This is an interesting purchase for Amazon on a number of levels. Obviously they have a huge delivery fleet that could benefit from self driving vehicles. And a giant company like Amazon must love an opportunity with an enormous addressable market such as this one.
But the timing is interesting too IMHO. Amazon played the long game a little bit here and waited until self-driving technology was in the "trough of disillusionment" [1] to pounce on a promising startup. It saved itself billions of dollars in the process, let some of the pretenders fade away, and gave itself better odds of success.
Self-driving tech is still a slog though, and success is not guaranteed by any stretch. But I give them kudos for this.
Considering how far behind they are from competitors like Google, I wouldn't exactly call it a winning strategy. And they are going to spend billions on the acquisition anyways, regardless of how good or bad the technology is. To me this move seems more like desperation than brilliance on Amazon's part.
Yeah... except this is more complex than Gartner's hype cycle captures. This isn't k8s or openstack.
There are real technical challenges to making this work, not that it doesn't magically solve all things - it can't solve anything yet. The long tail still appears quite long.
Sadly Detroit's Big 3 announced that they believe electric cars are the next big thing so they're cutting back on research for autonomous cars. They're blaming it all on the pandemic. I said publicly that Tesla and Comma.ai would eat their lunch. Now I can add Amazon to the list!
I recall Zoox dumped their CEO founder quite suddenly, this announcement reminded to look up the story. Food for thought re founders and controlling stakes in startups.
Employees will make peanuts to nothing if Zoox gets sold for < 3.2 billion. This will make many folks who work at Zoox very unhappy esp with the amount of stress that engrs are taking every day. I hope engrs make some money here otherwise bye bye loyalty.
If you want certainty and no chance of winning big, join a Fortune 500 company. If you want uncertainty and a small chance of winning big, join a startup.
For the engineers, they wouldn't likely win big on a sale below a certain value, but that doesn't make it a total loss. They'll have a job at a FAANG with likely a higher total comp than before. Replace stock option lottery tickets with regular RSU dumps at a decent paycheck.
These risks are the same as for any tech startup company. Many people have made a similar gamble with equity. And in the meantime, salaries for AV software developers, machine learning engineers are not exactly below market.
Wow, I never thought about it before but partial self-driving assistance makes a lot of sense for deliveries.
I live in a pretty quite neighborhood in a suburb of Chicago and I see Amazon Prime vans drive by all the time. The driver parks, hops out, runs around to grab the package, drops it off, drives on, and repeat.
This is a great scenario where you don't need total autonomy all the time: the delivery person could drive the highways and complicated parts, and then when there's a cluster of deliveries where the speed limit is 25mph and the roads are clear and regular, the van could go from house to house, with the delivery person hanging out in back or something.
It's not full self-driving, but it seems like it could make deliveries a little more efficient.
I was under the impression that freeways were the easiest target for self driving. The lack of pedestrians, bicycles and unprotected left turns makes it much more predictable. It's hard for humans because the speed is much faster than human reaction time evolved for, but that's not as much an issue for self driving. Tesla for a very long time only supportted self-driving on freeways.
Neighborhoods are the complicated parts: they're filled with pedestrians, cyclists, kids, animals, and uncontrolled intersections.
AFAICT, a self driving car has to pretty much assume that every pedestrian might suddenly dart into traffic, it can't really tell the difference between a normal pedestrian and a drunk heads down in their phone.
But keeping speeds below 20mph opens up a lot of opportunities simply because stops are basically instantaneous at that speed and because car-pedestrian collisions under 20mph are almost never fatal.
Can someone explain why non car companies are buying self driving car tech? Do they think they will end up with some sort of kit they can just attach to existing cars? Are they trying to pivot into being car manufacturing companies? Or is this just a play to develop the AI then sell that to someone else?
Amazon is now a major shipping and delivery company, in addition to everything else (as of last year, they were apparently delivering half of their packages themselves [0]). My guess is they would like to automate that.
My take about Amazon interested in this is that in a giant list of AWS offerings, there will be a sensor suite processor service called SDS (Self Driving Suite) which will be installed physically inside the car as a server box. Amazon AWS has some experience in building hardware, AWS Snowball for e.g. Ofcourse it will communicate over 5G for non-immediate processing and will run through 3 intermediate AWS services before running the actual processing job on EC2 instance. One of them being SFS or Sensor Fusion Service, which combines geospatial data (from AGS + IPS) with up-to-date road data from other SDS equipped vehicles(1). The billing will be a bit complex so it will require ABS to optimize the billing based on location, mileage and other AI-driven metrics.
Amazon previously purchased Kiva for their warehouse robotics systems [1]. It's safe to assume that Amazon wants to extend that out to the delivery side as well.
>Can someone explain why non car companies are buying self driving car tech?
Because warehouses and the freight terminal (truck terminal in the consumer goods case) that necessarily surrounds them have a need for things that can roll around autonomously while not colliding with things.
There are existing solutions for this but they're far behind the state of the art for self driving vehicles that aim to eventually be used on public roads and they have several weak points that make them less competitive in very dynamic warehouses (variable contents and throughput, as opposed to a warehouse that feeds a manufacturing operation which will have more fixed contents and less variable output) that mostly handle lightweight goods (which is part of why amazon uses so much human labor in its warehouses). I wouldn't be surprised if Amazon was trying to scoop up Zoox on the cheap for their own internal use. They already bought a warehouse automation company. The probably have a list of things they want to automate and figure that buying a company that already has competence in the space (even if they're not a front runner) is the best way to meet their needs.
Of course the delivery network serving the area surrounding the warehouse facility can also make use of self driving tech but self driving on public roads is a much harder and farther off problem and I don't think solving it is their primary goal for this purchase.
In Amazon's case they are a major investor in Rivian, an electric truck company that they are reported to be buying 100,000 electric delivery vans from over the next decade.
Ford another major Rivian investor has partnered with and invested in autonomy startup Argo AI
Rivian has said that their vehicles will have level 3 autonomy but if you look at the sensors the vehicles will include they would probably could be level 5 capable with the right software. Perhaps both Amazon and Ford are planning on bringing their own software to run on Rivian's hardware to turn the stock Rivian into a level 5 vehicle? It seems like it could be a smart play since the autonomy software will probably be the highest margin part of the vehicle.
Imagine Amazon with a radically different UI: a physical store on wheels that automatically stocks itself with products you’re likely to buy and then drives itself to your house. You go inside and take what you want; then the vehicle returns to base and stocks itself for the next customer.
Dark horse suggestion: Patent portfolio. It’s possible that self-driving implementations will be patent encumbered, so they’re actually buying leverage to keep the fees reasonable.
As much as a technology or retailer company, Amazon is a logisitics company and they are buying a logistics technology that may drive down their delivery costs in an extreme way.
Well Amazon has several business units that could probably do something productive with this. The obvious one is of course their delivery fleet, which at some point will start benefiting from autonomous driving. We're talking about a logistics operation that is large enough that they operate their own planes and airports.
Then there's AWS, which is also used for machine learning and other purposes. I imagine this company has a lot of the right kinds of skills they'd want to tap to offer off the shelf solutions in that space. Even if they'd ditch the product that could be valid.
But of course these solutions include bespoke self driving solutions for those car manufacturers needing to catch up in a hurry. They are going to spend billions doing so and that are going to be in need of exactly the kind of services and scale that AWS provides. IMHO most of them are going to not succeed with in house solutions and are going to be looking for something off the shelf. At least, I don't see the likes of Kia, BMW, Chrysler, etc. turning into the type of software companies that can actually do this half decently any time soon. So, that means AWS offering all or most of this as a service sounds like a good idea.
Of course the theory and practice of acquihires are two things. Most of them flat out fail and are nothing more than a big corporation scratching the backs of investors (e.g. on their board) by bailing out their failed investments. A lot of exits in the startup space fall in this category. Investors hate having their investments go bankrupt. An acquihire you can still spin as a success. A bankruptcy is much harder.
The automotive industry has plenty of OEM suppliers for various vehicle components, so there would be nothing unusual about a third party company providing self-driving hardware and software for traditional manufacturers' cars.
The companies are discussing a deal that would value Zoox at less than the $3.2 billion it achieved in a funding round in 2018, according to people familiar with the matter.
You have to remember it's a completely different situation for companies like Amazon, Cruise versus Rivian, Tesla etc.
The former has a restricted use case. Specifically, driving around specific suburban cities where if conditions are bad or issues are found they can easily just fall back to human drivers.
The latter is trying to do self-driving under all conditions, all scenarios and where it's much harder for Rivian/Tesla to determine if it's safe or not to use self driving.
I guess they are playing for a portfolio approach. With investments in a bunch of different companies, in-house drones, and robotics research. Very interested in how they integrate them into Amazon, if at all.
I feel it was only a matter of time. Amazon is competing on every front. And any transportation initiative would be core to their business. I won't be surprised if they bought an airplane company.
What are the chances that there is no rational thinking behind this deal? What I mean is, what if some Amazon exec said, "Hey, sounds like a cool project, let's throw a couple $b their way and write it off as R&D in the worst case?" Or maybe that exec is buddies with a VC who lost a ton of money on Zoox, and the exec decided to bail out the VC at the expense of Amazon shareholders? Or maybe someone at Amazon really likes Jesse Levinson?
Companies don't make billion dollar purchases with no due diligence with just one person involved in making the decision. And they don't buy companies just to immediately write them down.
Amazon is one of the most successful companies in the world. They didn't get there by acting like children.
[+] [-] tech-historian|5 years ago|reply
But the timing is interesting too IMHO. Amazon played the long game a little bit here and waited until self-driving technology was in the "trough of disillusionment" [1] to pounce on a promising startup. It saved itself billions of dollars in the process, let some of the pretenders fade away, and gave itself better odds of success.
Self-driving tech is still a slog though, and success is not guaranteed by any stretch. But I give them kudos for this.
[1] https://www.gartner.com/en/research/methodologies/gartner-hy...
[+] [-] paxys|5 years ago|reply
[+] [-] smachiz|5 years ago|reply
There are real technical challenges to making this work, not that it doesn't magically solve all things - it can't solve anything yet. The long tail still appears quite long.
[+] [-] rmason|5 years ago|reply
https://www.detroitnews.com/story/business/autos/mobility/20...
[+] [-] Gatsky|5 years ago|reply
https://www.smh.com.au/business/companies/why-self-driving-c...
[+] [-] socceroos|5 years ago|reply
[+] [-] ss45|5 years ago|reply
[+] [-] mabbo|5 years ago|reply
If you want certainty and no chance of winning big, join a Fortune 500 company. If you want uncertainty and a small chance of winning big, join a startup.
For the engineers, they wouldn't likely win big on a sale below a certain value, but that doesn't make it a total loss. They'll have a job at a FAANG with likely a higher total comp than before. Replace stock option lottery tickets with regular RSU dumps at a decent paycheck.
[+] [-] supernova87a|5 years ago|reply
[+] [-] nojito|5 years ago|reply
https://techcrunch.com/2020/04/07/nuro-gets-ok-to-test-its-d...
[+] [-] losvedir|5 years ago|reply
I live in a pretty quite neighborhood in a suburb of Chicago and I see Amazon Prime vans drive by all the time. The driver parks, hops out, runs around to grab the package, drops it off, drives on, and repeat.
This is a great scenario where you don't need total autonomy all the time: the delivery person could drive the highways and complicated parts, and then when there's a cluster of deliveries where the speed limit is 25mph and the roads are clear and regular, the van could go from house to house, with the delivery person hanging out in back or something.
It's not full self-driving, but it seems like it could make deliveries a little more efficient.
[+] [-] bryanlarsen|5 years ago|reply
Neighborhoods are the complicated parts: they're filled with pedestrians, cyclists, kids, animals, and uncontrolled intersections.
AFAICT, a self driving car has to pretty much assume that every pedestrian might suddenly dart into traffic, it can't really tell the difference between a normal pedestrian and a drunk heads down in their phone.
But keeping speeds below 20mph opens up a lot of opportunities simply because stops are basically instantaneous at that speed and because car-pedestrian collisions under 20mph are almost never fatal.
[+] [-] saos|5 years ago|reply
https://www.gov.uk/government/news/cma-provisionally-clears-...
Just how many acquisitions and fundings have they provided this year?
[+] [-] true_religion|5 years ago|reply
[+] [-] necubi|5 years ago|reply
[0] https://www.vox.com/recode/2019/12/19/21029932/amazon-logist...
[+] [-] fermienrico|5 years ago|reply
(1) Inbound bandwidth will be free.
[+] [-] michaelbuckbee|5 years ago|reply
1 - https://pitchbook.com/news/articles/ma-flashback-amazon-anno...
[+] [-] throwaway0a5e|5 years ago|reply
Because warehouses and the freight terminal (truck terminal in the consumer goods case) that necessarily surrounds them have a need for things that can roll around autonomously while not colliding with things.
There are existing solutions for this but they're far behind the state of the art for self driving vehicles that aim to eventually be used on public roads and they have several weak points that make them less competitive in very dynamic warehouses (variable contents and throughput, as opposed to a warehouse that feeds a manufacturing operation which will have more fixed contents and less variable output) that mostly handle lightweight goods (which is part of why amazon uses so much human labor in its warehouses). I wouldn't be surprised if Amazon was trying to scoop up Zoox on the cheap for their own internal use. They already bought a warehouse automation company. The probably have a list of things they want to automate and figure that buying a company that already has competence in the space (even if they're not a front runner) is the best way to meet their needs.
Of course the delivery network serving the area surrounding the warehouse facility can also make use of self driving tech but self driving on public roads is a much harder and farther off problem and I don't think solving it is their primary goal for this purchase.
[+] [-] pmorici|5 years ago|reply
Ford another major Rivian investor has partnered with and invested in autonomy startup Argo AI
Rivian has said that their vehicles will have level 3 autonomy but if you look at the sensors the vehicles will include they would probably could be level 5 capable with the right software. Perhaps both Amazon and Ford are planning on bringing their own software to run on Rivian's hardware to turn the stock Rivian into a level 5 vehicle? It seems like it could be a smart play since the autonomy software will probably be the highest margin part of the vehicle.
[+] [-] SheinhardtWigCo|5 years ago|reply
I want that.
[+] [-] coredog64|5 years ago|reply
[+] [-] topkai22|5 years ago|reply
[+] [-] jillesvangurp|5 years ago|reply
Then there's AWS, which is also used for machine learning and other purposes. I imagine this company has a lot of the right kinds of skills they'd want to tap to offer off the shelf solutions in that space. Even if they'd ditch the product that could be valid.
But of course these solutions include bespoke self driving solutions for those car manufacturers needing to catch up in a hurry. They are going to spend billions doing so and that are going to be in need of exactly the kind of services and scale that AWS provides. IMHO most of them are going to not succeed with in house solutions and are going to be looking for something off the shelf. At least, I don't see the likes of Kia, BMW, Chrysler, etc. turning into the type of software companies that can actually do this half decently any time soon. So, that means AWS offering all or most of this as a service sounds like a good idea.
Of course the theory and practice of acquihires are two things. Most of them flat out fail and are nothing more than a big corporation scratching the backs of investors (e.g. on their board) by bailing out their failed investments. A lot of exits in the startup space fall in this category. Investors hate having their investments go bankrupt. An acquihire you can still spin as a success. A bankruptcy is much harder.
[+] [-] ttmb|5 years ago|reply
[+] [-] naveen99|5 years ago|reply
[+] [-] chickenpotpie|5 years ago|reply
[+] [-] EpicEng|5 years ago|reply
Is Amazon, a company worth 1.4T dollars, trying to "pivot" into a car manufacturer? Seems unlikely, no?
Amazon has a massive logistics and delivery operation. I imagine any automation there will yield great returns.
[+] [-] ummonk|5 years ago|reply
[+] [-] samfisher83|5 years ago|reply
[+] [-] kevin_thibedeau|5 years ago|reply
[+] [-] kyle_morris_|5 years ago|reply
The companies are discussing a deal that would value Zoox at less than the $3.2 billion it achieved in a funding round in 2018, according to people familiar with the matter.
[+] [-] awjfowiefj|5 years ago|reply
[+] [-] toomuchtodo|5 years ago|reply
As my uncle used to say, you can marry more money in a moment than you can make in a lifetime.
[+] [-] wmf|5 years ago|reply
[+] [-] postingawayonhn|5 years ago|reply
[+] [-] justicezyx|5 years ago|reply
With Rivian already get investment from Amzn, the picture of autonomous & electric delivery vehicle can be quite appealing.
I guess the price would reflect how eager is Amzn to implement that strategy, and how realistic they think about the roadmap.
[+] [-] threeseed|5 years ago|reply
The former has a restricted use case. Specifically, driving around specific suburban cities where if conditions are bad or issues are found they can easily just fall back to human drivers.
The latter is trying to do self-driving under all conditions, all scenarios and where it's much harder for Rivian/Tesla to determine if it's safe or not to use self driving.
[+] [-] person_of_color|5 years ago|reply
[+] [-] organicfigs|5 years ago|reply
[+] [-] vetler|5 years ago|reply
[+] [-] presiozo|5 years ago|reply
[+] [-] sidcool|5 years ago|reply
[+] [-] baxtr|5 years ago|reply
[+] [-] arusahni|5 years ago|reply
[+] [-] mkchoi212|5 years ago|reply
[+] [-] person_of_color|5 years ago|reply
[+] [-] KKKKkkkk1|5 years ago|reply
[+] [-] threeseed|5 years ago|reply
Companies don't make billion dollar purchases with no due diligence with just one person involved in making the decision. And they don't buy companies just to immediately write them down.
Amazon is one of the most successful companies in the world. They didn't get there by acting like children.
[+] [-] blisterpeanuts|5 years ago|reply
[+] [-] Infinitesimus|5 years ago|reply