Ford Seeking Group Of Ranger Owners With Extremely Dangerous Trucks

I suppose that when Thomas Edison invented the light bulb, he may have considered that he was meeting certain obvious needs and addressing growing problems. In addition to the poor lighting provided by candles and kerosene lamps, fires were frequent. Poor lighting represented an encumbrance to productivity while fires represented a threat to safety. I doubt Edison’s zeal for invention was hindered by the obvious constraints of his day. How would his newly invented lights be powered? Would anyone be able to pay for the product after all the effort to manufacture it? In the same way, Henry Ford’s zeal for gasoline-powered automobile manufacturing didn’t seem to be hindered by the state of his world in 1900. How could his automobiles get anywhere on the country’s existing roads? How far could one of his cars travel before running out of fuel and having no way to refill the tank? How would these contrivances interact with the thousands of horses on America’s muddy streets? Modern industrial history is littered with examples of leaps in technology and lifestyle that are made in spite of seemingly giant obstacles. Things seem to happen even if a sober logistical analysis would have pitched the entire drawing board into the trash. Knowing this, one might consider the wisdom of keeping an open mind when listening to the growing drumbeat for replacement of all internal combustion engine vehicles (ICEVs) with autonomous battery-powered electric cars and light-duty trucks (BEVs).

Much improvement in ground transportation is absolutely needed. With little effort, any person commuting in a fairly populated region of the United States today can make a compelling case for change – even disruptive change. Consider these factors:

  • Traffic fatalities and injuries.
  • Property damage and financial loss.
  • Cost of car ownership and operation – maintenance, insurance, fuel.
  • Fuel consumption of cars and trucks.
  • Commuter delay due to congestion.
  • Wasted fuel energy due to congestion.
  • Loss of productivity due to congestion.
  • Internal combustion engine environmental impact.

Each of these factors can be analyzed and crafted into eye-popping and disturbing statistics. The numbers are at your fingertips with Google, but who needs numbers when you have your own daily experience? While remarkable improvements have been made in automobile safety features, any parent who is teaching a teenager to drive will notice with great trepidation that driving conditions are worsening - the ever increasing volume of cars mixed with heavy transport vehicles and the increasing complexity of driving conditions. The safety concern is intertwined with the exasperating and time-consuming congestion problem. Finally, accepted science tells us that our thirst for refined petroleum (the majority of which is used for transportation) has dire implications for energy sustainability, environmental health, and human health. We should all be nicely primed to listen to any ideas for credible solutions to these problems, including General Motors CEO Mary Barra’s “Zero, Zero, Zero” pitch for autonomous BEVs. “ Zero accidents. Zero emissions. Zero congestion.” This is a triple crown of excellent motivations. Mary Barra is hardly alone in her endeavors. The autonomous BEV concept is being developed and promoted by a powerhouse of entrepreneurs, academic elites, and industrial titans. Here’s just a short list: Intel (INTC), Apple (AAPL), Google (GOOGL), Qualcomm (QCOM), Nvidia (NVDA), Uber, Lyft, Carnegie-Mellon University, University of Michigan, GM (GM), Ford (F), Nissan (OTCPK:NSANF), Tesla (TSLA), Mercedes (OTCPK:DDAIF), Hyundai (OTCPK:HYMTF), and Volkswagen (OTCPK:VLKAY). Progress is being made, and the rubber has started to hit the road in places like Pittsburgh, PA, where you can see test vehicles regularly running the streets of the city.

Imagine emissions-free vehicles that operate themselves safely and efficiently to take us anywhere we need to go. Industry and academia seem to be telling us that the time has come - that we have arrived at the nexus of artificial intelligence, robotics, sensor technology, and the greening of America. Does it seem too good to be true? Are we ready to relinquish control of the wheel to an intelligent machine? In the famous words of Winston Churchill: “The pessimist sees difficulty in every opportunity. The optimist sees opportunity in every difficulty.” Yet, there is no rule that prohibits optimists from asking questions and challenging the ideas of others. In that spirit, let’s take the three-pronged autonomous vehicle pitch of zero accidents, zero emissions, and zero congestion for a test drive.

To get started, though, we need to first define what is meant by the terms “self-driving car,” “driverless car,” or “autonomous vehicle.” If you Google on these terms, you’ll realize that there is a range of definitions and descriptions. Included in the range are the following:

  • Semi-autonomous, requiring the presence of a driver with the responsibility to oversee the performance of the vehicle, and the ability to take control.
  • Autonomous, with the ability of a passenger to assume control, if needed.
  • Autonomous, with no ability for a passenger to assume control - but with a connection to a remote human operator who has the responsibility of intervening when a problem arises.
  • Autonomous, with no ability for a passenger to intervene and no mention of remote human operator.
  • Independent, self-standing vehicle, such as what we have today, that does not interface with other vehicles.
  • Interdependent vehicles that sense each other and have some sort of connectivity.
  • Some concepts appear to continue the current practice of customers continuing to buy and own cars.
  • Some concepts appear to consider a practice where car ownership is not needed – implying a service arrangement similar to what exists today with Uber and Lyft, except that there is no human driver.

There is a reason for the range of visions. Proponents of the self-driving concept have taken on an enormous challenge – both technical and societal. An incredible amount of variables are involved, and the safety and satisfaction of millions of people are at stake. This is not the same sort of engineering problem as building and operating a chemical processing facility where the boundaries and conditions are tightly constrained and controlled. A single autonomous vehicle could be produced to perform flawlessly in a well-defined environment, but such a vehicle could not be trusted to handle the variability needed to satisfactorily replace the role that the human-operated vehicle plays in our everyday lives. That is an inordinately tougher problem to solve. Therefore, there are bound to be different approaches to the problem. Some may consider an incremental approach that is somewhat gradual in the removal of human driver control. Others may find a gradual approach to be flawed. Either way, it does seem that the common goal among the competing developers is a fully autonomous vehicle that involves no human driver intervention. Such is implied by the following picture of the GM Chevy Bolt reconfigured with no steering wheel and no pedals. With that understanding, let’s give the “Zero. Zero. Zero.” pitch some healthy scrutiny. After all, the social and economic implications of the autonomous BEV vision are enormous. Because this is a substantial subject, I am including a road map for our journey. Note that all investment insights are consolidated in the final section, so you can scroll ahead, if you wish. But, if you believe as I do that you should never invest in something you do not understand, I invite you to follow my itinerary. Enjoy the ride!

I. Zero Accidents (C.E. Leach)

I.A Automotive Safety Data and Human Causation

I.B The Data Not Collected

I.C New Technology and Complexity

I.D Liability

II. Zero Emissions (C.E. Leach)

II.A Battery Charging Energy Requirement

II.B Data Storage Energy Requirement

II.C How to Achieve Zero Emissions

II.D Additional Pollutants and Technical Concerns

III. Zero Congestion (C.E. Leach)

IV. Near Term Expectations (C.E. Leach)

V. Investment Implications (D.S. Leach)


Part I. Zero Accidents

Let’s start out by putting ourselves in Mary Barra’s shoes. If we are to aim for zero accidents, the first thing we must do is examine the data on car crashes to size up the problem, figure out what causes it, then consider what we can do to fix it. Then, we should establish some criteria to define what constitutes success. Once we propose some fixes, we need to shake down the proposals to see if they truly support the success criteria. Then, we need to thoroughly analyze the proposal to evaluate whether it has any downsides. In other words, do all aspects of the fix lead to a net benefit when deployed? Or, would the implemented fix be more trouble than it is worth – creating unintended consequences, for instance? The GM pitch seems to imply that the heavy lifting has been done for us. The success criterion is “zero accidents,” and the proposed fix is to replace human drivers with machines that drive themselves. Let’s look at the data and consider if we’d come up with the same answer GM has.

I.A Automotive Safety Data and Human Causation

Intuitively, our everyday experience confirms that driving in the United States can be dicey. If we haven’t actually been in an accident, we certainly know someone who has, and we certainly have had the experience where we gasped at what felt like a near miss. The statistics support our collective feelings, yet there is a modicum of good news. Consider two figures published in the National Highway Traffic Safety Administration’s 2015 Motor Vehicle Crash Overview document (published August 2016). The first figure below shows the trend between 1965 and 2015 of crash fatalities. The second figure shows the trend between 1990 and 2015 for injury rate. From the perspective of overall trending, safety statistics have improved – although, we are seeing an uptick in recent years. The continuous green line in each plot shows results normalized to 100 million vehicle miles travelled (VMT). Therefore, the trend line is a rate not distorted by population change. Rate of deaths and injuries per mile travelled follows a downward trend and represents a significant improvement over past decades. The black bars show the actual numbers of fatalities and injuries. These bars demonstrate that despite the increase in U.S. population over the years, the absolute numbers of fatalities and injuries has decreased (see population and summary data in the following table).


The U.S. National Highway Traffic Safety Administration published statistics in February 2015 under the title “Critical Reasons for Crashes Investigated in the National Motor Vehicle Crash Causation Survey.” The survey considered two and a half years (2005 – 2007) of crashes involving light vehicles. An abbreviated break-down of the findings follows.


This data may be the source of the often cited statistic that more than 90% of all car crashes are caused by human error. The 2015 NHTSA Motor Vehicle Crash Overview provides an additional detail that cannot be ignored. There were 35,092 crash-related fatalities in the U.S. in 2015. This includes victims inside the crashed vehicles as well as victims outside the vehicles (e.g., pedestrians, motorcyclists, pedal cyclists). Of those fatalities, 10,265 involved a car driver or motorcycle operator with a blood alcohol concentration equal to or greater than 0.08 g/dL. This demonstrates that roughly 30% of fatal crashes involve alcohol-impaired motor vehicle operators.

A bit more data condensed from the 2015 NHTSA report provides useful perspective.


It is important to note that of the 22,441 passenger vehicle occupants killed, at least 9,874 (44%) were not wearing their seat belt. A last bit of data on comparative risk and fatality statistics is provided below.


A look at the data brings up some points to be considered.

Consider the Trend. The overall trend is definitely in the direction of improved safety. There is no drastic adverse change in the data to motivate a drastic change in practice unless the collective tolerance for driving risk has significantly dropped. Daily experience doesn’t seem to suggest that attitudes about driving risk have drastically changed. Most people seem to be a bit more concerned about climbing into an airplane than a car, despite what the data tells us about risk. People also seem to be a bit more upset about the nuisance and pinch of a flu shot than about getting into a car, despite what the data tells us. But, it is important not to get too glib here. People certainly do care about automobile safety. They like buying safer cars and they welcome improvements of safety statistics – probably as long as they continue to get what they want from their chief source of transportation. That is incredible flexibility and freedom to go where they want, whenever they want, with whomever they want, while listening to the music they want, announcing their taste and social status via the choice of the car they bought, etc. Yet, the third prong of the pitch for autonomous vehicles (zero congestion) suggests a possible challenge to some of those current freedoms and preferences (more on this in Part III.) Nonetheless, as reported by NHTSA, motor vehicle crashes consistently rank within the 10 top causes of death in Americans from birth through age 64. The quest for improved safety must continue. The question is how best to achieve it.

Consider the Data Specifics. The data indicates the driver to be the critical factor in 94% of crashes. It doesn’t take but a skip, hop, and a jump to grasp onto the idea that the solution is to get rid of the human. Enter autonomous vehicles. Judging by the media and industry buzz, autonomous vehicles will absolutely improve safety. This conclusion rests on the single assumption that removal of the human driver necessarily removes the primary culprit for accidents. After all, human (i.e., driver) error is shown to be involved in most accidents. But can this fact be extrapolated into a law or general precept that the only way to significantly improve road safety is to remove the human operator? Importantly, the human factors are quite focused, suggesting that a more direct solution would be to address those specific factors:

  • Alcohol Impairment. As noted, 30% of fatal accidents involve alcohol-induced impairment. Therefore, a successful effort to reduce or eliminate alcohol-impaired drivers could yield significant improvement. Note that the standard U.S. metrics for crash data include alcohol-impaired motorcycle operators who cause accidents, so efforts would need to address the behaviors of car drivers and motorcycle operators. We know of things that can be done today to reduce alcohol-impairment crashes, including the following: increased check-points, increased enforcement as deterrent, public information and education campaigns, breathalyzer interlock devices in cars. The point is that more can and should be done to address this safety problem, independent of the idea that cars should not have human drivers since some of them misbehave and drive drunk. In terms of the standard U.S. metrics for car crashes, autonomous vehicle promoters cannot promise “zero crashes,” at the very least, because the technology will not address the alcohol-impaired motorcycle operator who is included in the standard statistics.

  • Lack of Seat Belt Usage. As noted, in 2015, 44% of vehicle occupant fatalities involved failure to use a seat belt. An uptick in lack of seat belt usage occurred in 2016 for which the NHTSA notes: “Of the 37,461 people killed in motor vehicle crashes in 2016, 48% were not wearing seat belts. In 2016 alone, seat belts saved an estimated 14,668 lives and could have saved an additional 2,456 people if they had been wearing seat belts.” Therefore, a successful targeted effort to ensure passenger use of seat belts holds promise to improve road safety. And, these numbers do not even include the promise of fewer or less serious injuries in the case where less severe collisions do not cause fatalities. More can and should be done without delay. Effective actions could include increased check-points, enforcement as deterrent, public information and education campaigns, seat-belt sensor and interlock devices, and passive restraint systems. The need for passenger restraints cannot be eliminated. Promoters of autonomous vehicles cannot promise zero fatality or injury of vehicle occupants (more on this later). Whatever can be done to encourage or enable passenger restraint applies for human-operated cars and autonomous cars.

  • Inattention and Distraction. As seen in the NHTSA figures, there has been an uptick in fatal crashes and crash fatalities in recent years. The crash causation study shows that the majority (41%) of crashes involve driver inattention and distraction. The NHTSA reports an increase of distraction-related fatalities of 8.8% from just 2014 to 2015. The timing of the uptick coincides with the seemingly ubiquitous presence of the smart phone in our society. From class room to driver’s seat, the pull of the cell phone on our younger generations is an astonishing phenomenon with mostly negative impacts. Nonetheless, distraction and inattention are long-standing problems whether they involve cell phones, fast food, adjusting navigation settings, boredom, or spirited interactions with other passengers

    . As with alcohol and seat belt usage, behavior can be affected through increased public information and education campaigns, laws, and enforcement. Further, there are now “Do not disturb while driving” features on phones. Certainly, if we can invent driverless cars, we can dream up more ways to control our electronics in the name of safety.

  • Age. According to the CDC, “the risk of motor vehicle crashes is higher among 16-19 year olds than among any other age group.” The CDC further states that “. . . per mile driven, teen drivers aged 16 to 19 are nearly three times more likely than drivers aged 20 and older to be in a fatal crash.” Significant factors include alcohol, seat belts, and distraction. As I have already reasoned, seat belt usage will be important for any type of vehicle, autonomous or not. Further, autonomous vehicles would not obviate the need to address the serious societal implications of alcohol and drug abuse. It is a critical need right now to vigorously address substance abuse in our society, particularly for our youth. Obviously, it is not just an issue for transportation since current day tragedies such as hazing, date rape, and opioid overdose warrant urgent action. Similarly, youth addiction to electronic devices impacts many areas of life, not just driving. It is a safety problem for pedestrians and selfie-enthusiasts and a productivity problem for students and employees on the job. Again, driverless cars will not obviate the need to curb this problem in our youth. In fact, when we think about it, driverless cars may unwittingly further enable substance abuse and electronics addictions. Logically, that is what can happen when we work around our problems instead of address them.

Fatality and injury data (as well as property damage data) seem to easily support the simplistic notion of eliminating human drivers. The data does not, however, rule out the value of other strategies to manage risk and improve safety. Other strategies are much simpler and legions less disruptive than autonomous vehicles would be. Further, crash data alone is not a sufficient metric on which to evaluate the advisability or efficacy of autonomous vehicles. The costs and implications of the disruption, as well as the actual performance of the technology, would have to be weighed against the efficacy of other possible strategies to improve safety. Right now, slogans such as “zero accidents” are based on the implicit assumption of perfect machine performance, which has not been demonstrated and is simply unbelievable at this point. Some reasons follow.

I.B The Data Not Collected

I have no doubt that it is within the reach of modern technology to manufacture an autonomous ground transportation vehicle that handily out-performs any human driver for a given set of well-defined conditions. Exciting reports of early testing seem to bear this out. Therefore, it is logical for the hopeful among us to extrapolate test-scale results to success on a massively larger scale. Conceptually, integration and execution of the latest modern technology can be brought together to replace human drivers with machines that would seem to promise us the following:

  • Indefatigable and non-distractible driver attention.
  • Flawless and efficient acceleration, speed, and steering modulation.
  • Superior reaction time to unexpected obstacles and road conditions.
  • Superior execution of evasive and responsive safety maneuvers.
  • Flawless (and silent) navigation performance.
  • Ability to learn from driving experiences and to consistently execute improved driving performance based on new data.

Who wouldn’t expect improved safety performance? And who wouldn’t take a moment to daydream about an effortless commute to work in Atlanta while you finish breakfast, take a power nap, then run through your slides for tomorrow’s presentation – all without any concern about the upcoming detour at your normal exit. And, don’t forget to dream about your college kid who is en route to school after Christmas break without a car to park or a permit to renew this semester. Allow yourself to keep daydreaming and you’ll gain an acute appreciation for the large percentages of your worries that are actually related to your car. Maybe this is why we don’t want to consider the “difficulties of this opportunity,” as Churchill would put it. But we have to. With something this massive and this disruptive, we need to be assured that deployment of the new autonomous vehicle will be safer than what we have now or what we would have if we implemented simpler incremental improvements in safety. How can we be sure that autonomous vehicles will put us on a path to improved passenger safety? From a process standpoint, it seems that the technology developers need to do three things:

  1. Eliminate (or significantly reduce) the incidence of the crashes we have today.
  2. Replicate the human behavior responsible for all the accidents that don’t happen today.
  3. Be sure that they do not introduce new hazards that do not exist today.

The second item is tricky. The statistics tell us right now that an American has a life time risk of 1 out of 114 of dying in a car crash. The flip side is the 113 out of 114 chance that a person will not die in a car crash. What is responsible for the latter statistic and how can we be sure that it is replicated in an autonomous vehicle?

Consider some examples:

  • The act of operating a private motor vehicle consists of more than the negotiation of traffic signals and traffic conditions; it also involves the very decision to make a journey in the first place. Student drivers operate under restrictions for a reason. Parents impose additional restrictions until the driver masters a sense of when and where it is prudent to operate the vehicle. Anticipation of a coming snow storm, experience with fog, knowing about the black ice that forms on the road next to the steam condenser of the local power plant are example factors that affect driver decisions made before anyone even gets into a car. In a world of autonomous vehicles, humans will necessarily lose their driving experience and lose at least some of their perspective about safe road conditions. So it is reasonable to wonder whether autonomous vehicles will make up for that loss and know not to take risky journeys in the first place.
  • If a driver today sees a twister looming on the horizon, the intense human motivation for self-preservation and the protection of loved ones will result in an instant decision to avoid danger. An attempt will be made to leave the path of the twister, and if that is not possible, the driver may well decide to pull the car over and seek personal safety in a roadside ditch. So, it is reasonable to wonder if an autonomous vehicle will have the capability to sense the danger of a twister. What about a rock or mud slide that is just starting? Will an autonomous vehicle see and understand the implications of the trajectory of the twister, rocks, or mud onto the road? Will autonomous vehicles allow passengers the control to make life-preserving judgments and take actions? What if, due to the autonomous nature of ground transportation, there are blind or disabled passengers who are alone in the vehicle? Would such a scenario imply that there still needs to be at least one passenger possessing a defined minimum standard of human sensory competence in each vehicle at all times? That would be a pretty complicated requirement.
  • A driver today can see if a tanker truck has spilled its contents onto the road. Consider the possibility of an oil spill, or any spill of a substance that is chemically hazardous or creates a dangerous loss of road friction. Experienced drivers make decisions to stop, or drive around the spill, or to turn around and leave the scene. It is reasonable to ask if an autonomous vehicle will know the difference between slick oil and rain water on the road.
  • A driver today can see if a train car is derailed up ahead and to the side of the road. He can smell an acrid odor or see a cloud of green gas. In such a circumstance, a driver knows to immediately maneuver away from the threat – sometimes at any cost, such as in the case of a chlorine gas release. The theoretical napping passenger cruising into the cloud of toxic gas at 60 mph hasn’t a chance in this scenario.
  • A storm just blew through an area. Up ahead, a tree leans precariously towards the road and its trajectory would bring down a power line. A competent driver knows to consider the real possibility that the tree may come down at any moment. He will stop and turn around in any way possible. There is no cross street, but an experienced driver, under such an extreme condition, figures out how to carefully use the space around him to maneuver his car (perhaps using an open field, part of a side walk, etc.) to drive away from the hazard. Will an autonomous vehicle drive onto an open field or drive over a sidewalk curb in an extreme condition?
  • What if, just over the crest of a hill, out of sight, there is a washout or sink hole in the road. A person on the side of the road is waving a warning to oncoming vehicles. Will an autonomous vehicle recognize the intent communicated by the waving person? Would it be able to distinguish the difference between a person simply waving a greeting and a person communicating an emergency? The point is that a human can.
  • A driver sees children up ahead playing Frisbee in a yard. From life experience, the driver knows to expect the possibility of the Frisbee to fly into the street. So, the driver knows to slow down to a speed that can stop if a child runs out into the street after the Frisbee. Will an autonomous vehicle know to do that?
  • Drivers are taught not to drive through puddles of unknown depth, lest the car falls into a sink hole or is swept off the road into a waterway. Perhaps an autonomous vehicle will recognize a puddle and stop the car. That amounts to a lot of stopped cars on a rainy day. Perhaps after a few members of the autonomous fleet fall into a sink hole, machine learning will make the sensor logic so sophisticated that it will know how deep the puddle is and make a decision accordingly. That would be wonderful – but also a bit unbelievable.

You can imagine any number of similar examples where a hazardous condition existed and the unquantifiable life experience, instincts, and drive for self-preservation triggered human responses that prevented the accidents that could have happened but did not. Given the real-life variability of driving circumstances, it takes a bit of hubris to claim that there are near-term prospects for an autonomous vehicle that will replicate all the skills of an able human driver. It is the variability of the circumstances that is the issue. It is the ability to draw on a huge range of life experiences, distill them down into an instantaneous recognition of a threat, and the flexibility to take a compensatory action appropriate to the specific environment at hand. Enthusiasts of the autonomous vehicle concept claim that millions of hours of data collection done through on-the-road testing are being used to teach the programs that run the autonomous cars. It is the wonders of Artificial Intelligence (AI), they say. And, with that, there is the implicit assumption of perfection – or, at least, demonstrable superiority over human performance. But, there are those who observe that AI is neither artificial nor intelligent. Rather, it is created by humans. Exit the human fallibility from the driver’s seat and enter the human fallibility into the computer software. Once that happens, the burden of safety will be tremendous for the fully autonomous vehicle that is hauling passengers who have

  • no means to intervene if restricted by vehicle design,
  • no time to intervene because the onset of the hazard is faster than the time to transition control from vehicle to human,
  • no means to intervene because humans have lost or never developed the skills to do so,
  • relinquished control and responsibility for safety so thoroughly that they fully pre-occupy themselves or even sleep, or
  • diminished capacity to control their own safety due to age or disability.

Unless and until we reach the point where coding and hardware can exceed or match human flexibility to respond to the unexpected, it seems that there are only two outcomes. One would be a standstill for autonomous vehicles. The other would be a push forward but with the necessity to somehow simplify or restrict the autonomous vehicle driving environment so that it is more predictable. The latter would imply the very limited use of autonomous vehicles in places of greatest need and/or tightly constrained variability.

I.C New Technology and Complexity

It may be useful to consider that a fully autonomous vehicle is no more an incremental improvement to today’s cars than the light bulb was an incremental improvement to the candle. Candles and light bulbs both provide light, but most everything else about these two things is quite different – including their safety implications. I propose that the same can be said for human-operated cars and fully autonomous cars. Both provide transportation from point A to point B, but much else about these two technologies is quite different – including what can go wrong. Autonomous cars are not a simple logical improvement upon human-operated vehicles. The switch to autonomous vehicles (especially as envisioned in the three-part pitch by GM) isn’t merely a switch in what gets parked in your garage each night. In fact, there is much to suggest that there won’t be a vehicle in your garage. Individual car ownership may become a part of the past, as well (more on this in Part III). Autonomous vehicles will be pieces of equipment in a massive interconnected and coordinated system involving the introduction of technological disciplines that have not been brought together yet in the applications of taking Johnny to piano lessons, taking the family to the beach, racing to the hospital to give birth, hauling goods to the Wal-Mart Super Center, all on the same road, all at the same time, under every possible weather condition and every state of road construction. The idea that taking the fallible human driver out of this picture will necessarily improve safety is not a foregone conclusion and must be rigorously demonstrated. It is too early to make such a conclusion. (The developers have further complicated the safety analysis because they have also introduced the change of the power plant to battery - more on this in Part II.) The truth is that we don’t know yet what can go wrong in the operation of a vast autonomous vehicle transportation system. A vast autonomous vehicle system does not yet exist and we are betting on the come if we assume that the net effect of what could go wrong will be better than the problems we now have.

Much responsibility therefore rests with those who execute the design, manufacture, and operations of the vehicles. While the performance of the human driver has been removed, the importance of the performance of the human software developers has grown disproportionately. Seen in this way, the possibility of human error has not been eliminated at all. Instead, it has shifted to another group of humans – the developers endeavoring to create a technologically complex system on a vast scale. But the consequences of their potential fallibility seem to be hiding behind the excitement over the amazing variety of advanced technology to be used, such as AI and machine learning – as though these aspects of the design might prevent or erase the intrusion of fallibility. The reality is more likely that incorporation of AI and machine learning, for instance, may make it harder and harder for humans to understand and debug the technology they have deployed. This is not a rationale to reject the use of AI, it is simply the sober reality of technological complexity.

In theory, machine learning will replace the need for programmers to anticipate myriads of conditions that an autonomous vehicle could encounter. With the technology of machine learning, developers can theoretically deploy volumes of test vehicles to collect thousands of hours and miles of road experience and have their machines learn from it. It sounds like this should be an enormous relief to the safety engineers. It is implied that the burden of trying to identify and code for any possible driving condition is now shared with the hardware and software that replaces the brain of the human driver. Perhaps all of the learning will be integrated somehow in the cloud and will be deployed for logic execution in the nation’s fleet of driverless vehicles. Or perhaps this sort of talk is over-hyped and does not really reflect how the sausage is being made. Either way, some degree of machine learning is involved, which shouldn’t bother us in principle. Machine learning exists in many forms already. Much of the basic journalism we consume, such as sports and weather reporting, is written by programmed logic built upon computer absorption of many previously written articles. So, the principle exists and has been executed in practical applications. But has it been executed in such a demanding circumstance, with the extent of widely varying conditions, as private ground transportation involving encounters with thousands of separate but interacting machines? Complexity can enable extraordinary achievement, but it can beget unintended and unforeseen consequences. How much confidence do we have that something vastly more complex won’t actually suffer from new faults and unintended consequences?

Here’s the good news. The technology developers are very smart people. Hopefully, they are not so busy trying to replicate and/or surpass the good parts of human intelligence that they are forgetting to ask whether there is a possibility that their novel designs may introduce new and unforeseen hazards. Hopefully, they are looking at each and every new system and component and preparing integrated in-depth failure modes and effects analyses as well as plenty of fault and event trees. Hopefully, there are safety engineers who are working independent of the marketers and project managers, and hopefully these engineers have the authority to add time and cost to the development effort in the name of safety. With a significant degree of AI, however, it would seem that there is a risk that the human developers would become disconnected from the coding that governs vehicle operation and, therefore, feel less responsibility for the safety performance of the vehicle.

In addition to the complexity of the coding, other questions arise regarding the complexity of equipment and systems necessary for safe function of the vehicle. Hopefully, the developers will thoroughly evaluate and address all their failure modes and associated worst case scenarios. Hopefully, regulators are becoming informed and asking questions as early in the development process as possible. If the technologists working on autonomous cars assume that the civil servant regulators and the general public are ill-informed and even disinterested, pressure on them will be removed to consider safety in a truly methodic and rigorous way – especially when it is expensive and time-consuming to do so. The autonomous BEV concept raises many new questions such as the following. There are certainly others that would be surfaced through an objective and systematic process of hazard identification.

  • What are all the failure modes of the autonomous vehicles sensors? What are the consequences of single and multiple failures? What is the probability of failure?
  • Would the autonomous vehicle be vulnerable to mischief and purposeful sabotage, such as the covering of a sensor or the redirecting or distorting of a sensor image with a mirror, etc.? What would the consequences be?
  • How much of the control and navigation of the vehicle will depend on internet and satellite connection, and what sort of recourse will passengers have in the event of failure or hacking of these systems?
  • In the event of control malfunction, will there be a means for passengers to override the control system to bring the vehicle to a safe stop? Can passengers be expected to be able to do this?
  • Has the safety of newly deployed hardware been established? What about the system of spinning lasers atop the vehicle? What are the health implications for people amidst a throng of autonomous vehicles delivering the energy of many laser beams at once? Is eye safety assured?
  • What is the reliability and lifetime of the many new components (on board computers and sensors) used for the first time in the driving environment (e.g., range of temperature, humidity, dust and dirt, snow, hail, lightning, rough road surfaces, physical impacts, etc.)?
  • Since the autonomous vehicle concept is tied to the retirement of the internal combustion engine, questions of automobile battery safety become more significant. These include battery failure modes such as shorting in water. Other concerns include battery fires, difficulty in extinguishing them, hazards to first responders including toxic fumes and electrocution.

I.D Liability

Current versions of autonomous vehicles include the ability of a human driver to take control of the vehicle. Common sense tells us that continuation of this model cannot work. For obvious reasons, human intervention would have dangerous implications in the case where intervention is intended to be rare or intermittent. It makes no sense to expect people to be passive passengers the majority of the time but then expect them to exhibit exemplary skills on demand in emergency conditions. Over time, automation will necessarily diminish human driving skills. The newest vision for the driverless car realizes this, as there is no steering wheel or any other controls. This will obviate any need for learning how to drive a car. In fact, under this model, there may not be any cars left that have controls for humans to use. Having eliminated a human driver, all humans become passengers. Consequently, the elements of human fallibility and responsibility are removed. Or are they?

Perhaps regulators and future passengers should be considering this question sooner rather than later. It behooves us all to think like prospective customers and begin to challenge the thinking of those who propose massive disruption in the name of our safety. They are the ones who stand to profit most from the upcoming massive disruption. Therefore, it stands to reason that they need to embrace the idea that on top of the current level of product liability assumed by auto manufacturers, the autonomous vehicle industry must now assume the responsibility that once belonged to each and every human driver. Consider some examples:

  • If a pedestrian is struck by a driverless car (which has in fact happened), the question of responsibility will arise. The passengers in the fully autonomous vehicle are faultless. The pedestrian in a crosswalk may be at fault. But if there is any doubt, the performance of the driverless car will be indicted. Who then bears the liability? Who is left but the designers and manufacturers?
  • Consider a circumstance where the logic of the vehicle is forced to decide between striking another car, striking a pedestrian, or striking a wall resulting in the deaths of the passengers? The soundness of the choice rests with the design and performance of the machine. Therefore, responsibility will rest with the humans who design, build, and maintain the machine.
  • Just like children cannot appreciate why Mommy will not drive Junior to his play date during a sleet storm, passengers who have not had the direct experience of encountering slick driving surfaces may become unreasonably demanding customers. Who then bears the responsibility when a driverless car spins out on a patch of ice? Will autonomous vehicles have the insight to say “no” to a prospective passenger? They better, because the prospective passenger will no longer have a driver’s license nor the same sense about what constitutes dangerous road conditions. So, who bears the liability when the autonomous vehicle spins out on the black ice?
  • Who bears the responsibility if a driverless car fails to recognize a twister looming on the horizon while the passenger is fully engaged in reading a book or disengaged in a power nap?
  • Who bears the responsibility when a driverless car chooses to swerve to avoid hitting a deer but strikes a jersey barrier or tree resulting in passenger fatalities?

Here’s the rub. No manufacturer can afford to retool and mass produce autonomous vehicles without some understanding of the upper limits of its product liability. Even though today’s manufacturers cope with the consequences of product liability issues, the weight of responsibility for all ground transportation mishaps is unprecedentedly large and would seem to be an impossible burden for a commercial entity, for example GM. From a business standpoint, this means that driverless cars are a non-starter until some surety is established in the matter of responsibility and liability. Putting yourself in the shoes of the developers, you can well imagine their keen interest in spreading risk and liability. Until performance data is available for mass deployment of this technology, insurance companies will be reluctant to fully underwrite policies for the manufacturing and operating entities. In such a chicken and egg scenario, a non-standard approach to liability needs to be formulated. There are two logical targets for spreading any new risks and liabilities – the passengers and the government. Perhaps we can be convinced that driverless cars are inevitable and serve such a greater common good that whatever the risk they entail should be absorbed by all of us – like federal flood insurance.

Perhaps, we will be convinced that we should individually carry passenger protection insurance in exchange for having the privilege of using autonomous vehicles. Remember the passenger flight insurance sold in hallway booths of airport terminals 40 and 50 years ago? In the case of autonomous vehicles, as individuals, we may be convinced that gratitude is in order for the new technology. Imagine never having to teach a teen to drive. Imagine the money and trouble saved because we don’t have to buy a car, register a car, get a driver’s license, be concerned about traffic violations and liabilities, pay for car insurance, maintain a car, etc. Perhaps with the removal of all those expenses, we’d be happy to take out individual policies akin to health insurance. Or perhaps, the government will feel there is a critical national interest in the over-haul of ground transportation due to safety, emissions, and congestion. If so, perhaps lawmakers will indemnify the developers for a period of time in order to launch the technology. In that case, perhaps government would play a role in subsidizing passenger insurance. Reams could be written about the implications of these or any number of other strategies that could be proposed. But, I’ll make one point. It behooves us all to develop an understanding of this liability issue so that we do not find ourselves in a situation where a faulty new technology is foisted on us (think the internet and the scammers and hackers it has spawned) and a disproportionate amount of the liability is borne by the hapless passengers who have little option but to use the technology. In closing the subject of safety, I will make this observation. It should make us pause to think that a highly complex revolutionary technology is being put in place because of the belief that nothing more practical can be done to make more people wear their seat belts, put down their cell phones, curb their alcohol consumption, and obey the speed limits.

Part II. Zero Emissions

“Zero Accidents. Zero Emissions. Zero Congestion.” I guess this communications slogan would not sound so snappy if its second element was replaced with “Zero Emissions From Vehicle.” A bit clunky, for sure. But, that would be more accurate since we all know that the power for a battery operated electric vehicle has to come from somewhere. So, until the power grid of the United States is supplied by carbon-free generation sources, BEVs will contribute to greenhouse gas emissions.


It is seen that roughly 17% of greenhouse gas contributions in the U.S. comes from light-duty vehicles (60% of 28%). As this is the most discussed target for near-term electrification (cars, SUVs, light trucks), we will focus on this classification of vehicles. Considering the degree of disruption that autonomous BEVs promise to cause, one would hope that their implementation would at least be very effective in eliminating that segment of greenhouse gases. Let’s consider what it would take to do that.

II.A Battery Charging Energy Requirement

According to the U.S. Energy Information Administration, light-duty vehicles accounted for 90% of all gasoline consumption in the U.S. in 2016. That was roughly 352, 800,000 gallons per day (more than a gallon per person per day). That gasoline usage translated into a collective 2,849,718,000,000 light-duty vehicle miles in 2016 at an average of 22 mpg for the full array of vehicles in the class.

It would be interesting to know what the battery energy demand would be if this entire class of vehicles could be converted to BEVs. Since the combustion engine average mpg was only 22, it seems reasonable to pick a battery energy demand that matches a fairly large EV. For this exercise, we’ll use the 2018 Tesla Model X P100D SUV (curb weight 5,531 lbs and GVWR of 6,768 lbs). Per U.S. DOE fuel economy data, the combined city/highway MPGe (miles per gallon equivalent) for this vehicle is 85 MPGe. If anything, this is probably a generous stand-in for the theoretical BEVs in this analysis. Since there is currently no BEV counterpart to a vehicle such as the Ford F-350 gasoline truck yet, there aren’t other values to use in a computation for an average. A summary of data, assumptions, and calculations follow.


This energy (1.35 x 109 MW-h) would have had to come from the U.S. power grid. In 2017, the total combined utility-scale and small-scale electricity generation amounted to 4.04 x 109 MW-h. Therefore, the electricity demand for a full equivalent fleet of light-duty vehicle BEVs would have been equal to 33% of the U.S. 2017 total electricity usage. Of course, that is an astonishing quantity. It is straight-forward to verify that there is not sufficient reserve generating capacity in any region of the U.S. throughout all times of the year to accommodate such an increase in demand. Further, the share of electricity generation from fossil fuels in 2017 was 62.7%. And, the fastest growing source of newly installed generating capacity is fossil fuels - in particular, natural gas. Therefore, promise of “zero emissions” can only be achieved with the installation of new non-greenhouse gas emitting electrical generation capacity sufficient to meet the entire demand of the BEVs.

II.B Data Storage Energy Requirement

This is not the end of the story on electrical requirements, however. The quest for “zero accidents” requires the vehicle to perform consistently as good as or better than a really good human driver. The developers are not shy about discussing the incredible and amazing computing power required to achieve this endpoint. In fact, they seem to brag about it. This may lead some to wonder about the infrastructure and energy needed to support the BEV computing power. Most of us tweet and text away without a worry in the world about what enables our vacation selfie to show up on our Facebook page. Like magic, it just happens effortlessly. Except for the exercise of our fingertips, we don’t contemplate that there is energy expended to house and operate bank upon bank of data storage hardware in brick and mortar buildings to make this happen. But there is.

It may not be possible to have a very precise discussion on this topic right now given the early status of autonomous BEV development and given what is shaping up to look like a very commercially competitive race in a potential winner-take-all event. With this understanding in mind, let’s agree to just pluck some low hanging fruit off the internet to get a general idea of the issue. According to a June 2016 article found on, U.S. data centers across the country collectively consumed 70 billion kW-h of electricity in 2014, which was 2% of the U.S. total energy consumption. According to a December 2016 article found in, the CEO of Intel, Brian Krzanich, speaking about autonomous vehicles, stated that “the average driven car will churn out 4,000 GB of data . . . for one hour of driving.” He further said, “One can compare that to an average person’s video, chat and other internet use, which is about 650 MB per day and will escalate to 1.5 GB per day by 2020.” Continuously, while the autonomous vehicle is operating, streams of data from sensors (e.g., cameras, radar, sonar, Lidar) are generated, processed, and stored while streams of additional data are received by the navigation system, which uses extremely detailed dynamic mapping. All of this is done in the quest of replicating and enhancing human driving performance.

Given the numbers cited above, it is conceivable that one car would generate as much data as several thousand people in one day. To drive home a sharp image of this amount of data, a May 2017 Financial Times article points out that 4,000 GB of data (4 terabytes) would equate to 5,600 CDs. Further citing the CEO of Intel, the Financial Times writes, “Anyone who wants to build an autonomous car must have a network of data centers.” This reality is apparently not lost on GM’s Mary Barra. As reported in a January 2018 article in, GM built two data warehouses in southeast Michigan in the last three years at a combined cost of $288M. (By the way, according to the Boston Consulting Group, automotive data monetization will become a $28B business by 2035.) The point is that autonomous vehicles will have a significant off-road impact because of incredible data processing, delivery, and storage needs.

Prior to autonomous vehicles, data storage accounted for 2% of annual U.S. energy consumption. Innovation is improving data storage energy efficiency, but such improvements may not offset energy consumption associated with the anticipated autonomous vehicle data deluge. It seems that many more brick and mortar data centers will have to be built, complete with cooling modules, back-up generators, transformers, and lots of security. At stake will be the safety and mobility of all light-duty vehicle passengers, after all. It seems only logical to assume that the energy consumed by additional data centers will be multiples more than the 70 billion kW-h used for data storage in 2014. On top of that, we have not yet discussed the volumes of additional data and computing power that would have to be brought into the picture to manage traffic and achieve “zero congestion.” More on that in Section III.

The article includes the following figure with caption.>

The source of the figure is a June 2016 report titled “United States Data Center Energy Usage Report.” The report was compiled by authors from the Lawrence Berkeley National Laboratory, Stanford University, Northwestern University, Carnegie-Mellon University, and the U.S. Department of Energy. The figure projects that improvements in data storage efficiency will accommodate the rapidly growing volumes of data generated, at least through 2020. It was not readily apparent in the report whether the study assumptions included anything specifically related to autonomous vehicles, such as expansion of storage capacity in preparation for commercial launching by or beyond 2020. E-mail communications with one of the authors provided confirmation that the report assumptions do not include data needs for autonomous BEVs.

II.C How to Achieve Zero Emissions

We’ve established that retirement of gasoline-powered light-duty vehicles and replacement by a fleet of BEVs will require a significant expansion of the U.S. electrical power supply. While BEVs will achieve “zero emissions” on the road, they will not achieve zero emissions overall unless their charging sources, as well as the power sources for their enormous logistical support, is non-greenhouse gas emitting. Unless the autonomous BEV developers want to also get into the power generation business, they are not in position to claim “zero emissions.” What would it take to truly achieve zero emissions?

Consider the current (2017) composition of U.S. electricity generation sources. The table below was published by the U.S. Energy Information Administration (, March 7, 2018). Nuclear (20%), hydropower (7.5%), and wind (6.3%) are the top three non-greenhouse gas emitting sources. The entire contribution from solar is 1.3%. In addition to being relatively small contributors, wind and solar suffer from the problem of their inability to support a continuous baseload supply of electricity. The sun doesn’t shine 24 hours a day and the wind is variable. Hydropower expansion is not practical or desirable. Currently, nuclear power is the only substantial non-greenhouse gas source that can reliably deliver a considerable baseload of electricity.

Consider what a sufficient amount of increased non-greenhouse gas emitting electricity generation might look like. We’ve estimated a required new annual energy source to charge the BEVs amounting to 1.35 x 109 MW-h. Next, we’ve touched the tip of the iceberg on energy requirements for data processing and storage which could possibly be multiples of the current 70 x 106 MW-h used for all other data storage in 2014. Since the demand for data storage is growing, there is already market pressure to improve energy (and cost) efficiency. So, we can be assured that many very smart people are working on the issue of meeting the data storage and processing demands of the future. The CEO of Intel would not be talking about this challenge if he didn’t think he could be part of a lucrative solution to solve it. This does not mean that we have to give a pass on it for our current analysis, though. In fact, to do so would be counter-productive. As a placeholder, let’s just assume that all data needs for autonomous BEV transportation and passenger logistics doubles the data energy demand that existed in the U.S. in 2014. This may well be a generous assumption in favor of autonomous BEVs. Proponents may argue otherwise. They may also argue that the inherent assumption of the same number of vehicles as today is not valid. After all, the third leg of the autonomous BEV pitch is “zero congestion.” A key component of that part of the pitch would be a highly sophisticated ride sharing program. (More on this in Part III.) But this point (whether realistic or not) would have to be balanced with the proponents’ own assertions that many more people, like the disabled, could enjoy greater mobility. Additionally, there is population increase. Nonetheless, let’s conduct this exercise for two cases. One is where the number of vehicles is the same as it was in the 2016 data. The other is where we will arbitrarily cut the number of vehicles in half. The table shows the need for up to 155 new nuclear power plants in addition to the 99 currently operating plants that we already have (, April 4, 2018).


As can be seen, the provision of “green” energy to support either a full replacement of all light-duty ICEVs or even just half of all light-duty ICEVs would be a monumental task that cannot happen within the next 10 years. Solar at this scale is untenable. Just in case this is not an obvious conclusion, let’s consider a reference point. SAS, the analytical software company based in Cary, North Carolina, installed a 1 MW photovoltaic solar farm (solar farm fact sheet: on 4.8 acres. It generates 1,700 MW-h per year, producing power 16-20% of each day. Setting aside the inability to continuously produce electricity, a quick calculation (simply the total annual energy demand for BEV charging divided by 1,700 MW-h) shows that over 790,000 of these farms would be required to meet the power demand if each ICEV is replaced by a BEV. While there should be some economy of scale, the implication remains that a land area on the order of Connecticut or more would be required.

Like solar, wind cannot provide a predictably steady generation of electricity. However, wind does look like it may have an impressive trajectory of expansion, both in terms of capacity and geography. Data from the EIA and DOE can be used to estimate some inferences.


As can be seen, the net estimated power delivered by wind in 2050 exceeds our Case 2 energy requirement estimate (i.e., where the number of BEVs is half the number of existing ICEVs) and approaches Case 1 (i.e., where the number of BEVs is equal to the number of existing ICEVs). Of course, this ignores the intermittent nature of wind power availability. It also does not consider that the planned expansion for wind does not consider the steep demand for electricity that will be caused by autonomous BEVs – both to replace the energy provided by gasoline and to meet the new energy demands of autonomous vehicle data processing (the latter of which may be difficult to quantify at this time). Independent of the demand created by autonomous BEVs, new power generation has to be planned for a full host of reasons such as the following: to support population growth, to address other areas of economic growth, to replace retired generation facilities, to provide sufficient power reserve for peak demand seasons, to replace fossil fuel generation with renewable energy sources, and to provide diversity and reliability in our electrical power supply. Given the leap in electricity demand which would be caused by large-scale autonomous BEV implementation, additional new generating capacity must be planned. For any vision of near-term large-scale launching of autonomous BEVs, natural gas may be the only option. But, of course, that works against the quest for zero emissions. Also, disproportionately increasing the nation’s dependence on natural gas plants creates economic vulnerability to any emergent issues with natural gas supply and price.

II.D Additional Pollutants and Technical Concerns

The greening of America needs to consider more than CO2 emissions, of course. Batteries have to be manufactured and they have to be disposed of. The materials comprising batteries are hardly non-toxic. Drilling, transporting, and refining oil is hardly a dainty duty either. So, an honest rendering of the two is a worthy task. It is fair and prudent to ask about the health and environmental impacts of the full lifecycle of the battery, which includes obtaining the raw materials, manufacturing the battery, any recycling, and ultimately disposing non-recyclable material. After all, if we are to undergo an enormous economic and lifestyle disruption to deploy BEVs in place ICEVs, we should be sure that there is a net benefit to society. The challenge to do this appears to have been taken on in an Arthur D. Little (ADL) report published in November 2016 and titled “Battery Electric Vehicles vs. Internal Combustion Engine Vehicles.”

The ADL study asserts: “The production of lithium-ion battery packs creates more damaging pollution to human life than ICEs generate over the course of a vehicle’s life time.” Speaking to the net impact of replacing ICEVs with BEVs, the report concludes: “BEV has 3 times greater Human Toxicity Potential and 23% less GWP (global warming potential) impact than ICEV.” The reduced Global Warming Potential of 23% is aligned with a compact passenger vehicle. A value of 19% was derived by ADL for a mid-size passenger BEV, which is more in line with the model that would need to be used for ride-sharing. Further, the ADL study compared a single BEV to a single ICEV and did not consider the relative numbers of each type of vehicle. Therefore, it is inherently assumed that ICEVs and BEVs continue to co-exist and market penetration of the BEVs (currently around 1%) is not great enough to require additional electricity generation capacity. The Global Warming Impact for a single BEV was found to be more severe than a single ICEV for the first three years of life, owing to the implications of battery manufacturing. The net improvement in Global Warming Potential was realized over an assumed 20-year life-span that credits an averaged charging energy mix consistent with the goals set by the U.S. Clean Power Plan. Therefore, the ADL report does not model a full replacement of the ICEV fleet with a BEV fleet and its attendant demand for significant increases in electricity generating capacity. Further, it does not address autonomous vehicles and the associated data processing energy requirements. Accounting for these factors would erase much of the GWP improvement for an autonomous BEV fleet.

With respect to Human Toxicity Potential, the ADL report projects the growth of the BEV penalty from 3X that of the ICEV to 5X by 2025 owing to the trend to larger battery packs in order to increase driving range. The ADL conclusion on BEV Human Toxicity Potential states: “For the American driver, the decision becomes a trade-off between generating small amounts of pollution in one’s local community (or driving region) vs generating comparatively large amounts of pollution in regions where mining and manufacturing occur.”

In addition to potential issues with battery lifecycle pollution, use of batteries to replace the internal combustion engine introduces some unique technical challenges, concerns, and performance issues. Except for the most ardent BEV enthusiasts, people are going to expect BEVs to perform as well as or better than the ICEVs they are accustomed to. Where a performance characteristic may be inherent to the technology (i.e., something that cannot be fixed, changed, or improved upon), people will have to learn to accept a sub-par product, hopefully concluding that the net effect of BEVs is beneficial. Some of those technical concerns and performance issues are briefly discussed below. Most of these are popular topics in on-line articles and owners’ chat sites, so the following comments are very general and brief.

  • Range Variability with Ambient Temperature – We all expect variability in mileage performance from our ICEVs depending on driving conditions and weather. But, the range variability of a BEV with ambient temperature is surprising to people who don’t own one or don’t think in-depth about batteries. A 2014 test conducted by the AAA on three BEVs at city driving conditions ( produced the following results. People purchasing BEVs need to understand that the real-world driving range of the BEV is lower than manufacturer rated range due to ambient conditions.

  • Water and Flooding – A BEV would be more vulnerable to flooding than an ICEV because the battery, of necessity, is at the lowest point in the body of the vehicle. Once water reaches the battery compartment of a BEV, the battery shorts and the vehicle becomes inoperable. This is obviously a safety concern for the situation where the driver is in an advancing flood area and needs to drive out and away from the hazard.

  • Charging Time - A BEV can be charged on a regular 110-v house outlet. With this level of charging, through, the rate is 2 to 5 miles per hour. In other words, a 12 hour charging period would give you a range of 24 to 60 miles. The Tesla Model X P-100, per the Tesla website, could take up to 4 full days to charge on a regular house outlet. Further, under very cold ambient conditions, a house outlet may not be able to charge a BEV at all. A 220-v outlet (dryer outlet) will give a charging rate of 10 to 12 miles per hour. So, a 12 hour charging period would give you a range of 120 to 144 miles. A DC Fast Charger (240-v to 480-v), which one would not likely have at home, charges at 150 to 210 miles per hour on the charger. There are some among us who have been asked to give a lift in our ICEVs to our BEV-owning friends who needed to make a quick trip to the CVS or get their child to school on time while their car was still on the charger.

  • Fire Potential and Fire Behavior – Both BEVs and ICEVs can catch on fire. One does not expect an ICEV, however, to spontaneously catch on fire. If such a thing happens, it is not something that makes the news. As the BEV is a new technology, it draws much interest, and stories of vehicle fires do seem to get elevated levels of press. That said, the videos of EV battery fires are impressive. A gasoline fire is relatively easily extinguished and once extinguished, does not spontaneously reignite. BEV electrical fires are more difficult to extinguish because of the stored energy in the battery. BEV batteries have spontaneously reignited well after the initial fire was extinguished. Personal experience has a way of making a particular impression. Last year, I called 9-1-1 because I saw a battery fire start in an empty hybrid car parked on the street. The owner soon approached the car and opened the rear hatch where most of the smoke seemed to be coming from. The initial assessment was that his golf club shaft caused a short in the battery causing the car to catch fire. I moved my family away from the area pretty quickly because of concerns about what toxic hazards may be in the smoke. This made me think of the dangers to first responders – inhalation hazards, electrocution hazards, burn and property damage hazards from unexpected re-ignition.

  • Availability of Cobalt – The previously cited ADL study raises concerns about battery materials and manufacturing toxicity. Battery components present environmental hazards because of the mining, manufacturing, and recycling of components. There are also potential economic issues because some of the battery materials (e.g., cobalt) are limited resources. R&D for next generation batteries should probably consider the possibility of alternatives.

  • Driving in Remote Areas – The greatest range available in a 2018 BEV is documented to be 337 miles for the Tesla Model S, P100D sedan. The greatest range for a gasoline vehicle is 713 miles for the GMC Yukon. It is 731 miles for the Jaguar XF 20d diesel. And, with a gasoline or diesel vehicle, one has the possibility of bringing extra fuel. Especially in extreme climate or weather conditions, driving a BEV in a remote area is simply not practical.

  • Stranded Vehicle Scenario – When an ICEV quits because it has run out of fuel, the situation can be rescued by a gas can. A stranded ICEV can be restarted and make its way off a highway to the nearest exit and gas station. This is not the case for a BEV that has run out of charge. A stranded BEV cannot currently be recharged by a portable service. It would have to be towed to a charging source. If the journey to the charging source involves up-hill travel, a high friction surface, or hot/cold ambient temperatures, the towing vehicle may well need to be an ICEV.

  • Towing Capacity and Range – Existing BEVs do not compare favorably with ICEV towing performance. A 2016 Edmunds test ( June 9, 2016) of a Tesla Model X towing a tear drop trailer weighing 1250 lbs, which is less than half the rated vehicle towing capacity, resulted in the driver comment: “I’m not sure I ever want to do it again . . . The problems are numerous: towing speed, range, recharge times . . .” A round trip of 1,003 miles took 40.25 hours, of which 17.23 hours were spent in charging stations (11 total stops that averaged 1 hour and 34 minutes each). The teslarati website ( January 3, 2016) comments that the rate of energy consumption while towing can cut range by 60% or more.

  • Braking Distance – Due to the weight of the battery, a BEV has a longer stopping distance than a comparably sized ICEV. Stopping distance, of course, is a critical safety factor. Theoretically, autonomous vehicles will have superior safety performance due to extra-human response time. But, this benefit needs to be large enough to overcome the penalty of longer BEV stopping distance. A thorough assessment of autonomous BEV safety needs to consider this. (Note that Tesla-owner chat sites discuss the practice of filling tires to 40 psi and more to reduce rolling resistance and thus increase range. This of course causes further penalty to stopping distance, which may not be reflected in the data below.)>

  • Particulate Matter Pollution – Because of the weight of the battery in a BEV and because of the BEV’s increased stopping distance, there is logic to the speculation that tire wear and brake wear may increase per passenger mile. Those of us who still hand-wash the car are quite familiar with the very fine, very black residue that quickly accumulates on our vehicles. A thorough assessment of the environmental impacts of BEVs should consider this effect.

With time and more R&D, solutions to some of these issues may be found. In the near-term, however, it is hard to imagine that most Americans would choose to buy a BEV over an ICEV. If funds are no object, the well-heeled among us can enjoy the gee-wizardry of a Tesla, but still have the ICE Cadillac Escalade to haul the out-board motor ski boat to the lake for the weekend or to reliably get anywhere else at any time. For the other 95% of Americans, the story is likely to be a bit different. Owning, operating, and maintaining a BEV is still an expensive, constraining, and potentially problematic proposition. Perhaps the technology developers know this, and perhaps this feeds into their visions for on-demand transportation services that could obviate the need for vehicle ownership altogether. If passengers don’t have to bear the burdens and risks of owning the new technology, they may be willing to ride in it – as long as the ride provider charges a reasonable fare.

It is clear that there are many factors to be considered when evaluating the true environmental impacts of a technology or a practice. Elimination of gasoline from our transportation diet does not by itself equate to environmental health victory. All energy needs associated with implementation of the replacement technology must be accounted for. All the environmental impacts of new materials and components must be accounted for. All implications for the real world performance and utilization of the technology need to be accounted for.

Part III. Zero Congestion

It would be purely pedantic to rehash the data on U.S. traffic congestion. All can agree that congestion is a huge problem with huge social impact. It is intertwined with the safety problem and the environmental problem. It is a problem on our roads. It is a problem in our parking lots and parking spaces. Something needs to be changed, and no one can be faulted for thinking outside the box to try to fix it. The strategy of continuously widening roads and expanding parking lots is not working. So, hats are off to GM, Uber, Google, and all the rest for offering up anything that is new and different. Imagine solving our congestion problem without a constant state of road expansion, which has become such an immense blight upon the landscape. Imagine elimination of stress over inability to find a parking spot, or inability to stay in a parking space for more than 2 hours. So, just what is it that they are offering? How would it work? We hear about bits and pieces of a vision in media articles, TED talks, and NPR. It seems to boil down to something like this.

  • Theoretically, minutes and hours can be mined and amassed from the collective seconds of improved reaction time afforded by autonomous vehicles.
  • Theoretically, the superior driving performance of autonomous vehicles will also eliminate crashes, eliminating tremendous delays.
  • However, with the elimination of licensed drivers, more people will be able to commandeer a vehicle for a trip on the road. This will increase the number of vehicles on the road, which will add to traffic. Also adding to traffic will be normal rates of population growth.
  • To manage traffic and eliminate congestion without sacrificing the demand for passenger miles, ride-sharing will either have to be encouraged or become a condition of light-vehicle transportation.

The primary strategy for zero congestion is probably ride-sharing. It is a miniaturized and fine-tuned form of mass transportation, and an earnest effort to use it to eliminate congestion would amount to serious social engineering. A fully optimized system of ride-sharing would have to be managed by a ride-share service using the internet and data centers to collect and process a tremendous amount of data from the passengers. The ride-share service would have to coordinate the destinations and schedules of passengers in order to group them in a way that tries to fill cars. Therefore, passengers would not be driving cars, planning routes, or even controlling who is in the vehicle. If this is truly the case, there seems little reason for a passenger to own a vehicle. Instead, a passenger could use his smart phone to ask for transportation to a destination in a desired timeframe. It would be up to the ride-share service to collect the passenger and get him to his destination. For congestion control, the ride-share service would use computers and data bases to find other possible passengers to transport along the way. It would feel similar to sharing a taxi with other passengers one may not know. Perhaps there would be an option to ride alone during times of minimal traffic or if the passenger agrees to pay a greater fee. The bottom line is that we would be paying by the drink for our transportation, and we would forego all the hassles and costs of vehicle ownership. Additionally, we would shed the burden of personal liability. But, we’d have to closely rub shoulders with people we may not know.

Setting aside the issue of ride-sharing, there are some enticing advantages to a ride-service. In a perfect world scenario, once you summon a vehicle, it would arrive clean and charged. You’d bring some reading, knitting, or a sleep mask and ride along to your destination. The vehicle’s on-board computer would constantly check for level of charge and continuously assess miles left before needing to head to the nearest charging station for a recharge. Recharging, cleaning, and repairing the vehicle would all be the burden of the ride-share service. As long as ride fares stay at a reasonable rate and people can be convinced that it will cost them less to pay by the drink than to own and operate their own car, the system should theoretically work. There are plenty of hitches, however.

Given the range of technical challenges that face the developers of this new technology, it could take quite a while to match the degree of vehicle quality and performance we are accustomed to in our ICEVs. Much testing and real-life road experience is needed to work out glitches. So, centralized ride-services may be the only realistic way to introduce the technology to the public. The technology would be shaken out while providing transportation to passengers. As long as passengers do not feel that they are bearing the costs to work out the bugs, they may not object to participating in a beta phase of product maturation.

Nonetheless, anyone I talk to (which does not include any NYC dwellers) is horrified by the vision articulated above. It would end life as we know it! Here are example exhortations that I hear in my informal opinion surveys.

  • How would I shop for food? Where would I put my groceries, and how would I know if there would be enough room for them? Would I have strangers riding home with me and waiting at my curb while I unload my groceries? If I didn’t want to get wet from the rain, would I have the vehicle drive me and some strangers into my garage?
  • How would I take my dog to the vet?
  • Might I have to climb into a car with someone’s feisty pit bull?
  • How would I tow my boat to the lake?
  • What if I change my mind about my destination mid-trip while there are three other people in the vehicle?
  • What if I am traveling some distance with a baby? A weak bladder?
  • What if I’m contagious and I need to get to the doctor?
  • Could I get mulch, plants, and plywood at Lowes? Would I have to have everything delivered?
  • What if I don’t have a destination in mind? What if I just want the pleasure to explore a place in any random order?
  • It would be great to eliminate parking, but does that mean that I have to call for a new ride after each of my stops? What if I have a car full of stuff and have three separate stops I need to make?
  • How do I transport my harp?
  • What if I don’t want a particular ride-share passenger to see where I live? What if I feel intimidated or threatened by a particular ride-share passenger?
  • What if I can’t fit my family into one ride-share vehicle?
  • What if all passengers are going to the airport and all their luggage cannot fit into the ride-share vehicle?
  • What if I arrive at my destination and find that it is a wrong address? Will I be discharged from the vehicle to accommodate the programmed pick-up of another passenger?
  • Will I be able to swing by the drive-thru window to get a coffee when I see a shop along the road, or will I have to communicate that request in advance to the ride-share service? What about the bank drive-thru? Do I have to make my transactions under the scrutiny of a car full of strangers? Or, will I have to wait in a line at McDonald’s while one of my fellow passengers gets lunch?
  • When you summon your ride, will you have to specify the size and shape of what you will be bringing so that the ride-share service will know how to make sure that the combined ridership has enough cargo space? Or, will we have to summon the drones to take our cargo separately?
  • What if I get bit by a copperhead in the garage and need to get to the ER (which happened to my neighbor)? What if Dad slices his thumb on the band saw (which happened at our house)? What if I get stung by a bee and start to get an alarming allergic reaction (which happened to my friend)? If no one is around, a ride service is better than nothing. But, nothing is better than immediately jumping into the car with family, co-worker, or friend and quickly getting help.

The panicked questions are endless. Giving up car ownership seems to equate to giving up a lifestyle and, frankly, freedom. How much do we value the freedom afforded by individual car ownership? Safety statistics should give a big clue. Despite the safety risk and hassle of car ownership, we keep buying cars and driving them. Air travel is far safer than car travel, yet there are those who refuse to fly and those who choose not to because they love their ability to take the cat, pack inefficiently, leave when they want, and stop as many times as they want along the way. While air travel is extremely safe and provides tremendous time savings for long distances, it is constraining. When we fly, we agree to give up some freedoms. In this way, it bears some resemblance to the vision for autonomous ride-share cars. Someone else is responsible for our safety. Someone else determines the route. Someone else owns and maintains the vehicle. We don’t have to know the first thing about how to operate the vehicle. We can sleep, read, or knit along the way. We have to ride with strangers. We have to limit what we bring. We pay for each trip.

In fact, all these things are true for any form of mass transportation. This is why I contend that autonomous ride-share cars are simply a fine-tuned form of mass transit. The fine-tuning is what makes it truly different and truly a game-changer, however.

  • Unlike traditional mass transit, the ride service comes to your location to get you and takes you to your exact destination. As such it opens up the possibility of not having to own a vehicle.
  • Ride-sharing confines a small group of people (easily strangers) to a much smaller space than existing mass transit. The feeling of anonymity and safety-in-numbers is removed. Four or six people confined to a car compartment for anything more than the briefest of trips requires some degree of inter-personal interaction.
  • Ride-sharing focuses passengers on the points of departure and arrival of other passengers, which further breaks down a normal social barrier that provides a prized degree of separation and personal privacy.

These are difficult issues that point out the impracticality and social awkwardness of the ride-share model. When circumstances are impractical or awkward, people attempt ways to work around them or avoid them altogether. This can be a source of unintended consequences. Here are some to consider:

  • It is almost always the case that a problem can be solved with money. Those who do not want to contend with the socially awkward, impractical, or troublesome aspects of ride-sharing won’t have to if they have enough money. Those who have the resources will likely be able to pay for private rides or even be able to buy their own vehicle.
  • While it may be intended that disabled people will be liberated by autonomous vehicles, the pressure or cost-incentive to ride-share may have the unintended consequence of discouraging disabled passenger travel. A person who feels vulnerable may not want to be confined in the space of a passenger car with strangers.
  • The inability to come and go at random as one pleases may inadvertently encourage more and more people to consider motorcycles or mopeds for their transportation. From a noise, emissions, and safety perspective, this would not advance the cause of the zero, zero, zero campaign.
  • The limitations of BEVs, autonomous BEVs, and ride-sharing would seem to require many people to use delivery services or somehow manage to also keep an ICEV around or be able to rent one. In addition to the continued need for gasoline, there would be the confounding problem of people losing their driving skills but still needing to intermittently drive. That is a safety problem.

IV. Near Term Expectations

We should all be grateful that some of the sharpest and most ambitious minds in the country are immersed in the desire to eliminate car crashes, address the threat of man-made climate change, and effectively manage ground transportation traffic. No solutions will come for these problems without a great deal of ingenuity, creativity, hard work, and investment. Much will come of the work being done in the areas of BEV development, driverless car technology, and mass transportation conceptualization. We will learn a great deal about what works and what does not. For the reasons presented in this paper, I personally suspect that when the details are clear to the public, there will be considerable resistance to the current vision of the fully autonomous BEV ride-sharing model. I also suspect that there will be resistance to wide-scale replacement of ICEVs with BEVs. In particular:

  • A serious will on the part of the nation to address CO2 emissions would be reflected in voter and legislative action to drastically change the composition of our electricity generation capability. Yet, the populace and government have effectively spoken on this issue. Nuclear power plants are being shut down. Incentives to develop new generation nuclear power are slim to none. There is a growing “not in my backyard” mentality about wind power. Utility customers are pleased to have abundant and reliable electricity at reasonable rates enabled by natural gas. The will of the country is simply not behind the massively disruptive venture of autonomous BEVs in the name of eliminating 17% of our nation’s greenhouse gas emissions. And, expecting the means of electricity generation to change in response to the desire to run BEVs off of green energy sources is the tail wagging the dog.
  • The glimpses we have so far of the possible end product look inferior to what we have right now. GM, Ford, Honda, Toyota, and all the rest have done an amazing job of giving us products we absolutely love. They are high quality, reliable, and last a pretty long time. We practically live in our cars and feel very strongly about them – how they look, how they feel, even how they smell. As much pain as they give us at times in terms of traffic, accidents, licensing, taxes, etc., we love car ownership and the freedom it gives us to come and go as we please, with whom we please. The noble goals of safety and environment are apparently not enough to make us give up these freedoms. If most of us cared so much about safety and environment, we’d already be a nation powered by wind, solar, and nuclear by now. And, we’d be a nation of people with 100% voluntary seat-belt, sobriety, and speed limit compliance.
  • Unless there are more near-term leaps in BEV technology, there is no incentive for most people to willingly give up their ICEV for an inferior and more expensive BEV. People are accustomed to being able to fuel up in minutes and go for long distances before having to refuel. They are used to being able to load up their vehicles and not be too worried about a reduction in range. They are used to being able to have a vehicle that can reliably tow a U-Haul, boat, or trailer full of lawn equipment. And, most people probably don’t want to compromise on the stopping distance they already have. Further, improvements in ICE efficiency continue. The ICE is not a technology that has reached a dead end. In fact, Mazda is on the verge of the commercial launch of the Mazda3 spark-controlled compression ignition engine (Skyactiv-X) which is estimated to increase specific power output by 21% and fuel efficiency by 20% to 30%. That is amazing.
  • In addition to the freedoms of car ownership, so many of us involve our personal vehicles in how we make a living. Our ability to look for stuff, get stuff, carry stuff, and deliver stuff is pivotal to how we work. Additionally, so much of our economy is organized around the existing model of individually owned ground transportation. Whether it is the fast-food drive-thru, the bank drive-thru, the car dealerships, the car repair shops, the car wash, the auto parts store, gas stations, innumerable small businesses like maid services, etc. There would have to be considerable incentive to upset this apple cart - there would have to be some immediate semblance of hope for anyone whose livelihood is disrupted.
  • The no-ownership, rideshare concept is socially tone deaf. Except to the seasoned huge-city dwellers and very short trip takers, this vision is too socially awkward. There is a reason why people in the confined space of an elevator all face forward and stare at the doors. Enough said.

I don’t believe that any of these things will stop the tremendous momentum behind the autonomous BEV industry, however. As long as the commercial contenders and the investors keep funding the technical development, and as long as the political class sees the possibility for national economic benefit, the path to initial product launch should be clear. However, due to the need to firmly demonstrate safety and due to cumbersome issues related to infrastructure installation and transition, the introduction will be gradual. This may well be an awkward and potentially difficult period for the new technology for at least two reasons. The ICEV-centric economy will be in a de facto competition with the intruding autonomous BEV-centric economy. Additionally, there may be unexpected safety and liability issues associated with an increasing mixture of autonomous and non-autonomous vehicles on the road (important topic, but beyond the scope of this paper).

I suspect that the initial launching of the autonomous BEV will have to be through a ride-service arrangement in a limited area. First, the cost of the initial vehicles won’t be competitive with today’s lower and mid-range ICEVs, so individual ownership won’t be feasible for most people. But more importantly, only the manufacturer will know how to maintain the vehicle and diagnose and fix problems. The sophistication of the sensors, on-board computer, and connectivity for navigation will need regular attention – especially in the beginning. That could be achieved if the vehicles are owned by the ride-service company and are housed in centralized car barns to be made ready for dispatch for passenger service. Initial markets would also need to be selected to suit technical needs such as available areas for car barns, and reliable and sufficient power sources for charging. In any place chosen for initial launching, however, most people will likely be skeptical of the safety of driverless vehicles. Passengers participating in the initial launching will need to accept that they will be part of a beta test program. There is no way to arrive at statistical safety data without actually exercising the technology as it is fully intended to operate.

If technical success and customer satisfaction are achieved with ride-service, I expect ride-sharing to be launched as an add-on to ride-service. The initial markets for ride-sharing would also need to be selected on the basis of social acceptance – perhaps in cities that already accept and embrace mass transit. These launchings should determine the real viability of the technology and the business model. Today, the investment and commitment to the autonomous BEV concept is wide-spread and substantial. Thousands of high-paying jobs have been created. Intellectual capital is being expended to advance technology at astonishing speed. Integral opportunities such as passenger data mining are already being shaped and explored (as expressed in emerging slogans such as “Data is the new gasoline.”). If the full vision for autonomous BEVs never comes to fruition, other profitable applications for these technological advancements will come about. So, the question isn’t whether the efforts will continue. The question is what will come of them exactly.

V. Investment Implications

Vehicle Manufacturers

Of the collection of vehicle manufacturers participating in the development of self-driving vehicles, there are leaders beginning to emerge. The summary below provides our ranked recommendations for investors looking for the companies most likely to bring viable self-driving vehicles to the transportation market.

#1. General Motors - GM has successfully demonstrated Level 4 autonomous driving capability at speeds up to 25 mph on the streets of San Francisco and expects to launch a ride-sharing pilot program in 2019 using its Bolt BEV sans steering wheel and brake/go pedals. GM has some obvious advantages over the competition. GM already has the manufacturing capability to build driverless Bolts in a factory north of Detroit, MI. While GM is already well capitalized and has the cash flow needed to fund further development, GM also has the financial backing of Softbank (OTCPK:SFTBY) in this endeavor. GM also owns the Lidar technology developed by Strobe having purchased the company last year. Finally, a position in GM offers the investor a healthy annual dividend of $1.52 while waiting on the development of the GM driverless offering.

#2. Daimler AG - Daimler is the parent company of Mercedes-Benz and developed and offered adaptive cruise control in 1997 in its S-class models. Today, Mercedes offers its Intelligent Drive system in most of its lineup. The system can steer clear of pedestrians and avoid other types of accidents. For its next generation of vehicles, Daimler is working with Robert Bosch Gmbh using a self-driving system developed by Nvidia Corp. Mercedes has tested vans running the Nvidia system at Level 4 and Level 5 on the roads of Boeblingen, near the Mercedes research center in Stuttgart. The vans have been tested on a challenging circuit including morning rush hour traffic. Daimler's target date for offering self-driving technology to the market is 2020. Like GM, Daimler has the capital and cash flow to complete the development of its self-driving technology and offers investors a healthy $4.50 dividend.

#3. Volkswagen AG - Volkswagen is the parent company of Audi and Audi has the most advanced self-driving car available to consumers today in the form of its luxury A8 model. The A8 is offered with optional Traffic Jam Pilot using Lidar. An A8 so equipped allows the driver to be completely hands-free at speeds up to 60 kph (37 mph). Audi is also working with Nvidia in the development of a fully autonomous system with a target date of 2020 for a commercial offering. Volkswagen is the largest vehicle manufacturer in the world. Despite the recent diesel emissions debacle, Volkswagen has a strong capital position and the cash flow needed to fund its self-driving technology products. Investors in Volkswagen receive an annual dividend of $0.92 while Volkswagen completes its self-driving technology development.

Not Rated. Tesla - Tesla has generally taken a completely different approach in their development of self-driving technology. Elon Musk, CEO of Tesla, doesn't have anything good to say about the Lidar technology most other manufacturers are using in their self-driving programs.

They’re going to have a whole bunch of expensive equipment, most of which makes the car expensive, ugly and unnecessary,” Musk >told analysts in February. “And I think they will find themselves at a competitive disadvantage.”

Tesla, rather than focus on the self-driving technology, has prioritized the development of range stretching BEVs that are pleasing to the eye, sporty to the point of ludicrous, and operated primarily through a computer graphic user interface (GUI). To date, Tesla's vehicle offerings have been well-received but mostly by the well-heeled as Tesla's attempt to build a moderately priced Model 3 have, so far, failed. The focus on development and production of BEVs for global markets has, we believe, pushed Tesla's self-driving technology development into the back seat. Tesla does have an Auto Pilot system available but its capabilities are limited compared to the competition. Unfortunately, a few drivers who have stretched the current capabilities of the Auto Pilot system on the road have paid a very steep price. Drivers’ use of Tesla's Auto Pilot system have resulted in 3 fatalities to date as well as several other non-fatal accidents. Independent assessments of Tesla's Auto Pilot technology assesses it at aLevel 2 currently, well behind the competition.

Tesla has yet to produce a profitable quarter and has relied on debt and equity capital for its development activities. This has resulted in Tesla surpassing $10B in long term debt at the beginning of 2018 with additional debt being added at a ferocious rate. Based on Tesla's financial position, its current level of self-driving capability, and its lofty stock valuation, we believe Tesla is a better candidate for a short investment than as a long position.

Key Technology/Components

There are a large number of companies currently involved in the development of key technologies and/or components critical to the successful launch of self-driving vehicles. It would be an exhaustive task to identify all of those companies and provide investment potential summaries. We identify here four companies with leadership positions in the field that should benefit given a successful launch of the technology.

#1. Waymo - Waymo is a wholly owned subsidiary of Alphabet (a.k.a. Google). While Waymo is clearly the leader in development of a self-driving vehicle software/hardware system, it does not manufacture the vehicles. Waymo has been using Pacifica Minivans in its San Francisco testing program achieving Level 4 operation on city streets and at highway speeds. Waymo has demonstrated the lowest failure rate of autonomous driving (when a human needs to take over driving) by a large margin. Waymo also has the lowest accident rate of any company with only 3 collisions over 350,000 miles compared to next best GM with 22 collisions over 132,000 miles. Waymo recently announced the purchase of 20,000 Jaguar I-Pace SUVs and is working on an alliance with Honda Motor Co. (NYSE: HMC) with a focus on delivery services. Waymo is planning to launch a pilot program of self-driving vans that will provide passenger ride hailing service in and around Phoenix, AZ later this year.

#2. Nvidia - Nvidia is working with several self-driving vehicle manufacturers including Daimler (Mercedes), Volkswagen, Tesla, Chery (China), and Volvo in their development of self-driving software and hardware. It is noteworthy that both Daimler and Volkswagen have plans to offer self-driving vehicles based on an Nvidia software system in the next two years. Nvidia has beat analysts' expectations over the last 4 quarters by substantial margins and is well rated by analysts with a solid BUY rating. Nvidia does offer a dividend but it is not big enough to get excited about at $0.60/share and 0.25% yield.

#3. Intel - Intel is having success with its Mobileye technology in several self-driving vehicle platforms. BMW, Nissan, and Volkswagen are using Mobileye Road Experience technology to build and update scalable high-definition maps for use by self-driving vehicles and SAIC Motor will be using Mobileye technology to develop Level 3, 4, and 5 capable vehicles in China. Like Nvidia, Intel has been beating analysts' expectations though not by the same margins. Intel is also well ranked as a BUY by a majority of analysts that follow the company. Intel does provide investors with a notable annual dividend of $1.20 for a yield of 2.4%.

Ancillary Beneficiaries

The successful development and rollout of self-driving BEVs would fundamentally impact a large portion of the US economy, that associated with privately owned, operated, and maintained cars and trucks. There will be changes and new companies created that we cannot even envision yet. One area that will likely grow tremendously with the introduction of self-driving ride-sharing technology is data storage. Today, without the added volume of data from self-driving technology, digital data storage is growing at the rate of 40% per year. That is already a massive growth rate but self-driving ride sharing technology will likely bring a multiple factor increase in data storage growth. As pointed out previously, today the average person uses roughly 650 MB per day which is expected to grow to 1.5GB per day by 2020. The average self-driving vehicle will churn out on the order of 4,000 GB of data every hour of operation from sensor outputs and dynamic mapping and navigation inputs. The increase in data centers to collect, process, and store the data necessary to support self-driving vehicle technology and ride sharing is orders of magnitude greater than our current usage. Clearly, we will have many more digital data storage centers than we have today. With that expected growth, investors should be lining up to own data center REITs. Below are the three we would recommend.

#1. Core Site Realty (COR)- COR has a bit more than 3M square feet under lease. COR has consistently grown both its Funds From Operations (FFO) as well as its dividend, raising its payout by a nickel at the end of June to an annual rate of $4.12. Though COR has been aggressive about raising its dividend, the FFO payout ratio has been very steady at roughly 60%, quite low by REIT standards.

#2. Cyrus One (CONE) - CONE has roughly 5M square feet under lease putting it between DLR and COR in terms of footprint. Because investors have recognized CONE's outperformance, it is the more expensive of the three listed here and its $1.84 annual dividend provides a current yield of only 3.1%. Like COR, CONE's FFO payout ratio is a very healthy 54%. At the current valuation, we've got CONE on our watch list waiting for a little better pricing opportunity before initiating a position.

#3. Digital Realty Trust (DLR) - With its acquisition of DuPont Fabros, DLR is now clearly the 800 lb gorilla in the space today with over 27M square feet under lease. Like the other two listed above, DLR has grown both its FFO and its dividend steadily over the past few years. It might be that DLR will need to spend some time digesting its last acquisition prior to another dividend raise. DLR currently pays out an annual dividend of $4.04 to shareholders.

Disclosure: I/we have no positions in any stocks mentioned, but may initiate a short position in TSLA over the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: We are long COR and DLR.

Editor's Note: This article discusses one or more securities that do not trade on a major U.S. exchange. Please be aware of the risks associated with these stocks.

Source :

Terima Kasih for visit my website
An Investor's Test Drive Of The Autonomous BEV Pitch
Finding the Perfect Truck Shocks
2018's best stories: The Montana Standard news staffers choose their favorites
‘You want a description of hell?’ OxyContin’s 12-hour problem
The Kim Family's tragic journey
COVER- Glenn Barker: Serial killer or convenient scapegoat?
The biggest car news headlines in 2018
Ford CEO seeks to rally dealers as shares sink to new low
Full-Traction 4-Link Coilover Conversion System - Worthy Conversion
Exxon Mobil (XOM) Q3 2017 Results - Earnings Call Transcript