While some existing car manufacturers are going the evolutionary route with efforts to increase and integrate the abilities of their connected cars in a way to earn additional revenue and as a way to differentiate their cars, there are others who want to speed things up, aiming directly for full autonomous cars.
While there are many existing car manufacturing companies who do this as well, here we concentrate on the two most important newcomers, Waymo, part of Google (NASDAQ:GOOG) (NASDAQ:GOOGL) and Tesla (NASDAQ:TSLA).
They are probably ahead of the competition in the road towards the ultimate prize, complete autonomous vehicles, at least in one important metric, the number of actual test miles. For instance, Waymo's Firefly has racked up millions of autonomous miles, like Tesla's AV (from Bloomberg):
has collected more than 1.3 billion miles of data from Autopilot-equipped vehicles operating under diverse road and weather conditions around the world. And in the frantic race to roll out the first fully functional autonomous vehicle, that kind of mass, real-world intelligence can be invaluable. In that way, for now, the electric-car maker has a leg up on competitors including Google, General Motors Co. and Uber Technologies Inc.
Of course, developing and training algorithms through machine learning requires lots of data (and cloud facilities to execute, but these can be hired) so the newcomers are likely to have a considerable advantage here versus the car manufacturers embarking on a more evolutionary path.
Indeed, this advantage allows Google to offer these capabilities to the car manufacturers, from Business Insider:
Waymo is retiring its fleet of Fireflies to focus on putting its autonomous driving software in vehicles mass produced by the big car makers, company officials said Monday in a post on Medium.
By focusing on mass-produced vehicles like the Pacifica minivan, we'll be able to bring fully self-driving technology to more people, more quickly. The Pacifica minivans are equipped with our latest generation of custom-built radar, LiDAR and vision systems and an all-new AI compute platform, so they can see even further and sharper. They can also reach full speed (where the Firefly is limited to 25mph), and the interior is equipped with creature comforts that passengers expect in their vehicles today - which makes our initial fleet of 600 self-driving minivans a perfect fit for our early rider program.
These are part of a larger test fleet called Early Rider with the goal of learning how different drivers (or, perhaps we should say people, as they won't be doing much driving) experience these cars. It is inviting the public to partake in the test in Phoenix Arizona with an automated ride-hailing service consisting of the significant amount of an additional 500 of these minivans.
This isn't the first of these kind of tests (Delphi (NYSE:DLPH) and nuTonomy both have pilot programs in Singapore with six cars each. Uber (UBER) started carrying some passengers in Pittsburgh in the fall of 2016 and launched a short-lived San Francisco test in December), but it is by far the largest. The cars do have a safety driver on board, at least for the foreseeable future.
To our knowledge, only GM (NYSE:GM) and Lyft (LYFT), and Volvo (OTCPK:VOLAF) have comparable (albeit, in the case of Volvo, more restricted) test programs. Volvo's is called Drive Me, from the World Economic Forum:
Volvo Car Corporation announced its Drive Me project as it rolls out a number of features in the driver-assist category in early 2015. 38 In partnership with the Swedish government, 100 production-ready autonomous cars will be in the garages of customers by the end of 2016. The cars will be allowed to drive on 30 miles of roads in the city of Gothenburg in what would be the first public test of a self-driving vehicle.
Volvo has also partnered with Uber in Pittsburg and San Francisco providing ride-sharing test programs, but there are some problems here, from Bloomberg:
Each SUV is staffed with two employees, one ready to grab the wheel and the other on the lookout for pedestrians. (Uber made the move without approval from the California Department of Motor Vehicles, and state prosecutors have threatened to seek a court order to force the company to stop. An Uber executive said it's acting "just like Tesla.")
Unlike Tesla, Waymo doesn't intend to produce its own cars. The idea is simply to perfect its autonomous drivers stack and sell this to car manufacturers. Apart from partnering with Fiat Chrysler, it's also in talks with Honda (NYSE:HMC) and has an agreement with Lyft.
One could argue that Tesla is like Apple (NASDAQ:AAPL), going for the integrated approach while Waymo is like Android, wanting to get its AV stack in as many cars as possible.
In October last year, Tesla has introduced new hardware (Autopilot 2.0), from electrek (our emphasis):
Tesla's new Autopilot hardware suite consists of 8 cameras, 1 radar, ultrasonic sensors and a new supercomputer to support its 'Tesla Vision' end-to-end image processing software and neural net, which is the real star of the show here. What was really "unexpected by most" here is the fact that Tesla ditched the original Autopilot 2.0 suite that would have enabled level 3/4 autonomy and instead, it jumped directly to a suite that can eventually support level 5 full autonomy.
We're not there yet, but the new hardware includes 360 degree cameras (with inbuilt heating, so snow and ice won't be a problem) necessary for full autonomy. There is also an upgrade in Tesla Vision; it now runs on Nvidia's (NASDAQ:NVDA) Drive PX2 platform (a 40x performance increase over the previous platform), but the software is Tesla's own.
If the hardware is capable of at least level four autonomy, what's stopping it? Well, for starters, stuff like regulation and infrastructure, which aren't really under Tesla's control.
The other thing is that the algorithms have to improve to include basically all possible road scenarios and appropriate responses, which requires a steady feed of real life data.
And, this is just where Tesla's big advantage is situated. Tesla has a fleet of cars on the road collecting data even if the Autopilot system isn't switched on, operating in shadow mode providing stacks of real world data enabling the company to improve its algorithms which can then be live-updated over the air to its fleet.
It's basically the biggest experiment going on by far, so it's not unreasonable to think that Tesla will be the first to reach level 5 autonomy. Tesla itself and some analysts are already rejoicing with the prospects, from Electrec (our emphasis):
Before introducing the second generation Autopilot hardware, Tesla CEO Elon Musk said that once the first truly self-driving car is available, all other vehicles without the technology will have a "negative value". Echoing the idea, Morgan Stanley analyst Adam Jonas said this week that they started warning their clients that if Tesla is successful in enabling fully self-driving capability on its current vehicles equipped with the second generation Autopilot hardware, it could render all other cars obsolete.
We have to add that not everybody is convinced Tesla has an advantage. Take for instance, Chunka Mui, from Forbes:
Tesla fans could rightly point to the more than two hundred million miles that Tesla owners have logged under Tesla's Autopilot feature. Those miles are not considered here. (Autopilot is not defined as autonomous under California law, so Tesla is not required to report disengagements to the California DMV.) But, no doubt, all those miles means that Tesla's Autopilot software is probably very well trained for highway driving. What do those highway miles tell us about Tesla's ability to handle city streets, which are more complex for driverless cars? Not much, but the 550 miles that Tesla did spend on public road autonomous testing speaks volumes about its dearth of experiential learning on city streets.
Another critic, Edward Niedermeyer even argues that Tesla has been misleading. Most of the autonomous mile driven by Tesla in California in 2016 were for shooting a promotion video, and the disengagement reports Tesla has to submit to the Californian Department of Motor Vehicles showed numerous human interventions.
We're not convinced this is very serious though, as these critics rely on data that is limited in two respects. It's limited to California and it doesn't consider data delivered by Tesla's existing fleet from Autopilot, only that of separate testing. There is another tidbit of data that is interesting though, from Forbes:
Waymo also leads the others in terms of fewer "disengagements," which refers to when human test drivers have to retake control from the driverless software. Waymo's test drivers had to disengage 124 times, or about once very 5,000 miles. Other companies were all over the map in terms of their disengagements. BMW had one disengagement during 638 total miles of testing. Tesla had 182 disengagements in 550 miles.
This suggests, but no more than that, that Waymo isn't far behind, in fact, it might even be ahead, despite having clocked up much less in data generation.
We do wonder how much Tesla will actually gain from 'winning the AV race,' as the bulls have it. Tesla will likely run into bottlenecks varying from certification to legislation to necessary network and infrastructure upgrades to consumer acceptance (probably the least of Tesla's problems) to infrastructure needs.
These roadblocks to commercial deployment could very well allow others like Waymo to catch up, and then there is another disadvantage which we already hinted upon above. Tesla isn't likely to sell its AV tech to others (at least we have seen no indications of this, and there is some confirmation of that in a Tesla forum).
Now, we realize this might change in the future. Keeping the relevant software in-house is only rational if it confers Tesla with a significant advantage over its competitors. The longer the wait for all the hurdles for level 5 autonomy to materialize on the streets, the more time its competitors have to catch up and the less sense this makes.
One way they gain is winning the evolutionary race, that is, they simply offer the most sophisticated ADAS system (short of full level 5 AV), a way to differentiate their cars and earn additional revenue. Here Cho has a point:
Tesla will roll out feature additions that can be validated with its pre-existing fleet and test-vehicle fleet. Also, Tesla's semi-autonomous functionality is more practical for testing/validating feature add-ons than tests of fully autonomous cars. I still believe level 4 autonomy requires network enhancements, hence the path of development/implementation via over the air updates seems most compatible with the current regulatory environment.
So while Tesla is already prepared to offer full level 5 autonomy, for various reasons, it can't do so and is unlikely to be able to do so for quite some time. But in the meantime, it also offers the most advanced ADAS, offering it a way to differentiate and earn additional revenue.
One could wonder at least a bit how much this matters as Tesla already has plenty of ways it differentiates from most of the competition.
The case for Tesla being ahead in the AV revolution seems a bit stronger to us than that for Waymo. The latter's case depends on a more limited data set, primarily its lower rate of disengagements in California.
We're also not convinced how significant a lead at present is, given the numerous constraints that still exist and will likely to pertain for a considerable amount of time to introducing level 5 autonomy on US roads (or elsewhere) as this wait enables laggards to catch up through machine learning.
Both companies enjoy vast valuations. Tesla's seems entirely based on this, whilst Morgan Stanley argues that Waymo is worth $70B. Both could also be caught up by others, like Apple, Uber, and a host of traditional car manufacturers that are engaged in their own testing.
After all, the hardware can more or less be bought off the shelf, and the software can be improved through machine learning that requires raking up lots of miles. Tesla is ahead, probably closely followed by Waymo, but this race still seems open to us.
What could alter the picture is that a clear leader emerges which is also willing to license its solution to others, some of which would then give up their own expensive development efforts. This could lead to increasing returns and a clear winner. At present, Waymo seems much more open to taking that route than Tesla.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within 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.