Dear Elon: On Trucks & Autonomy, Etc. (in 2015)

---

Support CleanTechnica's work through a Substack subscription or on Stripe.

---

Recently, I wrote a piece for CleanTechnica on the economics of autonomy, seeking to explain the vast implications of negative depreciation. (The ability to buy a car that is worth much more to the customer over time than its sticker price and its total cost of ownership.)

While I thought it was possibly too much too soon to explain the next step beyond relative cost benefits of the Tesla Model 3 within the ordinary sales model versus the BMW 328i and the Toyota Corolla, Elon Musk apparently felt otherwise and released the Tesla Master Plan Part Deux, essentially confirming the business model for autonomy presented here on CleanTechnica.

Elon Musk, July 20th, 2016: “the fundamental economic utility of a true self-driving car is likely to be several times that of a car which is not” [but despite being worth multiples of an ordinary car, it costs no more to make or to buy than any other mid-priced sedan, and it competes with legacy vehicle production and the total cost of ownership on a per-mile basis of every other car on the road] — my words added in italics.

Here’s a feel for what this means from one of my comments on that “economics of autonomy” article, published July 1st, 2016:

If you buy that car for $35,000, I am suggesting an obvious business model in which Tesla gives you a button on your smartphone that says, “release my car to the Tesla Fleet.” If you hit that button, Tesla would be paying you 30% of that $1 per mile and for the remaining 70%, or 70 cents, Tesla takes care of everything — from handling the hailing and billing to supplying the energy, service, and valeting. If we stick with the round 1 million miles, you hit that button and you get $300,000 in the space of about 5.5 years in return for buying that $35,000 car.

Regarding Timescales:

As Elon wrote: “We expect that worldwide regulatory approval will require something on the order of 6 billion miles (10 billion km). Current fleet learning is happening at just over 3 million miles (5 million km) per day.”

It is worth reflecting on the fact that, on a straight-line basis, 3 million miles is about 5.5 years away to reach 6 billion. It is not a straight line. 3 million miles daily corresponds to a fleet of circa 120,000 vehicles; 120,000 Model 3 makes circa 6 million miles per day; 240,000 makes 9 million miles per day; and so on. This is discounting additional Model S and X vehicles that enter the fleet prior to the Model 3 debut. At a fair estimate, even including delays for contingencies, we are looking at 2018–2020. This is not what Tesla’s would-be competitors and “experts” across the automotive and finance sectors are anticipating. This affects the owners of the very first Model 3s. It would also explain why Elon has suggested ruefully that Apple missed an opportunity to get going sooner. Especially considering Apple is apparently pushed out for a 2021 launch (article coming soon).

We don’t yet know the split Tesla plans to make between customer revenues for hitting that button to join the Tesla Fleet and the amount Tesla will retain for running the service and building production capacity. Both numbers will be very significantly in profit for both the customer and for Tesla. The economics gravitate to this optimum economic model and this is an economic positive feedback loop to sweep away the internal-combustion-engine era in a breathtakingly short timespan. I have known and understood this for years and have by various means been determined to ascertain that Elon was aware of it. This is something that makes emissions standards and carbon tax initiatives a “nice to have” at best and a “corruption magnet” at worst.

This is the Real Deal: 

The massively distributed capital economy of autonomous transport and distributed energy. Two components of the Internet of Things Era. Not because we like to hug trees, but because, by comparison, oil and ICE absolutely cannot compete for either capital or consumers and, with that, standards of living rise and not fall with the end of oil.  (I am writing a book to explain how all of this works with a working title of The Future History of Energy and Transportation — just in case anyone might be interested).

I think while many people are still grappling with the significance of Tesla’s latest announcement, many others anticipated autonomy and autonomous robotaxis in general.

What perhaps nobody saw coming at all was a Tesla semi truck.

Well, almost. I referred to it as a Class 8 Truck back in 2015 in a letter to Elon.

Feel free to check it out (this letter has been very mildly redacted in the interests of confidentiality of some proprietary concepts):

[Editor’s note: the letter has only been slightly modified by me to make minor changes to formatting and add an editor’s note in one place. –Zachary Shahan]

Julian Cox
julian@liionica.com
October 14^th 2015

Elon Musk
Chairman, Product Architect and CEO
Tesla Motors Inc.

Dear Elon,

Thank you sincerely for taking a moment of your time.

Here is some prior precedent from July 11th 2015 regarding Tesla’s Autonomous Transport as a Service insights. A subject later enquired about by Morgan Stanley’s Adam Jonas on the August 5th Earnings Call: http://www.quora.com/Who-is-Elon-Musk-competing-with/answer/Julian-Cox

The link is just some jottings on an internet site. What was far more interesting and remarkable to me was listening to your silence for a full six seconds on an earnings conference call when those thoughts were posed to you as a question and described as insightful. Long enough to imagine that I might not be wasting your time by reaching out.

Naturally when seeking to transition the world to sustainable transportation, it will be deeply helpful for Tesla to progress from delivering evolutionary change (riding/driving down the cost of batteries to improve incrementally on the cost-benefits of electric vehicles versus internal combustion engined vehicles) to delivering revolutionary change with autonomous electric vehicles with central OTA monitoring, company-owned automated and globally distributed field charging and servicing infrastructure that in combination can serve to replace ten or twenty internal combustion engined vehicles per unit of Tesla production.

The latter seems like an economic model for personal transportation with a total cost per passenger mile that after fuel costs, reduces the market value of producing internal combustion engined vehicle hardware to approximately $zero.

I have been excited to see the pieces of this puzzle come together including fleet-wide collaborative machine learning for mapping and the ‘snake’ charger with its implications for automated re-parking at Supercharger sites and ultimately fully autonomous fleet charging. As far as I can see the last piece of the puzzle is the charge door. I think it just needs an actuator to close autonomously when the snake unplugs unless perhaps the snake can be programmed to push the charge door shut. (Note since the time of writing this, the charge port doors on newer Teslas close themselves with an actuator).

With regards to Uber and similar attempts to produce the Napster of automotive. It may very well be that a leading inefficiency of the prevailing automotive business model is asset underutilization however the counterbalancing inefficiencies that arise from using an ICE vehicle in terms of fuel, maintenance, depreciation and a whole range of issues surrounding drivers occupying seats on journeys to places they don’t want to go strongly rebuts that case. I suspect in reality that a lot of Uber’s economics are actually driven by exploitation of owner-driver underestimation of the costs of vehicle maintenance and mileage depreciation. While Uber and similar clearly connects an oversupply of casual labour with a latent demand for service, in the context of an autonomous transit system I cannot see that Uber can either compete with Tesla or in any practical sense add any value at all.

Just as it seems remarkably hard for the world beyond Tesla/SpaceX to grasp the concept of a ground-level stratosphere in a tube for wingless, externally powered short-haul aircraft (the Hyperloop), there is empirical evidence that it is not widely appreciated and understood that Tesla’s current vehicles function as collaborative machine learning scouts building the virtual rails of an advanced global transit system.

It has been a fascination to witness the financial community (at least in public commentary) fully fail to grasp the significance of Tesla’s unique ability thrive in a vehicle market that Tesla’s technologies may indeed shrink in terms total unit volumes by as much as an order of magnitude.

Media quote: Thilo Koslowski, vice president and automotive practice leader at Gartner Inc., said [regards ride sharing] ‘this is certainly something Tesla would be considering just like other global automakers, including Ford Motor and BMW that already are experimenting with shared mobility services in select locations. Uber’s success is inspiring a lot visions for other companies and it would make sense for Tesla to do this as well’. End quote.

If this is what stands for top tier automotive expertise it occurred to me that you might agree that a contrasting perception of reality is both rare and potentially useful.

In an effort to explore the concept I have shared some thoughts below. I have no access to insider information and many if not all of these ideas may already have been considered and are either in the pipeline or discarded for good reason.

For example, I have some ideas for the Tesla Powertrain Business Development group: 

Tesla AC Induction wind turbines.

Regards wind turbines, while they won’t meet SpaceX ‘useful for Mars’ test, for Tesla it seems to me like a natural fit to re-use the AC induction motor, inverter and their cooling systems, regenerative braking algorithms, mechanical components and OTA data communications, even a subset of the GUI to produce something state of the art. Such a device would create a significant additional product category within the Tesla Energy product portfolio — ambient kinetic energy capture. When twinned with Tesla Powerpack I see this as extending the market for what is essentially a hugely simplified derivative of Tesla’s pre-existing vehicle technologies and I think that this is helpful in rounding out Tesla Energy. In particular I think common platform monitoring, control and dispatch (smart-grid) protocols would make turbines easy to purchase from Tesla and to integrate with Tesla’s battery technology.

On a point of physics and economics I have a feeling that the complex and wide dynamic range AC induction control algorithms developed for the vehicles holds in prospect an elegant solution to wind turbine build-cost and efficiency issues traditionally encountered by approaches relying upon mechanical transmissions and brakes. Tesla’s AC induction technologies can handle fluctuating wind-force and centrifugal force limitations in code governing negative torque. A very large turbine could comprise multiple standard generators on a single shaft. As a bonus, it would be nice to focus public attention to the merits of AC induction – for example by defeating classic fossil fuel lobby touch points surrounding solar and overnight charging and criticisms of rare-earth mining and refining footprint levelled at traditional wind turbine generators that are based on the use of permanent magnets.

Co-location with solar would also appear to offer the prospect of a reduced ratio of storage to generation when aiming for an acceptably reliable 24/7/365 supply at a reduced levelized cost per kWh. Additional benefits would seem to include improved economies of scale and production load-balancing for the Tesla drivetrains and a Tesla AC Induction turbine would appear to compliment the SolarCity installation and finance aggregation business. Perhaps it is not too far fetched to imagine SpaceX contributing at least to the design and possibly the fabrication of blades or alternative aerodynamic generation mechanisms – including the possibility to explore Nikola Tesla’s turbine designs — and the design if not the fabrication of stir-welded tower structures. Tesla turbines would also appear to be a good thing to purchase internally rather than externally in relation to power generation for the Gigafactory and Superchargers, even Fremont, and as a strong visual cue for Tesla along highways and wherever else they are sited.

Finally on this subject. Unless I am mistaken, the Tesla Powertrain Business Development group appears from the outside looking in to be excessively limited in ambition to trying to support third party light automotive. A sector that is somewhat predictably intent on cooperating with being disrupted rather than cooperating in a more classical sense — whereas the ability to widen that scope to include some relatively straight forward productization, for example in the form of wind turbines and in other seemingly obvious opportunities such as offsetting hill-climb energy demand with regenerative descent for Class 8 trucks, would I think enable the Powertrain Development group to knock it out of the park – and be ready with self-funded capacity increases to accommodate internal and external demand for motors and controllers.

Marine Drive Train

For example there is no inherent reason I can see besides management bandwidth why there is no Tesla Marine propulsion derivative of the Tesla drive unit — and yet the risk suffered by any third party entity to bring a comparable marine drive to market is likely to remain prohibitive for many years to come — not just because R&D would need to start from where Tesla was in 2003 but also as a result of concerns that Tesla may one day get the idea to put a drive unit and inverter in a choice of outboard and inboard mounted packages, trim the software and GUI for marine use and obliterate any third party R&D capital invested overnight. This is perhaps a serious hidden barrier to market advancement, an invisible wall if you like, resulting from Tesla’s commendable but daunting technological lead.

It would be great for the world to finally have a speedboat with millisecond resolution variable torque linked to an accelerometer that actually ran level across waves. [Editor’s note: This has been requested by another CleanTechnica author as well.]

Class 8 trucks

In the case of Class 8 trucks I think there is a stunningly attractive market for highway lane-keeping autopilot with elevation-aware least-cost routing. This, together with hill-climb assist and regenerative descent (possibly emissions-free first and last mile operations too) could be as straight forward in hardware terms as developing a kit to attach a Tesla drive unit to a prop-shaft plus some pack-mounting brackets and a heat-exchanger. By my estimation there is a path to a customer value proposition here that is in the order of 500% of Tesla’s build costs in fuel savings alone – plenty of room to structure a deal to bring it to market in any way Tesla sees fit. Again if it is a case of management bandwidth availability to drive innovations like this then this is why I am writing to you.

This stuff is obvious to me at a gut level and I absolutely recognize and share the depth of your emotions that I have seen for example in your recent opening comments in Berlin. Obviousness of what is the right thing to do should lead more directly to action in society than it generally does. The disconnect I think is fear, and that can be addressed effectively by leadership [from the Tesla Power Train Development Group].

Many of these ideas for the Powertrain Development Group are clearly far more practical than Hyperloop to seed in return for relatively minor management overhead.

Some ideas in relation to Tesla vehicle Sales and Marketing. 

It occurs to me that there are lost opportunities that can be addressed in the central sales funnel of the business: The online Design Studio. I think the online and app presence could readily become a more vicarious ownership experience than it is because in my opinion highlighting gasoline cost savings and lowering the bar of refundable reservation commitment is not a catch-all for every personality type that is a natural customer for Model S or Model X.

For example, at a minimum in my opinion the user should have the ability to save a desired configuration between visits, which could and probably should be extended to create the possibility of a pre-reservation account [for all manor of reasons].

I think it is important to have a vicarious feel for ‘your dream Tesla’ to encourage the act of hitting the reserve button. As such I don’t think that the Design Studio should be overly self-selecting in appeal to impulse shoppers, those considering a refundable reservation of some few $1000s as close enough to $0 to be immaterial, or those willing to redesign their ideal vehicle numerous times only to have it evaporate in-between visits. I suspect that there is a significant contingent of the early mainstream that will only hit the reservation button having considered the purchase to the point of a final decision. I would suggest extending the welcome mat to fully engage these additional personality types on their terms.

Extending the Function of the Tesla Mobile App as an Interactive Media Channel [To incorporate pre-reservation sales and broad public/consumer education regards what it is like to live with a Tesla EV].

To take the concept closer to its full potential as a precursor to reserving Model S or Model X and also to engage the broad pre-sales market for the Model 3, I think there should be the option of viewing and interacting vicariously with a Tesla vehicle in its saved state within the Tesla mobile app…

[redacted content].