Apple, Autonomous Vehicles, & Final Mover Advantages
Why Apple's internal dynamics and broader approach to Product could be a death spell for their autonomous vehicle, Project Titan
There was a piece published over the weekend in The Information describing the difficulties of Apple’s self-driving car initiative Project Titan. In it, they describe Apple’s focus on shipping a car with no steering wheels or pedals, their failed ability to bring demos to the real world, and the team’s heavy turnover.
These problems are not new to autonomous vehicle companies, however Apple’s culture of product readiness/perfection is one that will likely kill any hopes of it achieving an AV in any near-term time horizon due to a cultural unwillingness to ship first-moving, minimum viable products.
The biggest thing that Apple has going for it as it enters new product categories is that they have an effectively infinite pile of cash to throw at new problems. However, within AVs, the difficulty that Apple faces is one of internal structure, not capital.
As you build deeply technical, highly ambitious technology, you start to see burnout occur internally at both the executive level as well as the IC/core team level over time. In Apple’s case, it’s likely that the demos and lack of progress paired with the capital expenditure is what’s making a variety of non-Titan execs eyeroll and wonder why the company’s hard earned cash isn’t going towards R&D on other, more core areas (whether that’s the AR/VR team that will likely launch a product in 2023, audio units, or mobile device teams etc.).
If we then zoom in on Titan, you have a group of engineers that have been banging their heads against the wall for likely years (or maybe a decade while at other AV companies at this point) without ever seeing their work enter the real world. When you take a “finished product” mindset as Apple does (they didn’t launch the first smartphone, they didn’t launch the first wireless earbuds, they didn’t launch the first VR/AR headset…but they likely will launch the best of all of those) you have very few milestones that get met in the interim that are considered “wins”.
Apple is a wasteland of unfinished or overextended projects that never reached the public (for better or for worse), and unfortunately AVs is a problem that needs sequencing of deployment in order to both generate revenue to make the R&D moderately sustainable, as well as to keep a growing but still limited set of talent engaged, especially when the best talent likely has been compensated enough in their career to retire relatively quickly or join a research lab and effectively retire.
The next problem is one that has plagued most of the AV world, which is a shift in technical approach as we’ve begun to push the limits of machine learning. Apple (like most) opted to begin with utilizing HD Maps, only recently shifting into going mapless. This change is clearly one that aligns with the “finished product” approach and is core to Apple due to it being inherently more scalable and able to handle 99.999999% autonomy. The difficulty is the talent required to execute on this is not structured the same as an HD Map team structure, and the whiplash that it creates internally is not conducive in a world in which at any given point, AV companies understand the finite pool of elite talent and this talent has cycled through a variety of Silicon Valley companies over the last 6 years (again, I believe this talent is expanding and is becoming less of a known universe).
The last point I’ll make that relates to both AVs as well as AR/VR dominance is that Apple is expanding what their elite competency has to be. Historically this competency has been in hardware development and vertical integration with their proprietary software. While hardware development has become re-engaged as a vector of competition in many areas (Apple’s M-series chips outperforming basically everyone, Tesla developing their own compute power for the AV stack, a variety of machine learning specific chips coming to market) the necessity for both of these emerging technologies relies on computer vision and machine learning (and of course doing so computationally efficiently in real-time), which is more of a core competency to Meta than Apple1, and to other AV companies than Apple.
In some ways, the economies of scale the Apple has operated with over the past decade has given them a massive structural advantage to other players in their areas of consumer electronics. Ironically, capital intensity is something that will or has killed a variety of players in the AV space, with Google’s ad business basically subsidizing Waymo, GMs business subsidizing Cruise, Ford subsidizing Argo, Amazon (maybe) now subsidizing Zoox, and a variety of OEMs desperate for innovation subsidizing other players via JV contracts. But as companies have ripped through tens of billions of dollars to try and solve one of the hardest technical problems in the world, it’s becoming more clear that (again) capital is not the most important differentiator.
In the end, Apple’s product-readiness may well be the right approach to take for a company that has routinely been able to enact a final mover advantage across its core categories by shipping a superior product experience, however bridging the gap from consumer electronics to automotive feels like it likely leads the company to being a final mover in things like chassis design or some sort of vehicle experience/design and more at the mid to late-stages of autonomous vehicles, and not an early winner in true autonomy2. And while this may be a profitable endeavor once autonomy proliferates throughout the world in 20, 30, or 40 years, it’s likely the lion’s share of value capture and perhaps the next behemoth of technology will be borne out of the first company to solve scalable autonomy globally.
(Disclosure: Compound is an investor in Wayve, an autonomy company)
I wrote extensively about both Apple and Meta’s approaches to utilizing multiple inflection points to build into AR/VR as a category in my essay On Inflection Points
The only thing that could change this is if Apple were to acquire Tesla. Which given the impending Twitter lawsuit…who knows.