Another contributor to TalkMarkets that I follow, Mish Shedlock, has been posting for years on autonomous vehicles. I have posted several times myself but not as extensively as him. He has a new post today on GM announcing a test fleet of vehicles they will use with their own employees. Mish's piece on this is well done and I recommend it.

Let me add a few other things

Regulations

Mish mentions the regulatory side, but has not in my mind given it the time it deserves.Regulations on autos are still mostly done at the state level (licensing, speed limits, inspections) and while most states have been grappling with regulations for autonomous cars (around 20+ have regulations and over a dozen more are considering them ), the resulting regulations are all over the board - which in my view is a key part of the problem. Last I checked cars tend to traverse state lines, so inconsistent regulations between adjoining states that make the vehicle stop in its tracks at the border are not going to be well received by the autonomous taxi passenger trying to get to the airport. 

Nonetheless, a lot of safety regulations, like airbags, are traditionally done at the federal level by the Department of Transportation, so there is precedent for the Feds taking over - like it or not. And while the National Highway Traffic Safety Administration issued "guidelines" for autonomous cars, those "guidelines" are not getting the job done. We need uniformity in this area. 

Due to these issues, it seems Congress is looking at the problem and perhaps may take action (don't get me started on whether the current Congress can take any action, but they are at least looking at it). Recognizing the need for uniformity, the Senate published bipartisan principles outlining what the regulations might look like, and a House subcommittee approved an autonomous vehicle package making it easier for regulators to act. The Act talks in terms of "preemption," which is government speak for states to take a back seat (pun intended) as the Feds are taking over. The focus is also, like most state regulations, on autonomous vehicle testing requirements more than their actual final rollout and daily use. Still, a step in the right direction.

Liability and Insurance

Beyond regulation, there is the whole liability/insurance issue. Expectations are for vehicular accidents to in time go down significantly. Even for cars with drivers the accidents should go down as they have fewer terrible drivers to deal with and avoid and if they are terrible drivers themselves the autonomous vehicles can react and avoid accidents with them virtually instantly. 

Auto insurers are facing an unprecedented change over the next decade or so, so if you are invested in them make sure they are preparing for the new future. If accidents go down even 50%, and higher is predicted, rates will go down drastically. People, especially in urban areas, will not even have cars and have no need for insurance. Thus, the number of vehicles insured should go down significantly. 

The insurance market will need to shift to product liability insurance for the manufacturers, who will face product liability exposure (assuming no regulatory protection). But you also have all the companies supplying electronics, GPS signals, programing and similar attributes to the autonomous vehicle operations that may be at risk. They will need insurance, so the market will respond and it is worth watching to see what insurers are first to the market in this area. Cyber risk insurance, for example, has over the past few years become a multi-billion dollar market, so insurers would be wise to be looking at this new evolving market as well.  Still I do not see the insurance protecting manufacturers and other related companies will generate anywhere close to the premiums that car insurance generates today. Insurance premiums are based on risk and risk is tied to historical losses, so if the accident, or even theft, rate goes down dramatically, premiums will have to follow or companies will simply self-insure. 

As a side note, cyber risk insurance may help supplement premiums quite a bit in this autonomous area. Autonomous cars, driven by computers and hooked into the internet, are at danger of being hacked. The risk of a terrorist hacking some cars and driving them off cliffs or into each other, might just lead the manufacturers to seek cyber risk insurance protection. At least they would be wise to look into it if it exists.

Auto Manufacturers

It is no secret that several auto manufacturers are fast and furious in their dive into the autonomous vehicle market. GM, which Mish discusses, is relatively new it seems to the game and the car shown in Mish's article is just plain bugly. Others, like Volvo and Ford, have been working on it for a while. And you obviously have the new kids on the block (relatively speaking), like Tesla and Google, trying to capitalize. 

For traditional car companies this is a must in my view. Autonomous vehicles will allow someone not too many years from now to hit a button on an app on their phone and have an autonomous car of their choosing there in minutes to take them where they please. No need for a car, much less two or three, sitting in the garage. No need for car payments. No need for car insurance. No need for car maintenance. No need for washing or polishing or cleaning out the car. NICE!

Car sales in theory will be cut drastically as will manufacturer profits - unless they can replace the sale profits with profits from people paying for cars to pick them up and take them where they need to go. Thus, the rush into the market. Deciding which companies will succeed and which will fail is key to investment results.