Why Stellantis is building up its software muscles

Ned Curic joined Stellantis in 2021 as chief technology officer from Amazon, where he led Alexa Automotive. The former Toyota and Microsoft executive is charged with executing Stellantis' plan to generate $20 billion in software-related revenue by 2030, from services and subscriptions, features on demand, and data and fleet services. Curic spoke with Automotive News Europe News Editor Peter Sigal and Correspondent Nick Gibbs in a roundtable with reporters at the Paris auto show in October on how software is shaping Stellantis' future.

The new Jeep Avenger has Android Automotive as its infotainment operating system, but you are still using TomTom for maps. When are you moving to full Google Automotive Services (app suite)?

We have decided not to move to Google Automotive Services. 

Why not?

We looked at what kind of unique experiences we wanted to deliver to the customers and how we can maximize the value of the relationship between the customer and the vehicle. We found that to deliver always-fresh and always-up-to-date -- and very simple -- experiences, we need to change the way we build digital cabin software, and that most of that experience should be delivered from the cloud. Our current software platforms did not allow for these always-on, always-changing, living experiences. The OS is pretty heavy, it requires heavy updates over the air, you need to download heavy bits, and you need to plan for that, so scheduling or managing large fleets of tens of millions of vehicles becomes challenging.

What was the solution?

We decided to completely rebuild that digital experience to be cloud-centered and cloud-deployed. In the meantime we are going to continue with the existing platforms and we are going to transition with our new vehicle platforms to that new cockpit experience.

Does that mean the next generation won't use Android Automotive?

The likelihood is that it won't be Android-based, it will be a much lighter operating system for the digital cabin. You can think of it as something similar but much lighter than what somebody such as Tesla has in its vehicles.

Hyundai has just announced it will spend $12.6 billion on software development. What is your reaction to that?

We have formed our own software organization a year and a half ago, and we are already building the teams for this next-generation platform. I think every automaker needs to create that capability. We are all going to do that a little differently but owning the software capability is an absolute must, so I’m not surprised that other car companies will invest in bringing that capability in-house.

And developing their own operating system as well?

You don't have to develop an operating system; there are plenty of them available. You do have to develop your own software experiences that are critical for the interaction between the vehicle and the customer. And they need to be better than Car Play or Android Auto. Whatever a carmaker builds, it needs to be as good or better than customers get on their phone.

How do you transform from being an automaker to being a services maker? 

It comes down to two things: 1) the software and 2) the business models. At the end of the day, we are selling cars, but as we move toward selling mobility services such as with Free2Move, you phase in capability to create those services. On the other side, you need to deliver meaningful experiential services, and some of those make sense to have a business model attached, like subscriptions. Such services have to add significant value. That is not an easy thing to do, but there are subscription services that in the future you can offer, such as autonomous features. You buy the vehicle with the basic ADAS -- that's good enough, it's safe, it controls automated braking and lane departure -- a fantastic service, but maybe I want to get hands-free for the highway. Finding that delicate balance between what you need to sell in a vehicle and exceptionally high value-added services, where the customers actually want to pay for them, is the magic that we all are trying to find.

Stellantis announced recently that it will open a site in Bengalaru, India, to develop SmartCockpit software. Is that a model for future development?

We have development centers in France, Detroit and Italy, but to scale up with speed there are not enough trained engineers in these centers. India has a lot of smart people and a lot of universities, so that is why we chose to open our new center there.

How are you recruiting software talent?

Recruiting high-skilled engineers is difficult for all of us. We need to hire both highly qualified engineers who can scale the software and also have a pipeline of new college graduates. A potential slowdown in the economy could be helpful.

Will we see underfunded startups start letting go of engineers?

I think we will see that. Money is really tight now, and I suspect it will get tighter. We have (our venture capital fund) Stellantis Ventures, and we are seeing a lot more demand from startups. Six months ago when we talked to them, they would say that they aren’t interested in raising money. Now they are coming back to us and asking us to invest. By the way, you will get a discount on your entrance (investment). So, we have already seen behavior that leads us to believe that money is very, very tight. I think we are coming back to more realistic scenarios. When money was super cheap, ideas that maybe were not great were funded. I think we'll see a lot more rigor.

Would you buy some of those startups?

I think we will see some acquisitions in that space from us and from other large enterprises. 

Going back to the cockpit operating system, you say you want a lighter system, but you don't want to develop it yourself. Can you explain that a bit?

There are a lot of Linux-based light operating systems out there that you can choose from. What we want to build is a high-level abstraction layer, where we have partnered with Amazon to build much more meaningful experiences. In our digital cabin space we have partnered with Amazon and Foxconn to develop software. And then we lay that software on top of Qualcomm hardware. 

When will the first cars have this system?

Late 2024, they will come with the STLA Brain. I don't think we have announced which will be the first car.

How will Stellantis procure a supply of semiconductors to build what you want to build in the future?

Stellantis was born as the merger of two companies that came with a lot of diversity. So, the very first step was to bring diversity down dramatically. We have accomplished that by choosing for every new program a few families of chips from a few manufacturers. We are designing in such a way that we can move the chips up and down and horizontally (across platforms and components), so we don't see the kind of shortages we see today. Through that model we have secured a significant amount of flexibility for every new program. We went from hundreds and hundreds and hundreds of different chips and pins and process nodes, different I/Os, different memory configurations, to a standardized set of chips. That's on sort of an entry level, on large set MCUs (microcontrollers).Then, for high-performance chips, we have secured a relationship with Qualcomm.

Will Stellantis be participating in chip design, as some automakers say they are doing?

They're not really designing chips, they're designing parts of the chip, which is very typical in the industry. If you want to optimize functions to live in a particular part of the chip, then you do an IP (intellectual property) block -- and so, yes, absolutely, we will be looking at where we can deliver an IP block to optimize certain functions.

That said, when do you expect to have a free supply of chips?

I don't have any idea what is going to happen tomorrow or six months from now. There are capacity limitations on the larger node sizes, such as 90 nanometers and above. The future is pretty tight, no matter what. So, what we have decided is to for most of our chips to move down to 40 nm and below.

Doesn't that increase costs?

Not so much, because there is such high demand for higher-node chips that they are actually quite expensive. If you can move down to a 40 nm or below, it's not so expensive because there is a lot more space capacity in the lower nodes.

Where do you stand on so-called Level 2 Plus autonomous driving, which is “eyes on the road, hands off the wheel”?

I don't know what Level 2 Plus really means, but I can tell you about our capabilities. It means nothing in terms of regulations; you are either Level 2 or Level 3. The way we think about it is: What kind of services and experiences are you trying to deliver? We will deliver experiences that people really like and that are simple. You can push a button and it will simply work. We are very much focused on what will customers get versus some kind of arbitrary definition.

Is hands-free something that customers are asking for?

We really don't see a huge demand for hands-free driving from customers.

How will Stellantis gather and use data from its vehicles and customers?

Data is very important to us as a car company, and it's super-important for the quality of the product or vehicle, and quality of the services that we deliver. For example, how do you know which technologies people are actually using? So, when you actually track usage through data, you find out that people don't use a whole lot of features. They drive, they go to a destination, listen to entertainment, make some phone calls and they do a few things around controls like HVAC. So, that's the head, then there's a long tail of things that aren’t being used very much. There is a lot of focus and investment on the long tail and very little spent on the head -- on entertainment, where you have to make sure that content is available and fresh, and if I want to make a phone call it's super easy to do so. If I'm going to navigate, it has to be super easy. So, the data informed us that is where we need to focus, and the long tail is not so important. On vehicle quality, data is very important for things like optimizing battery usage and predictive failures.

How can you use that data? People talk about monetizing it -- I like to say the biggest monetization is if you deliver a really good experience and really good quality. That's how you monetize the data, and we are focusing on that. Customers are the holders of their data, and we are the guardians of their data. If you want to do something else it has to be up to the customer. 

How do you plan to generate 20 billion euros by 2030 with these new software-driven services?

Some additional services will be included in the sales price of a vehicle. And there are subscription services, such as hands-free driving, that will be net new revenue. Customers will subscribe for 4.99 or 5.99 (euros/dollars) a month to get much more personalized digital services. When you add all that up, 20 billion euro is quite attainable. Some of our models even show that these services could be much more lucrative.

Are these services included in the new revenue figure? 

Customers will buy them at the point of sale, so they are a net new revenue for us. On a trim option, for example, you could add a particular capability as a one-time payment or a subscription.

When will you start offering these add-on services?

We have started in some markets. In North America we are selling connected services and that business is going pretty well. We have started on the commercial side in some markets, on the fleet side. And we will be launching a big set of services with the new vehicle platforms.

Where is the big value in these services-on-demand?

A lot of it is in conveniences. Hands-free is the big one. Parking services are another, where if you go into a city you can be guided to a parking space. Or you buy a subscription for 50 automated parkings, where you push a button and the car parks itself. Subscribing to that service rather than buying an expensive option might be a good thing for most customers.