The Autonomous Vehicle Founder Conundrum
How an engineer imbalance, difficulty of deploying, and VC dollars are creating a perfect storm for autonomous vehicle founders.
Over the past few months I’ve seen a wave of startups aiming to tackle full self-driving cars (Level 4 autonomy) and advance driver assistance systems (ADAS) (Level 2 autonomy) utilizing some form of deep learning.
There is certainly nuance to whether you go end-to-end or how you implement parts of the stack and make ultimate motion/path planning decisions, but that’s not important for this post. What is important is that clearly, many believe there is an opportunity here.
Companies including Drive.ai, Comma.ai, Adasworks, and a slew of other “stealth” startups are utilizing this approach, in addition to larger OEMs like Tesla. I’d estimate that today there are at least 12–15 funded startups and another 5–10 teams of engineers thinking of breaking off to tackle this problem, and those numbers will only increase over the next 6–12 months.
The issue is — where do you deploy this technology?
You have three options as a startup:
1) Aftermarket — Cruise paved the way with this approach by initially positioning themselves as a self-driving/autopilot kit for retrofitting Audis and other vehicles. They eventually pivoted away from this for a variety of reasons. Comma.ai is now doing this by selling the Comma One, an aftermarket product that will allow your car to get similar features as Tesla’s Autopilot on one often traveled stretch of highway for $999. Conceivably the One will also scale as Comma continues to capture more data.
The aftermarket approach makes sense because you are putting the buying power in (mostly) the consumer’s hands. And while large corporations won’t trust integration with a startup, consumer’s do all the time. Whether or not that is ultimately a good decision when they are speeding down a highway at 65 mph in a death box on a startup’s tech stack is unclear, but what is clear is that there will be some demand.
2) Integrate with an OEM or Tier 1 — By allowing somebody else to control the hardware (or by bundling with a hardware provider) you essentially are just selling the overall concept of Autonomy As A Service. Historically, OEMs spend a ton of money in R&D for a given vehicle and have long cycles between major upgrades. Because of this large capital expenditure, no large OEM is going to trust a startup to provide their production vehicle with a software stack. The risk is not worth the reward when larger companies like Bosch and Mobileye already own those relationships and are developing increasingly advance tools for autonomy in conjunction with larger players like Nvidia.
But history is changing.
Now there are a plethora of “startup OEMs” that are trying to capitalize on two major trends at once — electric vehicles AND autonomy. Some are backed by massive corporations, others are well-funded by VCs, and some are the aspirational projects of international billionaires. This presents an opportunity for those startups which are trying to beat existing OEMs to market with a better hardware product and might not have the capabilities to ship a polished software-side product yet. Instead, they’ll license the software/sensor fusion/autonomy stack and voila, they have an autonomous EV. Startups helping startups.
Both of these options are great and represent some form of what most startups are pitching today, but there’s actually a third option in the back of everyone’s mind:
3) Never deploy — This option is why most talented engineers sitting at larger automotive companies are at least thinking of starting a company, raising a “pre-seed” round of at least $500K-$1M, and riding the hype cycle perfectly. And I don’t blame them.
You read the insane article that states that AV acquisitions are happening at $10M/engineer and you run some math on how much money that is for you and your team. You raise that pre-seed, build something compelling enough, raise a more formal seed round to scale it to different types of production vehicles, driving scenarios, location types, etc. and then you pray you never have to go to market because options 1 and 2 sound like hell. So instead you find a nice OEM, a larger more well-funded team, or a tier 1 supplier to fit the bill for the next few years of R&D that it will take to get to market, in the form of an early acquisition.
Your VCs are probably a little pissed, but who cares, because every VC is also sitting there telling themselves that these AV bets are downside protected due to these exact scenarios. Until they’re not.
To be clear, this isn’t an attempt to disparage any of the companies targeting this problem. It is more to highlight what is happening in some cases and the thought processes echo’d by many founders. I believe that ADAS will go the route of the seatbelt and become required in all vehicles eventually, and am on record as being bullish for startups in this space whether or not autonomy commoditizes.
This post originally appeared in my notes.