Nvidia CEO says software will soon define the car, drive profit

Editor's note: If you would like to get stories like this, subscribe to Automotive News Europe's range of digital products. The newest is our monthly Focus on Mobility. Click here to learn more and sign up.

Nvidia CEO Jensen Huang wants the company he founded in 1993 to put together the "brain" of autonomous driving systems, and he expects the supplier's technology to play a key role in the shift to self-driving vehicles. He also predicts that it will be commonplace for a new car from a volume brand to be sold at cost within four years because profits will come from the automaker's proprietary software instead. He discussed this and more during a video chat with Automotive News Europe Associate Publisher and Editor Luca Ciferri and Managing Editor Douglas A. Bolduc.

How many fully autonomous cars will be on the road by 2030 and what percentage of them will have Nvidia technology?
By 2030, I would estimate that 20 percent of the cars on the road will have a high level of automation and most of them will have Nvidia's technology inside. Also, the vehicles making up the 20 percent will account for 50 percent of the miles driven in the locations where they are available. That is because autonomous vehicles will come in many forms -- passenger vehicles, trucks, taxis, shuttles, robot delivery -- and all of them will be able to work around the clock. They never get tired, they never lose their concentration, so, their utilization rate is going to be much higher.

When you say high level of automation do you mean Level 4 or even Level 5, which means the vehicle no longer needs a driver?
We will see Level 5 for the delivery of goods on campuses, at compounds or other areas that can be closed off. It is the most cost effective and safest option because, unlike humans, robots don't mind going slow. Robots have plenty of time. They don't get impatient. Therefore those area should be completely Level 5.
 

What about for passenger cars?
By 2030 I think autonomous vehicles will be largely Level 2, but these Level 2 vehicles will be incredible. It will almost be like driving with your mind. You think it and the car does it all while keeping you out of harm's way. By then parking and retrieving your car, highway driving and driving in traffic jams will be completely autonomous. Level 2 alone is going to be a completely transformative driving experience. AI [artificial intelligence] and robotics are not supposed replace the need for humans. They will enhance our work. They won't replace driving, but they will enhance driving. And, while I really enjoy driving, I also love using the Autopilot in my Tesla. It is very relaxing.

What are the biggest challenges facing the automotive industry over the rest of this decade?
The automotive industry is going through several revolutions at the same time. It's really quite phenomenal to see. One of the revolutions is the move toward electrification. Many people thought the EV buyers were motivated by environmental concerns but that was wrong. Yes, EV buyers care about the environment, however, an electric car is much more about luxury. It's ride is silky smooth, it's quiet, it doesn't make any bad smells, and its full every morning because you can charge at home. I haven't been to a gas station in years. That's a luxury. The industry initially missed this very important point, which is that this new functionality creates joy and delight through its convenience. The whole auto industry now understands that the EV is about luxury more so than energy savings. The second revolution is that the car industry will have to change the way it develops products in the future.

What do you mean?
A tremendous amount of the user experience, capability and functionality in the car will come through software. This will start the moment you sit down with the way that the car addresses you, engages with you and remembers your habits and preferences. The user experience is delightful because it almost seems like the car anticipates your needs whether you are driving or not. This will make you feel more comfortable and more confident in the car. There are many people who drive when they shouldn't do so. In the future, the baseline of safety will go up no matter how poor someone is at driving. The car will keep the rest of us safe and keep that person safe. The software capability inside a car will define the car.

How will that change the traditional automotive business model?
One of the most important things customers will value is the ever-progressing, ever-developing, ever-enhancing software. Since customers have the potential to be delighted by new software for as long as they own the car, the business model will fundamentally change. It is very likely that many car companies will sell cars at near cost and largely deliver the benefits to you through the software that you subscribe to or that you buy on a regular basis. When that happens and the car becomes software defined it will be imperative for the automotive industry to become excellent at software.

Why?
Because that's the only way to capture value again. We have seen this in many other industries and it is very likely we will see this in the automotive industry.

How will this transformation be different for automakers compared with other industries?
What separates the auto industry from the mobile phone and television industries is that the software that runs in the car is bespoke to the car. Therefore, the automaker will largely own the software opportunity, in addition to the car, for a very long time. That's fundamentally different than the mobile phone industry, where the mobile phone makers, with the exception of a few, largely don't benefit from the software on top. The car companies will benefit, but they have to build a car that's programmable. If the car is unable to accommodate great software, then the software that they could offer to customers will be limited.

What is the long-term effect of this change?
In the past, the functionality of the car was complete at the point of sale. In the future, the functionality of the car is largely incomplete at the point of sale, but it needs to have a great capacity for software. That way the functionality can be continuously enhanced by car companies for as long as the customer owns the vehicle. That creates tremendous economic opportunity. I believe that the world is now awakening to this incredible potential. All you have to do is look at the valuation of one car company in particular. When you do the simple math of installed base times software opportunity it is measured in hundreds of billions of dollars. This could very well be the world's largest single economy. It's a transformative time, but it's also a very exciting time. It's great for the industry.

What is the profit potential?
Let me just do some simple math. If a car company manufactures 10 million cars a year and sells them at cost with the potential to generate $5,000 of profit in software that is $50 billion dollars a year. Of course, you have to harvest that opportunity by writing great software, but it's still $50 billion a year of profit opportunity on top of making back the cost of the car. This profit completely belongs to the car company that built it. This simple math would suggest it is extraordinarily exciting.

Does the $5,000 come once or will it come every year throughout the life cycle of the model?
I'm talking about opportunity. Many people will decide against upgrading the software at the time of purchase. However, after two years they might be getting a little bit tired of the car. Then they find out that the car company has written a fantastic piece of software and for $1,000 or maybe $5,000 they can make the car feel like it's new again. This upgrade also could be made by the customer before selling the car. So, whether it's at the point of purchase or during the ownership of the car or to boost the car's resale value, there are so many opportunities to enhance that car. The return on investment is excellent.

Volkswagen Group is trying to develop its own software and it has struggled. Do you truly believe automakers will be able to transform into software developers?
The large car companies will have to own their own software. The reason is that they are no longer building cars, they are  operating fleets. Car companies will be like media companies because they will have systems on the road and they need to provide services to those systems. I have every confidence that transition will be made. However, not every company will make the transition in equal time and with equal skill. Therefore, the positions of car companies will change a little bit. However, there is no question in my mind that the companies that make that transition will be richly rewarded.

How long with this transformation take? Until 2025? Longer?
Most car companies will have to make a major transition within the next four years. The reason is because I believe mainstream cars will be sold at cost. There will be no profit margin for mainstream cars. The profit margin, however, will be enormously rich in software.

Why four years?
Because in four years' time it will be commonplace for a very good car to be sold at cost for $25,000 to $30,000 and you can decide later whether you will buy some software. It will also take four years for the technology to be developed. This is the new business model. That's why the traditional mainstream car companies must make it happen in four years. It's imperative.

Which companies have an advantage?
It's easier for startup companies to do this because there is no legacy to protect. It's a very challenging transition, but the reason why it is so easy to capture the courage needed to do this is because the opportunity is so great. Let me tell you why that makes sense: Not only do you get a car, you also get to rent a chauffeur with it. They come along with the vehicle so that you can have a more enjoyable and safer driving experience.

There is a fear that this transition will result in a massive reduction in jobs. Do you agree?
The car industry will absolutely hire more people. It just has to get to the other side of this transformation because where there are great economics there is great opportunity.

How will Nvidia and other technology-focused companies benefit from this change? Will they take over more control of the car?
We have changed the way we offer our technology so that we can enable car companies to create their own fleets and determine their own destiny. The idea that the technology industry would control everything inside the car makes no sense. It won't happen. The reason for that is very simple, the car companies are going to be fleet managers and service providers, not widget makers.

In this brave new world will Nvidia be a hardware supplier, a software supplier or a supplier of both?
Nvidia is a full platform supplier that works with the car industry however the different companies choose. For example, robotaxi companies want to develop all of their own software and operate the service themselves, but they would like to purchase Nvidia's computing solutions, and use our AI eco system and tools. We have no trouble with that. Some customers would like us to build a whole stack so they can have the capability by 2022. We are not a carmaker. We are a technology maker that wants to bring this world of AD [autonomous driving] and AI to the automotive industry. There are many different types of companies and they have different needs. We have to be flexible.

This also requires a long-term outlook, right?
We have the benefit of being quite patient. We have been working on autonomous vehicles for six years and 2024 is another four years away. This requires a company with a lot of determination, a lot of staying power, a lot of indigenous core technology that we can leverage. But we are here for the long haul.

Many auto industry executives have privately told us that when they have visited companies in Silicon Valley that the felt looked down upon. Nvidia, on the other hand, has fostered such a great working relationship that automakers freely mention your company by name. Why is that?
Nvidia has got great admiration and great respect for what the car industry has done. The leaders of these companies are doing an incredibly difficult job. We also recognize the challenges the companies face during this transformation. Never before has this industry experienced a technology that improves exponentially. In this world, software and computing compound. Then you add artificial intelligence. That means the larger your fleet is the faster you learn. The experience improves dramatically because everything is compounding itself. Every single year it's better than last year. To put this in automotive terms, when I bought a car in the past, the day I got it was the best the car would ever be. With a software-defined car, the day you buy it is the worst it will ever be. From that point forward, however, it's going to be like magic.