Connect Planet Earth (Daylight)


The auto industry is one of the largest in the world with global revenues in excess of US $1 trillion. It influences everything from global supply chains and military expenditure to safeguard them, to ways cities are laid out, how and where people live and work,  climate change and human cultures. Hundreds of millions if not billions of people depend on it for their livelihoods directly or indirectly.

In a fascinating interview, Ben Evans, a partner at Andreessen Horowitz, a venture capital firm in Silicon Valley, discusses how the assault of the tech industry on the car will change things -so profoundly that economies, societies and the way people live will be quite fundamentally transformed over the next 20-30 years. It turns out the key is servitization of mobility.


  • The auto industry is one of the few industries bigger than tech (which is driven primarily by smartphones and software) and therefore a prime candidate for tech companies to turn their attention and capital to to sustain growth rates and valuations. In fact Google, Uber and Tesla are already in the car business and Apple is rumored to be planning to join them. So it seems tech and cars are “colliding” in a way that tech and mobile phones collided about a decade ago.
  • Four developments coming together will be shaping this collision (though each on its own also brings significant change): i) the electric car; ii) on-demand mobility services; iii) autonomous (self-driving) technology and iv) car management software moving away from proprietary systems developed by car manufacturers to standard systems developed by software companies.
  • The point of the electric car from an economic (and tech) standpoint is not (only) the CO2/pollution reduction aspect, but the complexity reduction by an order of magnitude (electric motor v combustion engine). The reduction in the number of parts and components not only radically reduces maintenance requirements (how will car dealers make money? -they probably won’t and electric cars will be accessed more and more online -a disintermediation phenomenon), it radically changes supply chains, reduces capital intensity of the business, changes required capital structures and therefore total vehicle economics and the industry value chain, – but it also simplifies software requirements! Hence it makes market entry for new players easier. The current obstacle or bottleneck is the battery, both from a cost and performance (range) standpoint, but charging infrastructure is growing fast and battery costs and performance are actually improving at a faster pace than expected. Granted the plug-in vehicle (PEV, includes both pure electric and hybrid plug-in cars, but not hybrid cars that cannot be plugged in such as the Toyota Prius) market is still very small (up to 350,000 PEVs are expected to be sold worldwide this year, around 1% market share), nevertheless growth rates in some countries (China, UK) are surprisingly high in spite of the oil price plunge which affects vehicle economics. If battery performance continues to improve the combustion engine cannot compete. For example in 2013 the International Energy Agency (IEA) estimated that cost parity would be reached in 2020 with battery costs reaching US $300 per kilowatt hour of capacity.  But according to new research published in Nature Climate Change (based on a review of 85 cost estimates in peer-reviewed research, agency estimates, consultancy and industry reports, news reports covering the views of industry representatives and experts and finally estimates from leading manufacturers) the cost of batteries in 2014 for market leading companies was already lower than what was expected for 2020. The paper estimates prices will fall further to around $230 per kilowatt hour in 2017-18, “on a par with the most optimistic future estimate among analysts”. The crossover point where electric cars become cheaper than conventional vehicles depends on electricity costs, vehicle taxes and/or subsidies, prices at the pump and any additional potential regulatory costs/taxes on carbon. In the US, with current low oil prices, battery packs would need to fall below $250 per kilowatt hour for electric cars to become competitive (this is higher in Europe). Behavioral barriers to electric vehicle uptake present additional hurdles to widespread adoption. At current oil prices and regulatory environment the paper estimates that costs would need to reach $150 per kilowatt hour for electric vehicles to move beyond niche applications and begin to penetrate the market more widely, leading to a potential paradigm shift in vehicle technology. While battery technology “breakthroughs” remain quite uncertain, it is however certainly feasible that economies of scale such as those expected at the Tesla Gigafactory in Nevada (JV with Panasonic) will push costs towards US $200/Kwh by 2020. If renewable energy (wind, solar) costs continue to decline as expected, incentives for EVs are sustained (due to climate change and pollution -particularly in China) and EV ecosystems start to appear the take-off point might be closer to $200/ kWh than $150. And this is not counting potential vehicle income from vehicle-2-grid (V2G) applications, where EVs help electrical network operators balance the grid by providing power back to the grid in times of high demand or network stresses.
  • On-demand mobility services -in any manifestation, whether as conventional car-sharing (short term convenient rentals as provided by Zipcar, Car2Go by Mercedes-Benz, Autolib and many others in virtually every major city) or Uber-type applications, are a fundamental part of the technology driven sharing economy and are already having a profound effect on car ownership structures and demand patterns and starting to impact vehicle design and marketing. This of course fits in particularly well with…
  • Self-driving cars because what does the car do once it drops its passenger off at work or a restaurant for dinner? Does it park itself somewhere? (probably in a parking lot outside the city center where it is cheaper) or does it drive other people around for money? In fact what does self-driving mean for the lay-out of cities (incl. parking lots, gas stations, strip malls and the whole structure and configuration of retail and commercial real estate which is build around parking as well as driving requirements).  How will this contribute to increasing urbanization and the associated productivity implications (urbanization and productivity are very strongly correlated). Once cars are self-driving (electric cars are much better suited for self driving as, among other things, it is possible to recharge them automatically rather than manually, which is difficult with combustion engines) there is an explosion in the supply of vehicles for on-demand services. As there is no driver the cost of the service is much lower (insurance costs will also plummet as self-driving cars will have less accidents and reduced maintenance requirements will minimize those costs) driving higher demand. Where will the equilibrium be (balance between demand for mobility services v demand for owned cars) and who will actually own the cars and provide the services? Will it be independent fleet owners (the Zipcars of tomorrow), car manufacturers turned service providers or aggregators of privately owned vehicles such as Uber or Lyft (it should be noted that Uber, a 5-year old company, has now a market cap exceeding US $50 billion, which is higher than GM and Fiat Chrysler and slightly lower than Ford). While there are no answers to this as yet, if the cars end up being owned by large fleet managers providing services -how will the cars be bought and, as importantly, what kind of cars will they be? Anyone who has seen the Avis or Hertz (as well as many car-sharing fleets) line-up will understand (what does than mean for BMW, Mercedes Benz, Lexus or for that matter Tesla and Apple) -though of course once a driver  and manual controls are not required the design and layout of a car will fundamentally change. And the secondary effects might be just as important. If cars are autonomous and accident rates are low perhaps people would cycle more, even in dangerous cities (minimal chance of getting hit). Traffic lights might not be required, but cars could swarm through crossings very close to each other in a continuous flow with no jams (or accidents). Streets would be noticeably wider with no parked cars at the sides so traffic flow would be eased further and shorter journey durations would have an impact on city layouts and real estate markets over time. And of course there is the issue of what people would do during car rides when they are not driving. If they spend their time on a smartphone or in virtual meetings what does it mean for bandwidth requirements? If they stop listening to radio (radio is listened to mostly when driving) what does it mean for radio programming?  Another big question is what happens with public transport and where it leaves people with low incomes. Because if on demand services push prices lower so that many people can afford an autonomous car service, demand for public transport would reduce significantly meaning much higher subsidy requirements from governments to move people who cannot afford the service. On the other hand prices might drop so low that urban public transport becomes completely obsolete. Naturally the job impact of autonomous vehicles, EVs and on-demand services will be enormous.
  • Who and how will develop the logistics algorithms and systems (what in fact is the kind of AI required) to move around and position the huge numbers of vehicles that a fleet manager would require in order to satisfy demand at the expected quality levels and cost? Routing and mapping (both essentially in real time) competence is transformed into a strategic asset instead of an “accessory” (which places in a different light the reason Audi, BMW and Daimler recently paid US $3.1 billion to acquire Nokia’s Here Maps), a first step in the transformation of car manufacturers into tech providers.

The podcast with a lot more insight can be found here.

The time frame for this development to play out, with the exception of real estate restructuring and city reconfiguration, should not be more than 20 years -which in strategic terms is not all that long. The podcast nevertheless leaves open the question of who will be the winners and losers and who specifically will make money out of this. At the moment service providers such as Uber are enjoying phenomenal valuations, however if the market grows as expected competition will intensify. If prices of mobility-on-demand services drop, cost leadership will be crucial to sustain margins and this requires substantial scale. Given economic network effects market leaders have a strong advantage. Market leadership however will in turn depend on technological ability to develop the algorithms and systems to route and position cars in huge numbers (where they need to be at the time they need to be there and find the optimal route to get from A to B) which sounds more like the competence of a large telecoms service provider (the analogy is not perfect). In any case if we assume that large service providers are owning these mobility service markets, and autonomous cars are essentially computers (robots) on wheels, it is a safe bet to imagine that some form of ecosystem will develop around the product-service system to provide the usual contextual services for the car, the required infrastructure and the passengers.

The valuation of Uber clearly shows that Mobility-on-Demand as a Service (MaaS) is a killer App. Coupled with self-driving cars, particularly electric ones, makes it probably the mother of all Apps of the first half of the 21st century (unless someone develops strong AI that is). And what will be the killer Apps within that ecosystem? Smart businesses are probably already designing them somewhere -even if they don’t know it.


Titos Anastassacos is Managing Partner at Si2 Partners, a consultancy helping clients leverage services to win in industrial markets


If you are a Service professional (manager, practitioner, consultant or academic) in an industrial setting join our group Service in Industry on Linkedin