As a driver approaches her luxury SUV, the Apple Car Key feature on her phone unlocks the door with handsfree ease. Sitting in the driver’s seat, she’s greeted by a "no phone present" notification on the dashboard and realizes the personalization of the cabin is her husband’s.
Meanwhile, the owner of an EV compiles a list of his vehicle’s electronic glitches: a failed charging connector, false tire pressure warnings, delayed boot of the infotainment system, screen glitches, and a faulty backup camera.
Bugs, bugs, bugs...
Bugs are one thing; bad user interface (UI) is another. Over the past 20 years, as color displays have become a prominent feature in vehicles, we have seen few examples of well-executed user experience (UX) and UI in automotive design. Sixteen years after smartphones have paved the way, cars are rolling off the assembly line with complicated and unusable interfaces, frustrating features, and ineffective design. Some of the specific issues include:
• Messy menus and sub-menus
• Lack of self-explanatory designs
• Useless or underutilized features
• Dead space but small touch targets
• Outdated design or pseudo fancy graphics
• Too many steps needed to find the desired function
And the list goes on...
How did we get here?
Thousands of talented software engineers and UI/UX designers work at these companies, so how is the automotive industry still in this situation today? Years have passed since automakers started integrating more and more electronics into vehicles, painting themselves into a corner. Whether they realize it or not, the problem will just keep getting worse.
Car makers are assemblers. They’re mechanical and manufacturing companies, not electronics makers. They assemble control units the same way they assemble structures and body panels. But this approach doesn't work – it's counterintuitive and profoundly inefficient.
Modern vehicles are made with 40 to over 100 control units per vehicle. Some of these control units come from a single supplier, but easily more than 10 tier 1 vendors build them. This means that many of the control units are developed with chips from multiple vendors and various software stacks and hardware platforms, and the work is replicated numerous times.
At their core, most automotive control modules don’t differ much: they’re tiny computers with a communication chip with inputs and outputs that perform similar tasks. But the pieces don’t often work together as planned, despite months and months of validation, resulting in vehicles like the two examples above.
But that’s not even the worst part. The worst part is the OEMs can't fix the bugs. Why not? Because they don't own the tech and tier 1 vendors charge high fees to change a few lines of code. OEMs don't often know where to start since they don't have access to the code that’s running on their own vehicles.
The issue is the same for UI. Frankly said, the automotive industry doesn't know what to do with good software or UI talent. The processes, mindsets, and politics are counterintuitive to what software development requires. Regardless of the amount of talent invested in development, the results are almost always subpar.
Committees or A/B testing dictate the behavior of an interface, making it confusing to use. Multiple teams create various pages of a user interface, time is wasted in sync meetings, and massive, soulless hierarchies full of pseudo-experts kill creativity and deliver sub-par results.
So, can automakers fix this? The short answer is no unless the fundamental approach changes. Sure, OEMs talk about doubling down on bringing software in-house, but instead of developing a complete architecture from scratch, they only develop 10-30 percent and carry too much legacy to really optimize their systems.
Even 50 percent may improve things, but it won't solve the problem. The only way to build a true digital product with continuous bug fixes and feature updates that leverage the full potential of sensors and actuators is to start from a blank slate, build all electronics in-house, and strategically control the entire process from start to finish. The auto maker closest to currently achieving this is Tesla, and the real value of that company is in vertical integration and software. But even Tesla got its start with outsourced electronics.
The Future of Automotive Design
The next 10 years are going to be decisive for the automotive industry, and without vertical integration, it’s hard to envision how carmakers could deliver good products.
For almost a decade, Apple has had a “worst-kept secret” EV project dubbed "Project Titan", with hundreds of engineers working on mobility and autonomous technologies. A lot of partial information has surfaced, and Apple was rumored to utilize different OEMs’ platforms for their future vehicles.
There has been buzz for some time about changes in direction in this project and people leaving or being let go in droves, painting a dire narrative about the supposed failure of this project. However, I think the reality of the situation is simpler…
Apple is the paragon of vertical integration. I think it’s highly unlikely they would release a vehicle with a single line of code they don’t control. What is certain is Apple has the resources to build every automotive electronic in-house, as well as other things like traction motors. If they outsourced, it would likely be on the mechanical side, the exact opposite approach to current automotive design.
For now, what Apple plans to build is still a mystery, but I bet that whatever the form, factor, size, speed, or way it’s used, they’ll own and control the software and hardware like no other car maker. And it will be a game changer. Unless we get there first!
The Nxu Way
Nxu, Inc. is charting its own game-changing course in vertical integration. Since its inception in 2016, Nxu, Inc. (formerly Atlis Motor Vehicles) has made the conscious choice to build and own all hardware and software in-house with the goal of simplifying, optimizing, and delivering better experiences for our customers. We developed one hardware/software module called Omega that we use and will use in the huge majority of our electronic modules.
We’re on an incredible journey and the first results are encouraging. At present, we’re set to soon launch our first battery packs with battery cells, hardware and software all researched and developed entirely in house.
To learn more, watch this short video.