Tesla Model 3: More than Yet-Another-Car: Ushering in the Energy-Transportation Nexus

On March 31st, Tesla will unveil its much-heralded Model 3, their mass market electric vehicle expected to be priced at a more affordable $35,000 (prior to subsidies). Note that the car will be shipping in late 2017 only. The Model 3 will likely be more than a car: it is likely to be a catalyst in the transformation and deeper coupling (i.e. a "nexus") of the sustainable energy and sustainable transportation ecosystems. I will not comment on its wow-features as a car - but focus this article more on the Energy-Transportation nexus aspects.

The Tesla Model 3 announcement also comes at a time where a broader set of innovation trends are converging: post-sale software-driven upgrades of EV capabilities (pioneered by Tesla), autonomous vehicle technology (auto-pilot),  on-demand shared transportation services (Eg: Uber, Ola!, Didi Kuaidi, Lyft and their pooling options (eg: UberPool, OlaShare)), decentralized energy resources (EV & EV chargers, solar, energy storage) on the electric grid, and financial innovation (eg: yieldcos, securitization, p2p financing, mobile payments). This article speculates on a subset of the potential interactions of the Tesla Model 3 with these trends.

Model 3's Asset Value & The Case for Tesla Mobility

Electric vehicles from Tesla may depreciate slower than internal combustion engine (ICE) vehicles, and have lower OPEX and per-mile costs. Depreciation is a function of perceived value by the market, wear-and-tear, and supply-demand for the product in the secondary market.

First, lets consider "perceived value". Tesla updates features regularly remotely via software over-the-air updates. Even compelling new functions like autonomous capabilities are provided as a software update. What does software-driven updates of EVs mean? The users value the vehicles more, and the vehicle is better in terms of features than what it was when the user bought the vehicle. This likely increases perceived value by the user and other secondary market participants. The user would not want to sell their vehicle, since they are getting auto-upgraded; while secondary market participants will want the vehicle.

Second, consider wear-and-tear depreciation. Certainly mechanical components of the vehicle depreciate. But there are fewer critical moving-part components of the vehicle, and the electrical components (battery, motors are warrantied and actively managed remotely for long life). Maintenance requirements in EVs in general is lower than ICE vehicles, i.e. lower OPEX. The actual wear-and-tear could also be monitored via IOT techniques. Tesla also monitors and manages the battery pack remotely.

Third, there is likely to be a large range of secondary market buyers for the vehicle. This is also a function of the number of vehicles on the market relative to secondary market demand. Another class of users would be those who would like to use the vehicle "as-a-Service" (eg: lease, rent for long distance trips from Hertz, or via Uber / Lyft etc). For instance, if I were to go back to the USA for a holiday (and plan a cross-country drive, or significant driving journey) and Hertz offered me a rental of a Tesla Model 3 vs an ICE car, given the large supercharger network and destination charging options at hotels/motels & rest areas/malls, I would likely choose the Model 3. This would be more profitable for Hertz (perhaps in partnership w/ Tesla) since they could charge me a higher rental rate, and/or per-mile charge for super-charger use, and pick up a nice margin on the transaction.

Next, consider per-mile costs. In a brilliant, but simple arithmetic analysis, (see page 127) David MacKay shows that EVs like the Model 3 could deliver 15-20 kWh per 100 person-km. If you invert this you get 5-6 km/kWh of charge. A Tesla Model 3 is therefore likely to give 4-6 km for a kWh of charge. How much does a kWh of charge cost? In USA, it costs 6c - 15 c / kWh in most states. This implies a cost of 1c - 3c / km - a remarkably low unit cost, i.e. per-mile marginal cost. Even in Australia, where per-kWh costs can be 25 c (AUD), the per-mile economics ( 5c (AUD) / km)  are compelling w.r.t. petrol costs. My small petrol car goes between 10-15 km / litre of petrol, which costs about Rs. 70 (or $1.1), implying a per-km cost of at least 10 c (USD) / km. In other words, EVs can be 5-10X cheaper on a per-mile basis compared to ICE cars. {Note: With solar PV and net metering, the highest tiers of utility costs can be shaved off, even though EVs will increase overall electricity consumption (that is good for utilities). }

A lower rate of depreciation (lets say half the rate of a comparable class vehicle), combined with low per-mile OPEX and long life, has an important financial implication: in addition to being an attractive buy option (because of the lower CAPEX), it is also more attractive to lease the vehicle, and/or share it for maximum value extraction.

Why is this so? When you lease, you tend to pay for the depreciation, per-mile costs and opportunity costs of time (eg: see youtube video on leasing economics). The depreciation is low, and the opportunity costs of time can be monetized via aggressive asset sharing and pricing by-the-mile. Eg, via Uber or the rumoured Tesla on-demand Mobility service (see here also). Another quiet trend that is happening is the emergence of taxi operators who are using Tesla and other EVs (eg: Taxi Electric in Netherlands, Oslo Taxi, Yandex in Russia, Ecocab Taxi in USA etc). This will only accelerate with the availability of Tesla Model 3.

Note that my case for such a lease / mobility / taxi service expansion is not based upon autonomous driving capabilities, but purely on the financial dynamics of asset value, depreciation, low opex and per-mile variable costs. 

Monetizing EV's Electric Grid Services & Enabling Higher Renewable Deployment

An Electric vehicle also carries a huge "stealth" value in the energy storage embedded in the vehicle. This pack has a huge value as energy storage for the grid. {Aside: I have written elsewhere about the potential of EVs even in countries like India with weaker grids (Solar + Ola! = Sola! ... ; How Electric Scooters ... can spur adoption of Distributed Solar in India) }

The Tesla Model 3 will at least have 50 kWh of battery, and at 4 km/kWh (conservatively), it will give it a 200 km range. When 100,000 Model 3s are sold (~2018-19), this implies a 5 GWh of distributed energy storage (in addition to the at least similar amount of capacity available from Model S and Model Xs), available for orchestration to provide grid services. When a million Model 3s are sold, this jumps 10X to 50 GWh of distributed energy storage (in the 2020-2021 time frame), or 100 GWh, if you include the fleet of Model S and Model X likely by then.  There will also likely be 1-5GWh of stationary storage (eg: PowerWall, assuming 100K - 500K units of 10 kWh PowerWall packs sold by then) in residences and businesses (i.e. destination charging spots) by then.

What can you do with this energy storage? You can create a tremendous spatio-temporally distributed energy resource, managed opportunistically via software, predictive analytics and large-scale IOT monitoring/control. In other words, value will be created by "software" & "analytics" in this industry by orchestrating the underlying "hardware" assets for maximum yield in multiple ecosystems (energy, transportation, finance/insurance etc). It is no longer a simple durable consumer good sale business.

For example, consider the "duck curve" problem where solar penetration can lead to a trough of net demand during the day, and a huge ramp of net demand in the evenings (when the sun sets and evening residential electricity demand grows). Electric vehicles, along with smart EV chargers (and associated PowerWall battery packs) can be orchestrated so that charging can be done when there is excess solar; and avoid peak charging in the evening. More generally, when there is a surge of renewable generation, a fleet of Tesla chargers can fire up to ramp up their charging rates (and get monetization for such smart charging functionality). This "supply following" capability of smart charging, combined w/ non-trivial storage capacity will promote the penetration of renewables both in a centralized and decentralized manner.

You can do more than coarse-grained matching of demand-supply. It is possible to do fine-grained demand/supply matching and ancilliary services, with a largely decentralized control framework. For instance, IBM Research's work (in partnership with Universiti Brunei Darussalam, UBD | IBM Centre) on "nplug", (see full paper), is a way to do micro-demand response (and other ancilliary services) in a fully decentralized way, with no / minimal explicit grid signals. Such techniques combined with any demand source with inertia such as energy storage in an EV or a smart EV charger with PowerWall can be used to do balance the local grid feeders at a fine-time scale (see a longer article on this capability in an Indian context here). It is important to observe that this space-time demand shifting capability is possible purely by modulating the fine-grained timing / location of charging and does NOT require any electricity flow outward from the EV battery to the grid.

Tesla is setting up a large network of public charging stations: super-chargers (for long distance travel/re-charge) and destination charging (in partnership with hotels, cafes, parking lots, shopping centres etc) offering 50 kW charging rates. What people do not realize is that Tesla has 5X more destination chargers (1122 in Sept 2015) vs super-chargers (224 in Sept 2015), and a faster growth rate! Since these destination chargers are setup in partnership with hotels etc, they could be augmented w/ PowerWalls (to provide micro-super chargers with DC-DC charging), upgraded via software to modulate charging rates as a function of grid conditions and monetized for such ancilliary services / micro-demand response value creation. If combined with rooftop solar, presumably they can orchestrate such distributed energy resources in partnership with companies like SolarCity.

This enables some unique business models: buy the Tesla Model 3 chassis, but lease the battery pack, EV charger & PowerWall. This will significantly lower purchase price of Tesla Model 3, but give Tesla (in partnership with their financing partner) the ability to aggregate via software-control the decentralized resources (stationary battery pack, EV battery pack) and provide EV-grid services described above. Even if Tesla does not do this initially (i.e. they sell the car to buyers for a couple of years), they can offer a buy-leaseback program for the battery pack with the rights to provide EV-grid services (and perhaps share some of those benefits with the car owner). The financing partner could come in with sophisticated financial engineering (YieldCos, Asset securitization, Green Bonds etc) to drive down the costs of financing. It may also be possible for the electric utility or their unregulated subsidiaries to start offering financing / leasing of the vehicle / battery pack, in return for the ability to use the charging infrastructure to manage their grid smartly (in addition to higher overall demand for electricity).

In other words, the EV will produce both transportation services and produce a supplemental EV-grid cash flow for the car owner. When a (slowly depreciating) asset produces two separate cash flow streams, it becomes a lot more valuable. These economics, in turn, will drive up demand for Tesla Model 3, and in turn drive down battery pack costs (due to manufacturing learning curve effects, estimated at 20% for every doubling of production).

I have not covered autonomous driving impacts in this article. But even if you look at the above Energy-Transportation synergy possibilities, this is far more than what we expect of an ordinary car. Tesla Model 3 and the business models it will validate will stimulate an increasing electrification of energy (even beyond transportation). In parallel, the combination of flexible, software-controlled demand will promote the penetration of renewables worldwide, especially solar and wind (which are on rapid cost reduction learning curve trajectories as well).

While EVs will transform the demand side, new software-driven/cognitive IOT led technologies on the RE supply side are adding flexibility and yield improvement to further reduce levelized cost of energy (LCOE). Note that these IT / software led transformations are additive to the natural learning curve reductions due to manufacturing. These will accelerate the total cost-of-ownership reduction trends, and hasten the cross-over with fossil fuel led energy sources. The modularity, decentralization of these options, and the introduction of (stealth) energy storage via EVs will also drive an Internet-like transformation of the energy networks and supply chains.

And oh, this transition will happen much faster worldwide than most people imagine. Lets hope the Tesla Model 3 unveiling on March 31st, 2016 will be a worthy catalyst for this remarkable future.

Twitter: @shivkuma_k

ps: Post Mar 31st announcement. Tesla announced that the number of superchargers worldwide would be doubled from 3600 odd to over 7000; and destination chargers would quadruple to over 15000 by 2017 end! Currently it is positioned as convenience. With enough coverage on key feeders, it can provide an aggregated service as mentioned above.

pps: On June 21st, 2016, Tesla announced that it would make an offer to acquire SolarCity... Another example that there is really an Energy-Transportation Nexus play afoot here?

ppps: On July 20, 2016, Elon Musk introduced his Master Plan, Part Deux. Here he goes beyond cars to trucks, buses etc. On the car sharing model, it seems to follow a CarNextDoor (an australian startup) type business model + autonomy for car sharing. The points I make about depreciation etc should support this storyline. His discussion of solar + storage does not talk about the possibility of grid services by managing the timing of charging as I discuss above, and seems a little disconnected w/ the autonomous car + sharing story which is focussed more on the automotive / transportation economics.

ppps: This is part of a series of articles: "Distributed / Rooftop Solar in India: A Gentle Introduction: Part 1","Rooftop Solar in India: Part 2 {Shadowing, Soiling, Diesel Offset}", "Rooftop Solar in India: Part 3: Policy Tools... Net Metering etc..." "Solar Economics 101: Introduction to LCOE and Grid Parity" , "Solar will get cheaper than coal power much faster than you think..", "Understanding Recent Solar Tariffs in India", "How Electric Scooters,... can spur adoption of Distributed Solar in India," "Solar + Ola! = Sola! ... The Coming Energy-Transportation Nexus in India", "UDAY: Quietly Disentangling India's Power Distribution Sector", "Understanding Solar Finance in India: Part 1", "Back to the Future: The Coming Internet of Energy Networks...", "Tesla Model 3: More than Yet-Another-Car: Ushering in the Energy-Transportation Nexus", "Understanding Solar Finance in India: Part 2 (Project Finance)", "Ola! e-Rickshaws: the dawn of electric mobility in India", "Understanding Solar Finance in India: Part 3 (Solar Business Models)" . 

pps: List of all my LinkedIn Articles