Using The Cloud To Bootstrap Your Autonomous Vehicles Startup


Startups in the Autonomous Vehicles (AVs) realm are still hot.

Sure, there is an ebb-and-flow at times, including that self-driving cars have witnessed a recent slowing of new AV firms (see my industry analysis at this link here), but you’ve got to think more widely and include the full gamut of autonomous vehicle types such as self-driving trucks, self-driving delivery carts, self-driving motorcycles, self-driving drones, self-driving submersibles, self-driving surface ships, self-driving planes, and any other manner of transportation that is amenable to autonomous capabilities.

The big picture is looking mighty fine.

Let’s go over some of the crucial success factors that underpin an AV startup that wants to make its mark. Those AV startups aim to get a viable product avidly underway and showcase that they are more than just some flash-in-the-pan semblance of vacuous ideas and words on a page.

The odds are that a savvy approach to an AV startup is going to outlast the AV startups that are based solely on whim and guts. I mention the guts aspect because there is truth to the notion that embarking upon a startup, and especially in the tough and unforgiving world of AVs, takes a sturdy and hefty amount of guts.

That being said, guts alone are insufficient, sorry to say.

Sadly, way too many AV startups never see the light of day. I see them all the time, before their collapse or fade-out, and can readily discern which will make it and which will be no more than a dream that was never realized.

To help any AV startups that are open to helpful advice, I’ll include herein some handy insights that I’ve gleaned via my time as a top exec at a major Venture Capital (VC) firm, along with ongoing efforts as an incubator mentor to AV startups and a frequent pitch judge at local and global Shark Tank-style competitions.

One overarching theme that will become apparent is that the use of cloud computing is an important bootstrapping mechanism and ought to be on the top of your list for matters to be undertaken from the get-go of your startup (for more on the cloud and AVs, see the link here) .

Strap yourself into your seat and get ready for a startup boosting ride.

We’ll begin by clarifying what the AV market space consists of.

Classifying By AV Scope

When thinking about AVs, I’ve found it useful to classify startups into one of two buckets, the AV-totality bucket, and the AV-element bucket.

The AV-totality category entails startups that attempt to do the whole ball of wax, the A to Z, the end-to-end facets of creating an autonomous vehicle. This encompasses the essence of the hardware involved, along with the entirety of the software stack. These are the so-called big-chew visionaries that say they want to reimagine what the underlying vehicular facility is, trying to begin with a fresh look at how that form of transport ought to be devised.

Those of these veritable all-you-can-eat buffet AV startups are glaringly ambitious, outlandish in their aspirations, and are tackling oftentimes more than they can adequately get their hands around. I’m not saying they shouldn’t embark upon those outsized goals, only that it will be a lot harder to pursue than they might otherwise assume or outright claim. Good luck to them and one has to earnestly root for an underdog that fights way above its weight class.

More realistic are the AV startups that go the AV-elements route.

They carve out a segment or portion of the autonomous vehicle market and aim to become the go-to choice in that niche. This might be in the sensors arena, offering a new kind of sensor, or a sensor that is less costly than the prevailing products, or a sensor that is several times as efficient or effective, and so on. It might be that the niche is chosen has to do with software, such as providing a newly crafted AI capability, perhaps leveraging Machine Learning (ML) and Deep Learning (DL) innovatively and as applied to autonomous vehicles (for my discussion about ML/DL for AVs, see the link here). Etc.

I realize that the AV-element strategy might seem timid or narrow to some that insist on the grandiose AV plays, but any startup that is truly just a startup and not already backed by heavy hitters will have a much greater chance of survival and existence by taking the elements-oriented avenue. You can be an AV-element provider and still knock things out of the ballpark.

When pitching an AV startup, make sure to clarify whether you are the AV-totality or the AV-element gambit. Believe it or not, many startups make a pitch that is so jampacked that it is not readily feasible to figure out what their aim is. A startup might emphasize the AV-totality and turnout to be the AV-element type, and meanwhile, a startup that appears to be an AV-element goer is actually an outstretched attempt to be an AV-totality.

Handy tip: When you confound or confuse potential investors and mentors, you are doing quite a disservice to your own ambitions and are usurping the limited time-attention those funding and advice-giving sources have (it will backfire on you, tarnishing your reputational goodwill).

Of course, if a startup is embarking upon a step-at-a-time approach toward an AV-totality, doing so by gradually pursuing a series of AV-elements that will plug-and-play together, this is certainly worth mentioning and can provide some satisfaction for those investors that are seeking a grand scheme.

You also don’t have to pretend to have a veiled desire to be an AV-totality and can at least explain how your AV-element fits into the AV-totality efforts taking place elsewhere.

Let’s dovetail the cloud computing aspects.

You will want to indicate how the AV-totality or the AV-element will be designed, developed, tested, and fielded. Those activities are going to require the collecting of, access to, and use of data. Those activities are also certainly going to require the running of or crafting of software.

Where will those computing resources come from?

The days of buying a server and keeping it in your backyard garage as you gather together your fellow intrepid founders and crank out your MVP (Minimally Viable Product) are no longer in vogue. Indeed, such a proclamation of how you are intending to proceed is pretty much like waving a large-sized red flag.

An investor wants to know that the funds going into the startup are going to have legs, meaning that if the whole thing falls apart, there will be something salvageable. For whatever you’ve pieced together so far, there ought to be a secure and safe place for it to be found. That’s where cloud computing comes to the fore.

By placing your data into the cloud, it means that even if that endearing garage has a tree that falls onto it, the data and software will still be available, residing securely in the cloud. It also means that you and your fellow startup creators can access the data and software from wherever they might be, at any time, and not be constrained by having to be in the particular township that your garage resides in.

In the past, some startups did not begin their founding trek by using the cloud because it was perceived as overly costly or overly arduous to get underway. They believed that the faster and easier path was to establish their own servers and try and find any software that they could get their hands on, either legally or at times under-the-table. Scrambling and street hustle was the right buzz for any tech startup such as those in the AV realm.

That’s not the way things are done nowadays.

Investors will sleep better at night (that’s important if you want to attract them and keep them engaged), by knowing that the intellectual property assets of the startup are safely in the cloud and are legitimately licensed and properly-being utilized. Nobody but nobody wants to get into some protracted legal battles over licensed software, and you can be sure that whatever vendor’s tools you choose to use will come after those that aren’t in compliance (it is their bread-and-butter to do so).

Those are primarily protective measures and highlight how the cloud helps in that perspective. There are also the go-forward aspects that the cloud also provides enabling capacity for. We can touch upon those go-forward synergies in a moment.

It will be useful to first clarify the nature of AV functionality that is likely integral to your AV startup.

AV Functionality

An important means to establish what your AV pursuit consists of can be illuminated via the use of a typical AV-functionality indication (for an extensive depiction, see the link here).

In brief, nearly all AVs are going to have these five key functional aspects:

·        Sensing of the world around the AV

·        Interpretation of the detected surroundings

·        Virtual world modeling

·        Planning for vehicular movement

·        Carrying out AV motion

If you are an AV-element startup, make sure to clearly identify which of the AV functionality domains you are inhabiting. It is okay to be in just one. It is also okay to be in more than one. Furthermore, you might be devising your AV-element to initially be in just one but are doing so with the larger eye of incrementally expanding the AV-element into the other domains.

For an AV-totality, presumably, the startup is going to attain all of those AV functionalities. It could be that the startup is going to handcraft them all from scratch, which is the proverbial reaching for the star’s ambition. More likely, the startup is going to leverage other companies or other resources that entail some of those domains. In that case, make sure to explain how those will end-up dovetailing into your efforts and what if any chokepoints this might imbue.

Let’s now consider how the cloud is essential for the go-forward activities when pursuing the chosen AV functionality.

You and your buddies want to make a new kind of sensor and seek to showcase what it can do for AVs (this will be an example that can be equally applicable to the other AV functionality stages).

The sensor will be used to collect data. The data will have to be examined to ascertain that the sensor is aptly detecting the outside world. If the sensor is glitching, this needs to be figured out and corrected. Furthermore, the data is only useful if it can be turned into something useable by the rest of the AV stack.

Okay, so you need to decide where to put the data, housing it someplace that it can be kept secured, where it can be accessed globally by your scrappy startup team and will accommodate a likely large growth in the volume of data that will be stored (increasing in some proportional way to the growth of your startup).

The cloud provides a means to store the data, doing so in a flexible manner of both how to store the data (structures and formats), and at a cost that is commensurate with the amount of volume at-hand. You can start with smaller data volumes, paying less, and then ratchet to larger volumes when the timing is right, and the cost gradually ratcheting in concordance.

You’ll likely need software for examining the data being generated via your new sensor.

Most full-bodied cloud services have available a slew of data-related online tools. This includes the predominant AI packages for Machine Learning and Deep Learning. Rather than wasting time trying to figure out where to find the right AI packages, such precious and valuable time is usually better spent by adeptly using tools that are known to work, already at your fingertips, and then getting those packages underway expeditiously for your AV efforts.

All told, the notion is that the cloud provides a ready-made platform to readily proceed with your AV startup.

Going other routes will typically entail lots of added overhead, for which the savvy investor is going to want to know why your energies and attention are being diverted to such activities when the cloud is today such an instant-on, available, and cost-reasonable option.


There is an AV insiders sage line that some like to use, admittedly perhaps a bit over-the-top, but nonetheless catchy, consisting of the assertion that you can start with the cloud, grow with the cloud, and eventually exit with the cloud for an AV startup.

The assertion seems readily apparent about the starting and growing stages of an AV startup, but might cause you to scratch your head about the allusion to the exit stage.

Here’s what that refers to.

Assume that eventually, you will have some form of exit from your startup. This might involve an IPO or maybe using a SPAC to go public (Special Purpose Acquisition Companies are the darling of the market right now). Or you might agree to be acquired by another company that is directly in AV and wants to swallow-up and embody your AV capabilities. Perhaps you’ll merge with another peer AV firm rather than being acquired outright. And so on.

Those myriads of exits are typically contingent upon the startup being considered viable for an exit. If the startup has a demonstrative weakness or difficulty, the odds are this will be discovered during the due diligence process leading to the exit. The exit itself could be soured and fall apart. Another possibility is that the exit continues unabated, but the value of the startup is demonstrably downgraded.

Consider where the cloud comes into this picture.

Using the example of a startup with a new sensor, they are courted by a larger company that makes a full suite of sensors. The interest in the startup involves being able to quickly include the new sensor into the existing suite of this larger company.

This acquiring company is already using the cloud for its AV efforts. If the AV startup is also using the cloud, the odds are that the integration of this new sensor into the suite of the acquiring company will be relatively streamlined. And this makes the startup especially attractive. Time is money, as they say. Other similar AV startups that were not using the cloud would be considered riskier choices, and costlier to bring into the fold.

The same kind of business logic applies to the IPO or SPAC routes too.

A startup in the AV arena that has all their ducks lined-up, using the cloud, and able to readily assure that the data is there, the software is there, and the whole shebang is available without any missing components is going to get a green light to proceed. Auditors that have to dig into arcane corners and off-the-beaten-path computing resources will put up huge warnings that the entire core of the startup is potentially scattered, unreproducible, and not able to hang together.

The overall mantra for an AV startup ought to be that establishing the proper infrastructure is crucial.

This includes the initial phases of a startup, such as at the pre-seed and seed stages. Trying to after-the-fact get the right infrastructure is doable, though there is an added cost and delay that can frustrate and disrupt the growth path of the AV startup.

As Mark Twain was known for having stated, do the right thing since it will gratify some people and astonish the rest. Be assured that investors in the AV realm are eager to be astonished and are continually on the search for that next gobsmacking AV startup coming down the pike.

I hope that’s your AV startup.


Show More

Related Articles

Back to top button