Autonomous Trucking Software Upstart Embark Has Quietly Gone From $5B+ To Basically Worthless (crunchbase.com) 40
Out of all the beaten-down public companies in the autonomous driving space, Embark Technology stands out as a conspicuously terrible stock market performer. From a report: The San Francisco-headquartered company, which develops autonomous driving technology for the trucking industry, has presided over a roughly 98% share price decline since going public a year ago. In the process, it's wiped out close to $5 billion in market capitalization.
Today, Embark and a few others that carried out SPAC mergers are in that weird category of companies trading below the value of cash reserves. In Embark's case, the company's recent market capitalization of $110 million is actually quite a bit lower than the $191 million cash it had at the end of Q3. In other words, investors seem to think it's worth less than nothing. What happened? What's noteworthy in Embark's case, as opposed to some other venture-backed companies that crashed so mightily, is there's no high-profile scandal. There was also no giant earnings miss, as it's a pre-revenue company. Rather, a mix of factors seem to have contributed to its fall, including apparent initial overvaluation, a sectorwide downturn and a critical report from a prolific short-seller. Collectively, those factors have contributed to erasing billions in valuation from a company that once secured backing from the most famous names in venture.
Today, Embark and a few others that carried out SPAC mergers are in that weird category of companies trading below the value of cash reserves. In Embark's case, the company's recent market capitalization of $110 million is actually quite a bit lower than the $191 million cash it had at the end of Q3. In other words, investors seem to think it's worth less than nothing. What happened? What's noteworthy in Embark's case, as opposed to some other venture-backed companies that crashed so mightily, is there's no high-profile scandal. There was also no giant earnings miss, as it's a pre-revenue company. Rather, a mix of factors seem to have contributed to its fall, including apparent initial overvaluation, a sectorwide downturn and a critical report from a prolific short-seller. Collectively, those factors have contributed to erasing billions in valuation from a company that once secured backing from the most famous names in venture.
Not all business plans work out. (Score:2)
Seriously though, just because (almost) no one thinks it is worth their money doesn't mean someone that likes their plan and chances won't buy in as a speculator. If the company somehow makes it big over the long term then the stock price will adjust accordingly.
And if the company used those stock sales to keep afloat and they go on to make it to profitability then the stock sale did it's job regardless of what anyone else thinks of the stock price right now.
Re: (Score:2)
The company may have perceived long value while the stock still gets clobbered. If the company is going to burn through too much cash before reaching profitability, they may be forced to significantly dilute the stock to bring in more cash. If the company can make their case to potential investors that the company is on the right track, the dilution will be much less.
In times of uncertainty, though, being "pre-revenue" makes things more difficult. The value of the company is simply less liquid.
Re: (Score:2)
Re: (Score:3)
What is noteworthy here is that this is part of a recent trend where young tech founders ("when Embark's co-founders Alex Rodrigues and Brandon Moak formed the venture roughly six years ago, neither was old enough to buy a beer") are given VC money to do things in a field WHERE EXPERIENCE IS CRITICAL.
I am sure -- due to their age -- neither one has worked in the trucking industry and knew next to nothing about it before "pivoting" to trucking. The founders probably sold the VCs the story that somehow the co
I think there's another factor at play: (Score:5, Interesting)
dude (Score:1)
Re: (Score:2)
Re: dude (Score:1)
Re:I think there's another factor at play: (Score:4, Interesting)
Re: (Score:3)
I bet the simple answer is no. Humans must still be present to operate the truck during parts of the routes unsupportive of driverless mode. Plus, this mode will never be able to handle all of the various situations to be encountered along a route. Is the truck just going to shutdown in the middle of the road until someone has driven in from some big Urban Area to put out road flares so that road debris can be untangled from under it or negotiate tight pathways of congested and stopped traffic? Of cours
Re: (Score:2)
Driving is actually the easy part of this job. But (and this is the point I'm clumsily trying to make) it's not the only part.
I think that's true, though I do think there's a viable business model when a local professional does the loading & unloading and all the other stuff we can't automate and the truck drives itself between cities*.
Oh yeah, there's also this to chew over: why is it that the only successful tests of these automated trucks seem to be on four-lane highways (or ones with nice, wide shoulders); with clearly-painted lane markings and fog lines; and during fair weather conditions? Ever see one of these tests in lake effect snows? How about during one of these Biblical-level, apocalyptic thunderstorms we commonly get here in Florida during the summer? I'm betting some of these investors eventually started asking themselves these same questions.
I think this is closer to the point. I think people overestimated the difficulty of an AI driving a vehicle on a clearly marked road in clear weather. And so when antonymous vehicles started doing that they figured full self driving was just around the corner.
I suspect AI that can reliably and s
Re: (Score:2)
Those sub-categories already exist, though I don't know who makes them. Automated fork-lifts fit in that rubric. I expect the domain to expand, but I've no idea how quickly. (OTOH, there are sections of freeways that I avoid in bad weather, because NOBODY can drive them safely, but people get into them without knowing this, or perhaps because they have no practical alternative.)
Re:I think there's another factor at play: (Score:5, Informative)
I'm reminded of these edge cases in this excellent talk [youtu.be] (Building the Software 2 0 Stack (Andrej Karpathy) that shows just how complicated even detecting lanes lines is.
As they say the devil is in the (implementation) details.
Oblg. Office Space Lumbergh: "Yeaaaaah, good luck with autonomous driving when human's can't even make good judgement calls."
Re: (Score:2)
I've heard the comparison with OCR (turning a scanned page into letters). Imagine if you scanned one page every minute, and every typo was a carcrash. That's self-dri
Re:I think there's another factor at play: (Score:5, Interesting)
^this x1000%.
"Trucking is not just some redneck yahoo dragging a box down a four-lane interstate and I believe some people's hubris absolutely blinds them to this fact. "
Oh, I assure you they do believe that, and it does.
I've been in international logistics for 32 years and I thank the heavens every damned day for the hard-ass job being done by unappreciated truck drivers. I might be on the dock late in the day and see some LTL guy swing down out of the cab knowing he's done that 30 times in the last 8 hours...I'd be wiped out just from that.
I'd assert, personally, that your particular niche (car carriers) is exceptionally detailed, high value (thus high risk) and complicated (home deliveries all damn day? Christ I can only imagine how much excedrin you eat....)
FWIW my opinion is that they will likely automate the longhaul driving process, because it IS the easiest part. I can see drivers end up being like old time harbor pilots. They do city and local driving, but something like this:
Secured long haul loads are dropped at the end of the day at a massive freight terminal 30 miles outside the metro, next to a major interstate. Then they go home to their dinner and families, while robo drivers do the 'simple' drive onto highway, drive to terminal at next metro, drop trailer, wait for next outbound.
Re: (Score:2)
> There's a technology for that, it's called "trains".
There's a reason that a) the vast majority of the freight in this country still goes by truck, and b) there is nearly no price advantage to intermodal for most cargoes, when (as IRL) people take into account the time delay and service crappiness. Yes, intermodal works for very, very standard things in large volumes. But service the huge majority of actual domestic freight biz? Nope. It's still trucks for dozens of good reasons.
FWIW I *started* in i
Re: (Score:2)
What you do wasn't the actual goal for replacement. What they aimed to replace was parts that could be automated away. Mainly long range driving of mundane cargo. That is where overwhelming majority of hours are spent. You can get a driver to do the rest, and it's going to be a hell of a lot less hours you need to pay him for.
The thing that they actually really failed on though isn't "difficult circumstances". What they're failing on is inability to jump from level 2 to level 3 (the goal is level 5, google
Re: (Score:2)
That's a very good insight, but your examples represent a diverse minority of the field. For ever fuel hauler I see, I see maybe 5 car haulers, and then 500 container haulers. Your job sounds complex, but there are some actual trivial parts of your industry that are ripe for automation, especially the delivery of goods between warehouses and shipping freight.
I especially don't think anyone was proposing automating the transport of hazardous chemicals (I'm not sure that would even be legal, or rather even po
Re:I think there's another factor at play: (Score:4, Insightful)
Sounds like we should just invest more heavily into rail and use trains instead of trucks, for the longhaul part anyway.
Re: (Score:2)
That's a very good insight, but your examples represent a diverse minority of the field. For ever fuel hauler I see, I see maybe 5 car haulers, and then 500 container haulers. Your job sounds complex, but there are some actual trivial parts of your industry that are ripe for automation, especially the delivery of goods between warehouses and shipping freight.
I especially don't think anyone was proposing automating the transport of hazardous chemicals (I'm not sure that would even be legal, or rather even possible to make legal until the rest of the sector has proven it can be automated without risk).
Even with that there are still certain subsectors in which the AI proponents will be very much challenged due to the way in which the cargo is routed (if not the very nature of the cargo itself). A great example here is cattle hauling: ever notice how it seems many of these cattle trucks are abso-fscking-lutely SCREAMING down the road? There's a reason for that: most coe haulers are paid by the weight of their load when they get to the slaughterhouse. The problem here is that there are situations in which t
Question for a long haul trucker (Score:2)
I get that you have aspects of your specific load-type that aren't solvable by automation - this isn't about that.
I'm more curious about your thoughts on a way to solve the weather related aspects of your post.
Everyone is (or was) really keen to just assume that AI could drive the long haul routes, but you're rightfully pointing out that there are unsolvable challenges there.
So, what if AI was only used when it was capable? Ie, when the weather was nice, the sun was up, the road lines were clear, and the pa
Re:I think there's another factor at play: (Score:4, Informative)
Speaking as a trucker myself, I think they underestimated the task they were trying to automate and at least some of the investors started to realize that. Allow me to explain: trucking, despite how brain-dead simple everyone thinks it is, isn't some stupid, monolithic field. There are very different sectors of trucking, with very different requirements of its practicioners - and, as a result, one software suite won't rule them all.
Allow me to elaborate: I currently haul privately-owned automobiles cross-country for owners who (for whatever reason) either can't or won't make the drive themselves. I pull an 8-car trailer (that is, a trailer with a capacity for 8 standard-sized autos). This means actually I drive the semi to your house or other location; load the car by driving it onto the trailer (it should be obvious how those cars get onto the trailer but, trust me, a lot of people have never even thought about it); securing the vehicle by strapping the tires to the deck; and rearranging the vehicles such that the semi doesn't exceed federally-mandated gross and per-axle weight limits, while still somehow relocating the vehicles so that the first vehicles off are the last loaded. And doing this 7 more times, at 7 more locations within 2 states. And, once loaded, I get to reverse that process 8 times at the other end of the U.S.
Driving is actually the easy part of this job. But (and this is the point I'm clumsily trying to make) it's not the only part.
Then there's the fuel haulers. The oversize/overdimension freight haulers. And what about hazmat? You really want to share the road with some robot hauling nuclear waste? Trucking is not just some redneck yahoo dragging a box down a four-lane interstate and I believe some people's hubris absolutely blinds them to this fact. There's a lot of things going on that no one's been able to automate (yet), and the only "androids" we have now can only show us Tik Tok videos.
Oh yeah, there's also this to chew over: why is it that the only successful tests of these automated trucks seem to be on four-lane highways (or ones with nice, wide shoulders); with clearly-painted lane markings and fog lines; and during fair weather conditions? Ever see one of these tests in lake effect snows? How about during one of these Biblical-level, apocalyptic thunderstorms we commonly get here in Florida during the summer? I'm betting some of these investors eventually started asking themselves these same questions.
Slashdot can be funny about carriage returns, I've put them in where you intended them (I.E. you can see when you quote him). However I'd like to say in addition to the excellent post above, there is another huge reason that haulage will be one of the last professions to be automated, liability. When someone is paying the OP to get goods from point A to point B, regardless of what said goods are, someone at point B is expecting them at a certain time. This is important because every autonomous car is designed (and for good reason) when it's unsure of what to do, to simply stop and make itself safe (and this is, as mentioned a good thing). A lorry sitting on the side of the highway isn't just not earning money, it's losing money as someone is expecting that load to be delivered, the load may be perishable and most deliveries are time sensitive. Some one is liable for that.
In addition, if an automated lorry just stops in the middle of the road, someone is liable for that too.
So someone will be needed to ensure such failures can be addressed in short order.
We've had automated trains for decades now. From many mines in Australia and other parts of the world we've had fully automated trains for decades, on tracks where one train only ever runs. However a fully trained driver is always on board and his job is mainly to press a button to prove he's still awake. Mainly, if the train stops for any reason he's there to get it started again and to the destination, manually if need be, bec
Well that was $315M well spent. (Score:4, Informative)
Not a bad web page. [embarktrucks.com]
It looks like they took the intuitive/conventional engineering route and built their framework on a potpourri of sensors: camera, LIDAR, RADAR, sonic motion detector and GPS. Their graphic showing the coverage looks just like something about the Model S from 5 years ago. That's where they seem to be.
They claim to be able to operate in any weather conditions, but don't give any details on how they achieve that.
Tesla seems to be pretty much alone in pursuing a vision-only system. It goes against most peoples' (engineers and otherwise) intuitions, but no matter how outrageous it seems, the logic for it remains solid. Provided, of course, that you have the on-board and mother-ship compute power to pull it off. And nobody but Musk seems interested in investing in that.
We will see how the Tesla FSD wide-release goes. If they are successful with that (and odds are that they will be) then that pretty much will demonstrate why enterprises like Embark are ill-fated. (or, as Musk said, relying on LIDAR makes you "doomed to fail.")
Re: (Score:3)
Back in January, Musk said he would be "shocked" if his cars weren't fully self diving, more safely than a human, by the end of this year. He's got one more month before we can add it to the very long list of overly optimistic predictions he has made about self driving.
Re: (Score:3)
Musk is full of shit. He's been promising robotaxis and FSD summon from across the country "by the end of the year" for like 5 years now.
Re:Well that was $315M well spent. (Score:4, Informative)
Seven years, in fact. First claim was 2015. Original prediction was 2016, then 2017, then 2019, robotaxi in 2020 for sure, and now he's about to be shocked that it didn't happen in 2022 either.
Autonomous Trucking Software? (Score:2)
Not to be confused with the dog DNA testing company Embark [embarkvet.com] -- or maybe that's the problem. :-)
(I saw a TV commercial about them and it's what I first thought of when reading TFT...)
Re: (Score:2)
Smart people working on stupid projects (Score:4, Insightful)
Re: (Score:2)
Re: (Score:2)
Certainly in the UK, there is a major shortage of truck drivers, and it is only going to get worse, because younger people don't want to take it up as a career. Unemployment isn't a huge problem, they are finding work elsewhere, and it isn't the only sector of the economy where there's labour shortages.
Re: (Score:2)
Why would I encourage any young person that wants a career to go into an industry that another industry is actively working on making obsolete? Also, most trucking jobs are not glamorous and they come with long hours. My uncle was very successful as a truck driver, but he owned his own rig and did all his own repairs and setup his own business. He only did local routes and he still regularly put in 12 hour days, 6 days a week.
My truck drivers that work for Albertson's regularly work 12-14 hour days and many
Re: (Score:2)
Why would I encourage any young person that wants a career to go into an industry that another industry is actively working on making obsolete? Also, most trucking jobs are not glamorous and they come with long hours.
But mostly, it's a dying industry if we are trying to actively automate it away. Why would I want my child wasting time on that field?
Don't forget that there are people working to automate pretty much everything - from front desk reception / hotel concierge / bar tending and other supposedly customer facing jobs to software development / legal research & lawyers / doctors and other supposedly skilled jobs and everything in between.
What makes you think there will be jobs in whatever industry you are in or you are thinking of for the future in a few decades?
Re: (Score:2)
As a human our very best skills are going to be the soft-skills. Dealing with other humans. If you do that well and can learn other stuff quickly, you will continue to be useful.
Besides, if all industries are gone in a few decades, the civilization won't be long to follow. Not like all our needs will magically become met and if they did, we would get bored and start fighting for shit to do.
Maybe you are right though. Maybe all the jobs will be gone in 30 years.
Truckers can't be replaced, the driving part can (Score:5, Interesting)
Not weird (Score:2)
That's not weird. It just means investors expect the company to burn $11 million plus whatever the non-cash assets can be sold off for before they give it up and liquidate.