I’ve covered automotive technology since the early 1990s and been an enthusiast since the 1960s. Over that time, I’ve covered three big problems that automakers run up against. These problems have resulted in cars people don’t want to buy, poor customer loyalty and a lack of sustained and reliable income for both automakers and their dealers. At NVIDIA’s GTC conference this year, I saw the future of the automotive market, and much of that future addresses how to fix these endemic automotive market problems.
Let’s talk about those problems this week and how tools like NVIDIA’s Omniverse will eventually be used to fix them.
Problem #1: Time-to-market
When I was a competitive analyst, one of the training test cases was on General Motors which was thought to have the strongest competitive analysis function in the world. The company was concerned about the rise of Toyota, so in the mid-1960s, GM bought a Toyota car and did a full tear-down analysis, The results went into building GM’s Toyota killer which showed up in 1970, around five years after the car it analyzed had been in the market.
The car GM built was the Chevy Vega which remains one of the worst cars the manufacturer ever built. The problem was that, by the time GM could bring out a competitive car, Toyota had massively improved its model. The Vega wasn’t remotely competitive to the Toyotas of 1970 even though it was, as designed, far better than those available in the mid-60s.
Problem #2: Customer Loyalty
Customer retention is a huge problem in the auto industry. Switching cost between car manufacturers is very low. Customers can easily be convinced to switch from the brand they are driving to another brand. I recall my grandfather, who bought a new high-end Lincoln every three years (he had even owned the most exclusive model in that line, the Mark II Continental) suddenly deciding that Mercedes was his new brand. He abandoned Lincolns without any pushback from Ford which likely wasn’t even aware it had lost a lucrative customer.
Customer churn is a massive cost for the industry, yet manufacturers do a poor job of customer retention overall because people only seem to engage with the brand when buying a new car or having a problem. Some car makers do have advisory councils, but they are uneven and more focused on gaining insights than on truly engaging with the customer in a meaningful way. Car clubs attempt to fill the gap, but they tend to be decoupled from the manufacturer and often focused on owners who collect older cars they have often purchased used, not from the vendor.
Problem #3: Sustained income
Car makers tend to make most of their money selling new cars, while car dealers tend to make money from services and selling used cars. While a subscription model has been tried several times which could have connected the buyer more deeply to a seller, the concept of a “car as a service” only exists with apps like Uber and Lyft which generate little for the car companies and dealers.
This creates huge hill-and-valley revenue problems because, if the financial market tightens (as it is now in many markets, including the U.S.) due to high-interest rates, car companies bleed revenue. Manufacturers don’t currently have a good mechanism to weather those storms. Even leasing doesn’t help because the revenue stream generally flows to the financial institution while the carmaker gets a one-time, up-front payment.
Fixing Problem #1:
The problem is sourced in the time it takes to go from concept to finished vehicle and the inability to project future competition. NVIDIA’s solution is a combination of Omniverse and generative AI which could be used to provide buyers with an app that would let them specify what things they find most valuable and interesting in a future car. Using generative AI, they could create what their ideal car might look like several years out. In addition, generative AI and Omniverse can be used to project what competitive cars will look like several years out and factor in that information to the car design and potentially pass back the result to a core of car buyers to gauge interest and create a higher probability that the result will be competitive and attractive to the market once it arrives.
In addition, Omniverse provides a vehicle to fully prototype and present a virtually drivable result, skipping interim steps like model building and streamline testing because you are able to test in the metaverse and iterate the tests at machine, not human, speeds. This should result in a faster time-to-market and more vehicles that future buyers will find compelling.
Fixing Problem #2
Customer loyalty is generally tied to customer engagement which comes down to how often and well customers and manufacturers interact. NVIDIA has been demonstrating in-car AI capabilities for some time that would tend to connect the owner of the car to the car’s AI far more tightly. Initially, the AI would be more like a talking pet but would morph eventually into more of a digital friend or companion. The ability to build and work with the AI before it is even put in the car is a future enhancement to this capability, as will be taking the personality and personalizations related to the AI from car to car.
Thus, you’ll want to stay with the same car brand because the AI you’ve become comfortable with knows your unique interests, sense of humor, driving, and entertainment needs and preferences. This relationship will tie you more tightly to the brand and the car you drive. Done right, this should make customer retention a no-brainer for most buyers who won’t want to abandon the AI relationship they’ve built over years and eventually decades to move to another brand.
Fixing Problem #3:
Many of the coming future car services, like self-driving and streamed entertainment, will be subscription based. This should allow car makers to create better cars-as-a-service bundles and thus a sustaining revenue stream that not only ties them more tightly to their buyers but allows them to better weather market downturns like the one we are experiencing now (and the one we experienced during the pandemic). It also creates the opportunity to create revenue streams that move beyond the car into smart homes and home entertainment as drivers will want to have similar experiences and capabilities that are consistent between home and car and might even include hotel and other travel options.
Car companies are already moving away from using AIs like Amazon Alexa and favoring their own AIs with a tighter connection to their brand. It shouldn’t be long until they realize the full value of that AI relationship and move their AI, or at least the experience, into digital assistants so they can better sustain the relationship with their customers who, increasingly, want a consistent interface among all their technology. Since it’s not a car company or a consumer electronics company, NVIDIA is one of the few firms that can enable this kind of AI mobility. Jensen Huang, NVIDIA’s CEO, told me his vision is that eventually we’d own our own experience and the data that defines it. The ability to have a single AI experience, like Jarvis in the Iron Man movies, that moves with you could be incredibly compelling. It implies an eventual blending of our AI experiences until there is only one. And, regardless, of whether you are commanding your car or your home appliances or even your PC, it feels like you are interfacing with the same AI personality that knows you well and can better anticipate your wants and needs.
Wrapping up: NVIDIA is fixing the automotive market
The current automotive market is broken and has been for some time. It seems unable to properly maintain the customer relationships that define it. AI will eliminate the related problems over time, not only tying the buyer/user more tightly to their cars and car companies, but eventually expanding that experience to other parts of their lives. NVIDIA’s GTC event this year showcased the elements that will make this all possible. NVIDIA’s not only moving aggressively toward this AI future, it’s also one of the few firms driving it.