Computer chips don’t grow on trees.
Imagine if they did.
In that dream world, you could be walking in a forest and find all the micro-chips that you need for your smartphone, laptop, video gaming devices, and the like.
Heck, no need to roam a forest, since the trees lining your local neighborhood would undoubtedly have those prized electronics hanging freely and ready to be plucked whenever so desired. One supposes though that the dangling computer chips might only appear on a seasonal basis, so you’d want to plan for the proper harvesting time and be ready before the chips begin precipitously falling to the ground on their own (ouch, yet another micro-chip bit the dust by a sizable fall from the top of that towering redwood tree!).
Switching back to reality, chipmakers are the ones that make our revered semiconductors, doing so in production facilities and so-called fabs.
Turns out that it is pretty costly to make those sought-after chips. In addition, there is a substantive lead time required from the initial concept of what a chip will be and subsequently manufacturing that chip, ultimately becoming available for everyday use. No magic wands will make this any easier and nor any faster. That being said, there are increasingly sophisticated ways to speed-up the time from design to produced chips, and ways too to bring down the costs, but all-in-all this is going to continue to be a relatively cost-high and long-delay-oriented process.
You’ve undoubtedly seen the blaring headlines recently that decry the existing chip shortages. This is especially bumming out the automotive sector.
Modern-day cars need gobs of computer chips.
Those ubiquitous electronics are found within nearly all aspects of a contemporary automobile. The braking system, yes, likely has some micro-chips. The transmission, sure, indubitably has micro-chips involved. We can go on like this all day long. Beyond the driving-oriented elements, there are chips needed for those fancy in-car infotainment systems that blare out those wonderful tunes and display captivating cat videos while you are on a driving journey. Etc.
If cars were the only source of demand for chips, perhaps there might be large stockpiles waiting at the doors of the automakers. Of course, the automakers need to compete with the hunger and thirst for computer chips as exhibited by all other electronic-using contraptions. Smartphone makers want chips. Refrigerator makers want chips (we certainly do need “smart” refrigerators). Anybody that makes something that can be enhanced or stand tall by adding a computer chip is either already on that bandwagon or eagerly soon going to join the parade.
In short, automakers need to stand in line with all those other fevered buyers of chips.
This brings up the equally important factor that ordering computer chips is a dicey game.
Suppose an automaker wants computer chips for their cars and realizes that the timing involves a years-ahead prediction of how many chips will be needed, based on the volume of cars anticipated to be made and sold. So, a guess is made at the need for some high number of chips to be delivered. These will ultimately be included in the cars during their assembly and manufacturing of the cars.
Meanwhile, imagine additionally that the upcoming future market for car purchases is starting to dry up and the initial rosy predictions were off-base. Darn, all those computer chips are not going to be needed. But that won’t matter to the chipmakers that are churning out the chips since they will insist that the automaker ordered them and will need to receive them, regardless of what the automaker opts to do with the vaunted pieces of electronic wizardry.
You might be thinking that it would be wiser for the automaker to have originally lowballed the number of needed chips. Thus, if the prediction for future sales began to diminish, they would not be high-and-dry with those unused computer chips. The tough problem of course is that if demand is high, and you don’t have the chips, this means that those cars aren’t going to be made that could have been selling like hotcakes.
Though this admittedly is a somewhat simplistic portrayal of the global situation, you likely get the gist that there is a fine line to walk in terms of how many chips an automaker should be ordering. If you aim too low, this could mean that you won’t be able to have enough chips to meet demand and will lose out on selling cars that people would have readily bought. If you aim too high, you could end-up with pallets upon costly pallets of unused computer chips that you’ve paid good money for and are now overpriced doorstops.
For the overall nature of today’s supply chain for computer chips, predictably is king.
Returning to the woes facing automakers right now, they had to make predictions last year about their future car sales and accordingly adjusted when the pandemic came along. Some assert that their pullback in chip orders has now left them somewhat flatfooted. When the demand for chips is coming from all angles, the automakers cannot just hop to the front of the line for those savory micro-chips.
The catchphrase being used is car-chip pain.
Others are describing the whole matter as a self-inflicted wound.
Those are fighting words for some, and they would argue that the problem has to do with insufficient available capacity to produce chips. Were there more capacity, it wouldn’t matter that the lead time forecasts were misaligned. Assuredly, if you want to get into an unpleasant fistfight with someone in the automotive industry, merely bring up the semiconductor “shortage” issue and be ready for a quite bruising discourse on the heated topic.
Anyone that is not currently in the market seeking to buy a car is probably amused by the whole conundrum and could care less that the automotive sector is feeling the squeeze of insufficiently available micro-chips. People that own stocks of automakers are likely watching the unfolding situation with rapt attention. Obviously, the automakers are anxiously trying to figure out what to do and probably are having daily nightmares, staying awake at night over the situation, and ever hopeful of dealing adequately with the dilemma.
Shifting gears, keep in mind that the future of cars will consist of true self-driving cars (I’ll explain the word “true” in a moment).
I bring this up because the same type of conundrum or challenge will undoubtedly occur whence self-driving cars become available. You see, self-driving cars are going to need even more of those micro-chips than a normal conventional car requires.
Think of self-driving cars as a micro-chips bonanza on steroids (for more details, see the link here).
Assuming that the same or similar issues about the supply chain and the back-and-forth of chip availability will prevail in the future, this means that the same disruption or perturbations could impact the making of self-driving cars. A maker of self-driving cars will need to render a prediction of how many self-driving cars they plan on making in some timeframe ahead, and chip makers will produce the requested chips accordingly. But, suppose the self-driving car predicted volume is too low or too high? Once again, akin to today’s woes, it could be that the number of self-driving cars able to be produced would be undercut by not enough chips, or the makers of the self-driving cars could get stuck with too many chips.
Some pundits claim that this will not happen because the number of predicted self-driving cars will be readily predictable, meaning that the automakers and the chipmakers will be seamlessly aligned and dancing with great aplomb. There won’t be any chip shortages for self-driving cars, and there won’t be any untoward overages of micro-chips for self-driving cars, so they assert.
This is the proverbial Goldilocks envisioned future for self-driving cars and micro-chips. The porridge won’t be too hot and it won’t be too cold. Instead, it will be just the right temperature.
Sorry, that seems like a wee bit of a fairytale.
This brings up today’s intriguing question to ponder: Will AI-based true self-driving cars get caught by these miseries of prediction-affliction and at times suffer the angst of car-chip pain, or will they skate along freely and be in micro-chip heaven?
Let’s unpack the matter and see.
Understanding The Levels Of Self-Driving Cars
As a clarification, true self-driving cars are ones that the AI drives the car entirely on its own and there isn’t any human assistance during the driving task.
These driverless vehicles are considered Level 4 and Level 5 (see my explanation at this link here), while a car that requires a human driver to co-share the driving effort is usually considered at Level 2 or Level 3. The cars that co-share the driving task are described as being semi-autonomous, and typically contain a variety of automated add-on’s that are referred to as ADAS (Advanced Driver-Assistance Systems).
There is not yet a true self-driving car at Level 5, which we don’t yet even know if this will be possible to achieve, and nor how long it will take to get there.
Meanwhile, the Level 4 efforts are gradually trying to get some traction by undergoing very narrow and selective public roadway trials, though there is controversy over whether this testing should be allowed per se (we are all life-or-death guinea pigs in an experiment taking place on our highways and byways, some contend, see my coverage at this link here).
Since semi-autonomous cars require a human driver, the adoption of those types of cars won’t be markedly different than driving conventional vehicles, so there’s not much new per se to cover about them on this topic (though, as you’ll see in a moment, the points next made are generally applicable).
For semi-autonomous cars, it is important that the public needs to be forewarned about a disturbing aspect that’s been arising lately, namely that despite those human drivers that keep posting videos of themselves falling asleep at the wheel of a Level 2 or Level 3 car, we all need to avoid being misled into believing that the driver can take away their attention from the driving task while driving a semi-autonomous car.
You are the responsible party for the driving actions of the vehicle, regardless of how much automation might be tossed into a Level 2 or Level 3.
Self-Driving Cars And Those Micro-Chips
For Level 4 and Level 5 true self-driving vehicles, there won’t be a human driver involved in the driving task.
All occupants will be passengers.
The AI is doing the driving.
AI is most definitely not a free lunch.
To have a suitable AI driving system, you need lots of computer chips. For example, there are specialized computer chips for the Machine Learning and Deep Learning aspects of AI capabilities. Also, computer chips are running the AI algorithms for keeping track of the driving scene and planning the actions needed for undertaking the driving task. There are computer chips needed for the sensors that are used by a self-driving car. You’ve got the video cameras, radar, LIDAR, ultrasonic devices, and the rest of the extensive sensor suite, all of them utilizing various computer chips.
Chips, chips, and more chips.
And that’s just the micro-chips that are devoted to the AI driving system. A self-driving car is still a car and thus this requires all those other computer chips that go into the same uses that occur in a conventional car. For whatever amount of chips embedded into a conventional car, including a car with ADAS, make sure to increase that number by the additional chips constituting the AI driving system facilities for a self-driving car.
I don’t think anyone can argue that AI-based true self-driving cars aren’t going to be chip gobbling beasts.
One aspect that silo-focused pundits oftentimes don’t do is think about the world beyond the realm of self-driving cars. In short, self-driving cars won’t be the only seeker of micro-chips. You can certainly expect that the incorporation of chips and AI-based capabilities will inhabit all manner of other types of electronic systems.
The point is that once again, there is going to be fierce competition over the availability of computer chips, by just about everyone that walks and breathes. Self-driving cars will be amid the computer chip insatiable demand.
The question is whether self-driving cars will be able to get their portion of the chip pie.
This is where some paths begin to diverge.
Some believe that the adoption of self-driving cars will be a smooth curve that gradually builds and can be perfectly predicted.
In this scenario, there are initially those timid and tiny tryouts of true self-driving cars. Out of those tryouts becomes confidence that self-driving cars can be expanded in terms of their adoption. There will be a somewhat slow up-curve as efforts ramp-up, and then it will climb like wildfire.
The reason for the leap in the curve will be that people will crave the advantages of self-driving cars. No longer will you need to find a human driver to give you a lift. Self-driving cars have a driver already included, namely an AI-based system that doesn’t tire, doesn’t need food breaks and will work tirelessly and 24×7 without any qualms or complaints. The public will tend to eschew human-driven cars in favor of using self-driving cars.
Furthermore, it is anticipated that self-driving cars are not going to drink and drive, nor drive while distracted. There are currently about 40,000 annual fatalities in the United States alone due to car crashes and around 2.5 million injuries as a result. Those horrific stats are principally attributed to the foibles of human driving. The hope and assumption are that the number of car crashes is going to drop enormously by the use of self-driving cars and a concomitant reduction in the use of human-driven cars.
Overall, the emphasis related to computer chips is that the rollout of self-driving cars will be as easy to forecast as it is to fall off a log.
By having a quite predictable semblance of how many self-driving cars will be produced, year after year, in the many years ahead, it will be relatively straightforward to line-up the needed chip supply. Chipmakers will know far in-advance what capacity to put in place and can nearly fall asleep at the wheel as they let their plants crank out those wonderments of technology.
You can even put aside the difficulties associated with competition for computer chips, since the makers of self-driving cars will be able to “guarantee” what their demand consists of, and the chipmakers will have a nice steady sales flow to the self-driving car businesses, meanwhile contending with the fluidity and staggered fluctuations with the rest of the chips-desiring marketplace.
Those self-driving cars are the ideal cash cow for chipmakers.
No more car-chip bellyaches or pains, and you can relax on the sunny beaches of Hawaii as you delight in the car-chip nirvana involving self-driving cars (as an aside, for my discussion about self-driving cars in Hawaii, see the link here).
Hold on for a moment, that above picturesque imagery seems rather quaint, akin to an enchanting storybook. Since I earlier invoked the fable of Goldilocks, perhaps now might be a handy moment to recall the famed story of the emperor that had no clothes. I mention this because the aforementioned idealized vision for the adoption of self-driving cars seems just a tad over-the-top.
Let’s revisit how the future rollout of self-driving cars is more likely to occur, including all the unfettered grime and sordid gristle.
First, realize that the existing tryouts of Level 4 self-driving cars involve an Operational Design Domain (ODD). The ODD refers to the aspect that the self-driving car will only function within a stated operational domain, such as perhaps within a boxed geographical area and only when particular weather conditions exist (sunny versus rainy, snow versus no snow, etc.), and so on. The ODD varies by the brand of a self-driving car and can vary too by the model of a self-driving car within a given brand.
The point is that even if a particular self-driving car brand can showcase sufficiently that it seems safe to be used, this is going to primarily be bounded by whatever ODD the self-driving car is operating under. This does not immediately imply that the self-driving car can be used anywhere else that you want to adopt that same brand, and in fact, the notion is that it won’t be readily applied to other areas that are beyond the tested ODD.
This becomes a kind of bottleneck in expanding the self-driving car brand adoption. Only until the ODD is widened and tested sufficiently would the self-driving car be viable for adoption in other areas that require a more suitable ODD.
That is one gotcha as it relates to forecasting the volume of self-driving cars.
Another catch will be the safety topic. This one is a mouthful and I’ll just briefly explain it here (for more of my analysis, see this link here).
Right now, almost everyone in the self-driving car industry is holding their breath. The reason for this being on edge is that we all know that if even one instance occurs of a self-driving car that somehow gets involved in a car crash and someone is killed (recall the Phoenix prior instance, as covered extensively), the full weight of regulators and public opinion about self-driving cars could turn on a dime, overnight.
The counterargument about being overly reactive to one such instance is that the use of self-driving cars is projected to ultimately save lives, in terms of reducing demonstratively the 40,000 car crash-related deaths occurring annually in the U.S. In short, suppose that the number of car crash-related fatalities dropped in half (just for sake of discussion), we should all be heralding the vast number of lives saved. In that sense, if one person let’s say is harmed, and certainly, all lives are precious and vital, the matter ought to be viewed in the context of the overall number of lives being saved, one might assert.
Reality suggests that even one such instance will be a staggering blow to all self-driving car initiatives. There will be an immense inspection of the self-driving cars arena in a manner that heretofore has not taken place. Currently, things are relatively quiet since there haven’t been any recent clear-cut self-driving car blowback incidents.
The outcry and clamor will be deafening if triggered into action. And, mark my words, that time will come.
From the computer chips perspective, this suggests that the predictability of self-driving car adoption is once again hampered. You can make all the predictions you want, but if the big freeze suddenly happens due to regulatory oversight or public outrage, those forecasts go out the window.
We also don’t know how things are going to go in terms of whether people will avidly want to ride in self-driving cars. The argument made about the ease of going for a ride and not having to find a human driver is one side of the coin. Other sides of the coin include that perhaps some people won’t trust the AI driving system, some people won’t like interacting with an AI system and will prefer to have a human driver, and so on.
There is a sinking feeling among some that this building of a better mousetrap might not be grandly and immediately accepted by the world at large. We all tend to assume that by-gosh if self-driving cars were to exist, they would be irresistible and they would be needed by all. This is not the kind of assumption that you can necessarily take to the bank.
Do not misinterpret the somewhat doomsday appearing remarks as though self-driving cars aren’t going to make the cut. I am a firm and fervent believer in the future of self-driving cars.
The real point to the portrayal of the self-driving car adoption lifecycle is that it is not going to be as rosy as some might imagine or that they are pitching with glee. The real-world of rolling out self-driving cars are going to be a battle. There will be moments of great success, and other instances of setbacks and difficulties.
Trying to make a nice-looking curve out of the starts and stops of self-driving car adoption does not comport with what will really happen. It is going to be jagged, ups and downs, and not something that you can layout as though the numbers of self-driving cars and the needed numbers of computer chips are superbly going to be aligned.
As the tale of Little Red Riding Hood informs us, a seemingly peaceful and uneventful trip ahead is not always in the cards. There are wolves assuredly in the way of self-driving car widespread adoption, and will inexorably upend any expectation of a delightful and smooth romp in these woodlands.
Might as well face it, there will be computer chip shortages that get entangled in the arrival of self-driving cars. The happy side of that face is that self-driving cars will persevere and there will ultimately be a massive boon, in both self-driving cars and the shortages of computer chips for them.
If only we had magic beans that could be gently tossed into the ground and we would readily have self-driving cars that sprouted out of the ground (well, perhaps that’s a fable for another day).