Featured Post

This is the Kodak Moment for the Auto Industry

Plug-In Drivers Not Missin' the Piston Electric vehicles are here to stay. Their market acceptance is currently small but growing...

Monday, March 20, 2017

The Model 3 Won't Ship in July (and that's OK)

Optimistic fans hope Tesla will ship 80,000+ Model 3 this year. Goldman Sachs, on the other hand, downgraded Tesla's stock rating due to Model 3 delivery concerns. The truth is likely between the two. Musk has given many hints to the sources of delay that will likely impact Model 3. We examine his statements and, in light of these, attempt a realistic estimate.

Full Story
Despite the popularly held belief, Tesla will not likely start to ship the Model 3 in volume in July of this year. But if you live on the west coast, you might see yours arrive this year.

There is a lot of concern about the Model 3 and when it will ship. Goldman Sachs recently downgraded Tesla's stock to sell, citing near-term challenges including the launch of the mass-market Model 3.

So when will it ship and what do we know? Tesla provided a lot of information in their earnings call on February 22, 2017. At one point, discussing production risks, Elon Musk said, "I'm just going to tell you everything I know so you can have the same model in your head that I do." Of course, Musk could never convey all that he knows in this realm in one short phone call, but it did provide a glimpse of how he views risk management.

One of the large relevant chunks of information was that Tesla has asked suppliers to deliver 1,000 parts per week starting in July. This is to be followed by 2,000 parts per week in August, then 4,000 per week in September, ramping to 5,000 per week by the end of the year. If all of these parts could be assembled and delivered with a snap of the fingers, Tesla could deliver about 90,000 cars this year. But, this assumes all the suppliers will meet this delivery expectation and those cars magically assemble and deliver themselves. So, for 2017, we'll consider this the "lightspeed" or the impossible-to-reach upper bound of vehicle delivery.

At a high level, there are three stages: receiving parts, assembly, and delivery. Let's look at each stage and see what obstacles Tesla must overcome in that area.

Stage 1: Receiving Parts 

Supply chain logistics is not as easy as it might seem. Tesla has to deal with dozens of suppliers. The figure below gives you an idea of what they deal with for the Model S.

Model 3 will be a simpler vehicle than Model S, but it will still depend on dozens of other companies. Tesla can only deliver at the rate of the slowest one. This is the theory of constraints. When you start looking at second and third level suppliers, the list quickly becomes thousands of suppliers. It becomes a web of dependencies. So you must have contingencies and non-correlated alternatives... It takes a lot of planning and management to keep all the parts flowing.

Tesla has told all of their Model 3 parts suppliers to start delivering parts in volume in July. Will all the suppliers be ready and deliver in July and ramp at the rate that Tesla wants them to? The simple answer is 'no'. There will be some that are late. Musk referred to this as the "Term Paper Problem" in the Q4 2016 Results Earnings Call. Here's his quote from the transcript:

We have, what I call, the term paper problem. I was a teaching assistant in college and no matter what date we set for the exam paper, when the term paper was due, there's always some number of people that are late. It's just the way it goes.

Musk goes on to explain that when you have a global supply chain, you inherit "force majeure" risk from around the world. In other words, if there is an unavoidable major disaster that impacts one or more of their suppliers, then it also impacts Tesla's ability to deliver cars. They are taking steps to minimize risk where they can but if you don't know which supplier will be impacted or what will hit them, it is hard to prepare.

At Code Conference 2016, Musk explained this in more detail:

Think of any natural disaster you could care to name—all of those things have happened to our suppliers. A factory has burnt down, there’s been an earthquake, there’s been a tsunami, there’s been massive hail, there’s been a tornado, the ship sank...

These risks don't even have to be a major event. Musk also told the story of trunk carpet that shut down the production line for a few days.

There was a shootout at the Mexican border—no kidding—that delayed trunk carpet. The Border Patrol wouldn’t give us the truck because it had bullet holes in it. We just wanted our trunk carpet... That shut down the production line for several days.

That's right, one of the most technologically advanced things you can buy today was delayed for something as simple as trunk carpet. So, if you assume at least one or more supplier will have some issue and miss the July starting date, then there is the first delay. To address this, Tesla has several options: find another supplier, make it themselves, work with the supplier to address the problem(s).

Key factors: Theory of Constraints, Term Paper Problem, Global Force Majeure.

Let's make the very optimistic assumption that all of these issues can be resolved in three weeks.

Stage 2: Production

Tesla is a manufacturing company. You might think of them as a car company, a design company, or an innovation company. Yes, they have these aspects, but if all they could make was low volume bespoke cars, like the Roadster, then (no matter how great these cars were) they could never reach their mass-market goal.

Tesla learned a considerable amount about designing for manufacture with the Model S and X. These cars were designed, for the most part, without much consideration to manufacturing. They are hard to make and they have had delays and issues related to this manufacturing complexity. Tesla has brought their designers into the factory to see these issues as they come up. The designers get to talk to the people on the line and see the issues that their designs are causing. They have applied these learnings to Model 3.

The massive roof opening of the Model 3 might be a cool design feature, but it's there first to allow the robot arms access into the car's interior during manufacturing. Ease of ingress and egress will allow the arms to move faster and allow for multiple operations to occur simultaneously within the body.

Here's Musk from the Q4 2016 Results Earnings Call:
Model 3 is designed for manufacturing. It's considerably simpler than Model S or Model X. Model 3 has 1.5 kilometers of wiring. Model S has 3 kilometers of wiring. A lot of the bells and whistles that are on Model S and X are not present on Model 3. We don't have self-presenting door handles, for example, or falcon-wing doors.

For Model 3, Tesla has turned their attention to "the machine that builds the machine." Musk said that Tesla is applying the rocket equation to manufacturing. The rocket equation considers mass efficiency and rocket velocity; but in a factory, it's volumetric efficiency and the exit rate (velocity) of products. This is a novel approach to manufacturing that only a rocket engineer would conceive.

From the earning call, Musk said "I've refocused most of Tesla engineering, including design engineering into designing the factory. I think in the future, the factory will be a more important product than the car itself," and from Tesla's blog "our factories are so important that we believe they will ultimately deserve an order of magnitude more attention in engineering than what they produce."

Musk has said that ultimately a Tesla factory will look nothing like any factory that has ever existed. He said it will look like an alien dreadnought. The dreadnought changes that they have made will likely pay big dividends when they want to move to 500,000+ vehicles per year, but doing something in a way that has never been done before could have a startup price to pay as they bring the alien dreadnought online for Model 3.

Key factors: New Design, Boot the Dreadnought, Production Ramp Hell

Let's assume these factors have a minimal delay of 2-weeks above and beyond the above delay.

Stage 3: Delivery 

Vehicle delivery is that all-important final step. Even here, there can be delays. Musk told a story of a cargo ship carrying Teslas that was not allowed to dock due, ironically, to excessive smog in the port.

Tesla will avoid issues with boats by starting delivers to employees near the factory first. This allows them to easily bring the car back in to examine any failure and, if needed, make changes to the production line.

Starting with Tesla and SpaceX employees will mean that non-employee customers will have to wait for these deliveries before they move up in the queue.

Musk from Q4'16 Results call:
The initial cars, sort of Founder Series, actually go to company employees, because it's important to have a good feedback loop on the product we're making. And if there are any issues, bugs, or things that need to be addressed that we can address those before customers experience them.

After the SolarCity merger, Tesla has about 30,000 employees. SpaceX adds about another 5,000. That's 35,000 people that could be in line ahead of the first customer that does not work for Musk. How many of them have ordered a Model 3? It seems likely that at least 2,000 people that work at Tesla/SpaceX would be excited to get one of these cars.

Key factors: Employee deliveries first

This pushes the first delivery to someone outside of the company at least another month.

So When Do I Get My Car?

When pressed in the conference call the Tesla executives consistently said that the product ramp is not possible to predict. There are too many unknowns. But I see no reason for that to stop us from guessing.

We know that parts will start arriving in July. Recapping our potential delays above, there's the Term Paper Problem, Booting the Dreadnought, and Employee Deliveries. Assuming each of these are resolved quickly we could see non-employee deliveries starting in mid-August.

On the pessimistic side, if one of the items hit by the Term Paper Problem is a long lead item, this impact could be much bigger. Then, continuing on the pessimistic track, the employee deliveries could provide feedback that requires tweaks to the car and/or factory. This could further delay the non-employee deliveries until much later in the year.

Once the cars start shipping, there are several factors that determine your place in line.
Vehicles will go to current Tesla owners that live near the factory that reserved a car on day-one in a Tesla store. Then cars will go to people on the west coast US with the same qualifiers.

Generally, Tesla delivers cars that have a higher price tag first. This allows them to collect money for higher revenue vehicles sooner. This might not be the case for Model 3. E.g., if they are motor constrained, they may opt to deliver single motor cars before dual motor vehicles so they can deliver more cars. Similarly, if they are battery cell constrained, they may opt to deliver more units of the smaller pack vehicles rather than fewer large pack vehicles (assuming there are pack size options). This is unlikely, but possible depending on the constraints that present themselves.


What Tesla is trying to do is very hard. They are trying to take a car that is not yet in production and deliver tens of thousands of them in less than half a year. I've seen estimates that Tesla could deliver up to 80,000 Model 3s in 2017. This is nearly impossible (not totally, but nearly). Let me put it into perspective. In 2016, Tesla had an entire year's production and they delivered about 76,000 vehicles (source). Tesla had challenges with Model X production in the first half of the year and they had short-term production challenges in Q4 with the transition to the new Autopilot hardware 2.0. So yes, Tesla has delivered ~80,000 (with production challenges), but this was for the entire year.

For another comparison, in the first 2 months of 2017, GM is only delivering about 1,000 Bolt EVs per month. Tesla wants to get to 5,000 per week.

Model 3 will have ramp-up pains and less than half a year of production. As we've seen with every EV that has been delivered in the last decade (Tesla's included), the production ramp has been slow. Given this, I estimate that there will only be 2,000 to 4,000 Model 3 vehicles delivered in 2017. This is far fewer than the 80,000 unit prediction that others are making and still fewer even more than the lightspeed 90,000 number would allow. I hope this is woefully low, but given all the unknowns, it seems reasonable.

These 2017 units will be primarily consumed by Tesla and SpaceX employees with a few going to current Tesla owners on the west coast with day-one reservations.

In 2018, Things Turn Up To Eleven

In 2018, all these production issues should be worked out to allow Tesla to slowly turn the production volume knob up to 11. Tesla will hit their 5,000 per week production goal in Q2 of 2018 rather than December of 2017. A second production line in the second half of 2018 will continue to ramp up the volume of vehicles produced allowing Tesla to move to more than 7,500 vehicles per week by the end of 2018. This will allow all the current reservation holders (as of this March 2017 publication date) worldwide to receive their car by end of the year 2018.

Monday, March 6, 2017

Musk of Mars

Disclaimer, Disclaimer, Disclaimer: The below article makes inferences, speculations, leaps of logic, and several WAGs. This is just a blog, not PLOS ONE; please treat it as such.

It's not a stretch to say that Elon Musk is obsessed with Mars. He wants to make life multi-planetary (read populate Mars and beyond). He has even said that he wants to die on Mars, "Just not on impact." In 2002, he started SpaceX with the explicit objective for the rocket company to become the primary means to populate the red planet.

It's important that we attempt to extend life beyond Earth now. It is the first time in the four-billion-year history of Earth that it's been possible, and that window could be open for a long time - hopefully it is - or it could be open for a short time. We should err on the side of caution and do something now. ~Elon Musk

SpaceX is not Musk's only company. The other companies that Musk runs promise to make Earth a better place with renewable energy production and storage, zero-emission transportation, and high-speed rail in low-pressure underground tunnels. What if these Earthly benefits are only a side-effect and not the true reason Musk funds and helms these companies? What if these are pieces to a bigger secret plan?

All of Musk's other current companies were formed after SpaceX; or more specifically after colonizing Mars was his stated goal. These companies include Tesla Motors (2003, now Tesla Inc), SolarCity (2006, now part of Tesla Energy), and most recently The Boring Company (2016). Additionally, in 2013, Musk was the impetus behind Hyperloop. He penned the 58-page first draft* for the idea for a "fifth-mode of transport" in low-pressure tubes. Musk is not involved in any of the companies currently working on Hyperloop, but SpaceX does host a student competition twice each year that draws in over 800 students including teams from California-Berkeley, Carnegie Mellon, and MIT.

Why did he create each of these efforts?

I don't create companies for the sake of creating companies, but to get things done.   ~Elon Musk

What if, like SpaceX, each of Musk's companies had an explicit Mars mission statement. Let's look at each of his current endeavors through the red-colored lens of Mars and see what they might "get done" there.


A colony on Mars will need energy. As far as we know, Mars does not have deposits of oil, coal, or methane. There are no flowing rivers that can be dammed. The thin atmosphere would not power wind turbines very well.

Mars is farther from the Sun than Earth is, so solar panel energy output will be about 40% lower than it would be here. However, the panels will not have to deal with cloud cover or weeks covered in snow. These factors and the cool ambient temperature helps to make up for the increased distance from the giant fusion reactor known as the Sun. The solar panels that go to Mars will not be the same type that you put on your roof. They will be the most efficient full-spectrum panels that we can produce. One more solar consideration is that pesky dust problem. Looking at all these factors, solar panels will likely be a significant energy source for a Mars colony.

The people of Mars may eventually derive their primary power from nuclear or a fuel source extracted from the soil. Even if solar is not the final primary energy source, it would be useful for excursions and as they expand into new areas, before the infrastructure for other methods is setup.

Solar panels may not be the only energy source, but they are very likely a significant energy component for this Mars colony of the future. When SolarCity's Gigafactory** facility in Buffalo is complete, they will be able to produce the necessary solar panels for Mars.

Tesla Motors (i.e., the transportation division of Tesla Inc.)

A colony on Mars will need transportation. There is no significant atmosphere on Mars, so an internal combustion engine, like the ones that power most ground transportation here on Earth, would not work on Mars without the supply of air to suck in. Just as the Lunar Rovers were battery powered electric vehicles, so too will be the vehicles for Mars excursions.

Similarly, if the colony becomes large enough that you need transportation within habitation areas, then a polluting internal combustion engine indoors is a bad idea when every liter of breathable air must be scrubbed. Without miles of atmosphere above you, running an internal combustion engine would be like running a gas car in your garage with the door closed. It's a very bad idea.

So both surface transportation and intra-hab transportation will be electrically powered on Mars. Tesla Motors will be able to make vehicles for these needs.

Tesla Energy 

Batteries will play a vital role in both energy and transportation. If you live in a solar powered world, you need energy storage. Batteries will provide nighttime energy needs such as lighting, refrigeration, and heating. Batteries will also be important to power the electric vehicles mentioned above.

Tesla's Gigafactory (eventually Gigafactories plural), will be able to produce batteries for Mars colonies. Mars colonies will eventually need to be able to produce their own supplies of things like batteries, but supplies from Earth will kickstart them.

Perhaps the surface of Mars will eventually be dotted with a matrix of solar powered, battery-based Supercharger stations. This could be our chance to actually have a single universal planetary standard for fast charging (on at least one planet).

The Boring Company 

Musk has recently started a company to dig tunnels. He has stated that his engineering team can reinvent tunnel boring and create a machine that will bore through the earth five to ten times faster than any of the massive earth-boring rigs that exist today.

Would you need tunnels on Mars?

Without a significant magnetic field, like the one we have here on Earth, Mars is far more vulnerable to cosmic background radiation and highly energetic events that emanate from the sun such as solar flares, coronal mass ejections, and coronal holes.

On Mars, the easiest way to avoid solar energetic particles (SEPs) and galactic cosmic rays (GCRs) is to live underground. Five meters of soil should provide a level of protection similar to that which we enjoy on Earth.

This means that, at least initially, our Martian colony will be living underground. Natural caves might provide an initial starting point, but the colony will want to expand as the population increases with more areas to grow food, living quarters, recreation areas... Drill, baby, drill (horizontally)!


Hyperloop is the idea of an ultra-high-speed train in a (mostly) evacuated tube. The ~one-mile long test track on the SpaceX campus is the second largest vacuum chamber in the world (second only to the Large Hadron Collider).

In 2015, Japan's maglev train set the world record at 374 MPH. Musk hopes that Hyperloop will be able to double this speed. But it's turning out to be a very difficult problem.

The long tubes, that Hyperloop requires, makes maintaining a low-pressure environment very difficult. The tubes expand and contract with temperature changes. This means that they have to have some sort of expansion joint to avoid buckling. This makes maintaining pressure seals very difficult. One solution to this is to go underground where the temperatures are more stable. See The Boring Company above.

However, in places like California, where the earth has been known to quake, people may not be excited to travel underground.

Hyperloop is turning out to be very difficult here on Earth. Would it work on Mars?

On Mars, things are much easier for Hyperloop. Low-pressure environments are not hard to come by. And tectonically, Mars is stable. With the planet's likely solid core, Marsquakes are not something that Hyperloop riders would need to worry about above or below the surface. Although a Marsquake may have occurred in 2012.

Mars - The Puzzle Pieces Fit Better

Batteries, electric cars, & solar panels work well both, here on Earth and on Mars. These products make sense for both planets. Looking at tunnels and Hyperloops, the earthly benefit is not as clear.

Musk has said that he wants to drill tunnels to alleviate traffic. Studies have shown that adding more traffic lanes only invites more cars. This will be true if the lanes are in tunnels or on the surface. I'm sure Musk is aware of these studies. Perhaps his vision is to add far more levels and lanes than any historical road expansion, so prior studies of roadway expansions that added a lane or two to an already congested area just don't apply. On Mars, if we're living underground, the roadways could be designed from the beginning to be multi-level.

Looking at Hyperloop, there're several teardown articles and videos that discuss many of the problems of hurling people through an evacuated tube at near the speed of sound. Again, I am sure that Musk is well aware of these challenges. Hyperloop low-pressure levels are not easy to maintain (on Earth).

Musk is a genius. I don't think that he has simply overlooked the things that a YouTuber has pointed out. I think he has a different, Martian endgame in mind. If you were building Hyperloop on Mars, the near vacuum comes for free. The lower gravity will even make the train levitation easier.

Mars - The Secret Plan 

Musk's real secret plan: develop technologies to be utilized on Mars. If they also make Earth a better place, great. Shhh, it's a secret. Don't tell anyone.

Here's the Mars vision: Solar panels collect energy, batteries to store the energy, electric vehicles to move about on the surface, in the hab, and for farming drones. Tunnels between sites (farms, habs, loading docks...) with Hyperloop trains to move people, equipment, and supplies. And, of course, SpaceX will take all the people and supplies to Mars. It all fits nicely.

It's important to note that Musk's companies may not be the ones to ultimately deliver these technologies to Mars. Rather the mere founding of Tesla, SolarCity and the others would help to create a worldwide market for these technologies that would then drive the industry to invest billions of dollars into R&D to create better batteries, solar panels, high-speed trains, and even tunnel boring machines. All technologies that Mars will need. If another company can do it better, faster, cheaper than Musk's, Mars (and Musk) still win.

If Mars is the real mission of all these efforts and it just happens to make life a little better on Earth along the way, that sounds good to me. Maybe I'll sell my home in 2030 and retire to Mars. 😃

Where does OpenAI fit in? I'll leave that as an exercise for the reader.

* Arguably Musk revived an old idea. See Robert Goddard's vactrain concept. 
   As a rocket-man, Goddard is someone with which Musk is familiar.

** The SolarCity Gigafactory is now called Gigafactory 2 and Tesla has announced plans for Gigafactories 3, 4, & 5.