Car might look like. For the (slightly) more serious reasoning on how I came to the parameters of the possible Apple Car, please click here and for the performance characteristics click here. For the specifications of existing micro-cars I used as reference points to inform the parameters, please click here.

Duplo model courtesy of my 4 year-old daughter

Is this what the Apple Car might look like? (Duplo model courtesy of my 4 year-old daughter.)

In this piece, I seek to flesh-out and illustrate the likely ‘envelope’ and specifications of the Apple Car. In an earlier post, I described the broad characteristics of what I imagined the Apple Car to look like, drawing upon the thinking of well-known Apple observer and analyst, Horace Dediu.

Primary Parameters for the Apple Car

Together, Dediu’s criteria and my own reasoning pointed towards the primary characteristics relevant to visualizing and specifying the Apple Car as being:

  • A small vehicle, likely a ‘microcar’ or ‘autocycle’
  • It would fit only one or two people – we will assume one person here
  • It was a given that it would use a large amount of ‘smart’ technology e.g. autopilot, collision prevention, auto-balancing/leaning technology etc., but only that likely to be available at its speculated time of release in 2019-2021.
  • It would likely be electric
  • It would be unlikely to compete with the specifications of a conventional vehicle, making performance trade-offs to more specifically focus upon the job to be done (taking a person from ‘A’ to ‘B’)

    A comparison of a (very) rough scale model of the Apple Car to the contemporary Mini Cooper. In the photograph, the stack of Duplo blocks at top right is a rough proxy for a 1.75m (5’8″) person. For my calculations on the models and scales click here.

Dimensions of the Apple Car

Consequently, I arrived at the following dimensions for the future Apple Car (assuming of course, one is ever made):

Length: From 1.2 to 1.6m long or comfortably less than half the length of the average modern family sedan[1]. An important criteria is that the vehicle can park ‘nose to kerb’ and not be wider than a conventional car.

Width: Approx. 1 metre; or more than half but less than 2/3rds the width of the average modern family sedan. This is to enable the division of the regulation traffic lane into two, hence potentially doubling the carrying capacity of existing infrastructure.

Top view of a scale Apple Car model to the Mini Cooper. Note: Four Apple Car’s could be linked together in a 2x2 pattern and be roughly the same width and length as a family sedan. No more arguments over air-conditioning temperatures!

Top view of a scale Apple Car model to the Mini Cooper. Note: Four Apple Car’s could be linked together in a 2×2 pattern and be roughly the same width and length as a family sedan. No more arguments over air-conditioning temperatures!

Height: 1.35-1.6m or around 10-15cm less than the average modern family sedan. Note this dimension is one of the most constrained due to the assumption of a normal seating position. Going too far from a normal seated position risks alienating many people (the old, inflexible, tall, overweight, unfit, unusually proportioned etc). Historically, this is something Apple has sought to avoid.

Side-on view of a (roughly) scale Apple Car model to the Mini Cooper. Note, having owned a Mini Cooper, the seating is quite low. It will be difficult to push much below the 1.4m height of the Mini Cooper, unless the driver’s position is reclined steeply.

Side-on view of a (roughly) scale Apple Car model to the Mini Cooper. Note, having owned a Mini Cooper, the seating is quite low. It will be difficult to push much below the 1.4m height of the Mini Cooper, unless the driver’s position is reclined steeply.

Figure 5. Rear view of the Apple Car model compared to a model Mini Cooper

Rear view of the Apple Car model compared to a model Mini Cooper

[1] For comparison, the Honda Accord is 4.86m long. See the blog post here for the vehicles I have used for reference.