Auto OEMs Face New Competitive Threats

Automotive design and manufacturing are undergoing a fundamental shift to the left as cars increasingly are electrified and chips take over more functions formerly done by mechanical parts, setting the stage for massive disruption across a supply chain that has been in place for decades.

The success of Tesla — a company that had never actually built a chip or a car — was both a surprise and a slap in the face to the established players. The company didn’t call on traditional automotive Tier 1s such as Bosch and Continental as suppliers. Rather, it turned to companies like Mobileye, Nvidia, and others for its chips, setting in motion a rethinking of everything that goes into a vehicle, from architecture to design through manufacturing. And while OEMs have said they’ll still buy from Tier 1s, they’re also now talking to Tier 2 and Tier 3 suppliers directly.

“All of that is really the disruption in the supply chain,” noted Tom Wong, director of marketing for design IP at Cadence. “The OEMs are concerned they’re going to fall behind. And in the business of electric vehicle/autonomous driving, the industry is going to change drastically such that the first movers will obtain all the benefits, and those late to the party will only get breadcrumbs.”

Stewart Williams, senior automotive vertical marketing manager at Synopsys, agreed. “It’s the Teslas of the world that have clearly changed the world, and the other auto manufacturers, even if they’re not quite sure what happened, they feel that something’s happening and they need to respond. We see that in the various OEMs. Certain OEMs like Volkswagen are probably amongst the more proactive, and others are still trying to process what happened. At the end of it all, there will be a shift in the industry. Various companies will shift to various degrees, without a doubt, they won’t all become Tesla, but they will learn those lessons and integrate them.”

This is a new way of thinking for OEMs, and it’s being accelerated by the race to add more intelligence and sophisticated electronics into a vehicle. That requires truly taking ownership for the entire vehicle design, from the ground up, and understanding how to optimize each system within the vehicle, how to prioritize and partition data, and how to make all of this reliable, secure, and keep it up to date.

“The challenge now is, if you’re thinking about building an engine control unit to control the fuel injection in your car, the OEM doesn’t know any more than a Tier 1,” said Wong. “So at a knowledge level, they are equal. Likewise, the OEM doesn’t know more about gear boxes than the Tier 1 because it’s a well-proven technology everyone has learned, and everyone is at the same level. But with autonomous driving and ADAS, Tier 1s aren’t ahead of the OEM, because everybody was at the starting gate in 2009. For autonomous driving levels 1,2,3,4,5, everybody’s learning at the same time. This means the OEM says, ‘I’ve been doing this research for 15 years. I’m probably 5 years ahead of my supplier,’ so they actually know more than their supplier. The supplier is trying to keep learning fast, so that they continue to look credible to the OEM. Advanced SoC suppliers believe they can still supply to the OEM, and the OEM believes that the Tier 1 is at the same level they are and doesn’t know much more than the OEM. ‘Do I rely on them or do I invest in myself to try to get ahead of the game? Because I’m the OEM, I’m not trying to compete with my supplier. I’m trying to compete with the other car makers.’ The main concern becomes what they are doing.”

OEM vs. ODM
These dynamics have left room for new players to throw their hats into the ring. Case in point: Foxconn Technology Group, the primary assembler of the Apple iPhone, announced in October 2020 a software and hardware open platform for electric vehicles (EV), and detailed its EV technologies including the manufacturing of key EV components.

“Foxconn is a company with a million employees, and everybody knows they provide Apple with their manufacturing,” said Thierry Kouthon, technical product manager at Rambus Security. “What they’re doing is extremely specific and general, depending on how you look at it. It’s very specific to the EV industry. It is general because they want the entire EV industry. They want to be the Android of electric vehicles. I’m sure that there will be some success because they’re basically trying to lower the barrier of entry into the electrical vehicle market — up to the point where maybe one day you’ll be able to start a car company just by taking care of the upholstery. Everything else will be provided by Foxconn.”

While this strategy works well for smartphones, it’s not clear if it will gain traction in the automotive industry. If Foxconn or others are successful, it would be the start of original design manufacturers (ODMs) in the automotive industry. Unlike an OEM, which manufactures designs based on specs from a buyer, an ODM designs and manufactures its own products.

There is a lot of work to be done before that happens, however. Kouthon noted even though Foxconn has an EV drive train, there is a long road ahead. The automotive industry is heavily regulated, has a number of safety guarantees, and safety certifications that ODMs will need. “I’m sure that it will not be as easy for them to deploy in the German market as it is in South America or Africa or anywhere else.”

Who wins in this battle is uncertain at this point, and there are arguments for both. On the OEM side, the foundry industry has greatly accelerated the development of the fabless semiconductor business and enabled ASIC and COT business models, prompting OEMs to bring chip design back in-house. Amazon, Google, Facebook, Alibaba, and a host of other giant OEMs are now designing their own chips, and in the automotive space Tesla has been actively designing everything from the chips that are used for autonomous driving to the electronic cooling systems in batteries.

But ODMs may play a role here, as well. “The establishment of ODMs has greatly accelerated the development of desktop and laptop PCs, as well as all forms of digital consumer products,” said Cadence’s Wong. “Today the excitement in consumer desktops and laptops and smart home appliances is in industrial design and the user experience. It’s much less about the plumbing inside, which just works.”

Wong believes the same trends may be coming to the automotive space. “You can contract out manufacturing to an ‘automotive foundry’ that will build a car for you based on your own design and user experience (defined by you). An automotive platform is analogous to a reference design for a motherboard. In the case of the computer business, the motherboard is the basis for establishing branding from a ‘user design and user experience’ perspective. Is this a new trend for the automotive industry as well? How cool is it to design your own car and have the automotive foundry make 10,000 copies with your own brand?”

Wong pointed to electric bicycles (e-bikes) as an example of how this can unfold. “I can start an e-bike company without a manufacturing team or a factory. I can start with an e-bike reference platform and pick and choose the components I want, and make it my own brand. The ‘bicycle foundry’ will do the rest for me, including fulfillment. I don’t even need a storefront.”

David Fritz, senior director for autonomous and ADAS at Siemens Digital Industries Software, noted this concept is not new. “Not only has Foxconn done this with smartphones, but it’s been tried with automotive. Magna and Aptive have been trying to do this for several years. They are called Tier 0.5s, because they are not OEMs or Tier 1s. They are in between.”

The premise for attempting this in automotive certainly is different than for smart phones. For smart phones, an ODM approach began making financial sense when the market began to commoditize. For automotive, the ODM approach is driven by the need to address large increases in complexity that OEMs are not ready to address.

“When complexity is the driving force, an ODM must take on the huge task of driving the technology forward quickly and cost-effectively, as would happen naturally over time as the market matures,” Fritz said. “This has understandably been a monumental challenge. The costs involved in the rapid creation and subsumption of vast amounts of technology, and the methodologies required to manage them have the unfortunate potential effect of making ODM solutions challenging to justify financially.”

For commoditization to occur, vast amounts of advanced R&D costs must be recovered by nearly everyone within the ecosystem. This generally happens by charging higher prices to early adopters, while more effective solutions drive prices down over time as sales volumes increase. If an ODM tries to introduce products too early, this well-worn path to profits is in jeopardy, so there tends to be a good bit of resistance in the system.

“This often manifests itself as an inability to differentiate from the competition, resulting in lost market share,” Fritz said. “A standing joke is, ‘How do you make money selling a car when your only differentiation is in the number of cup holders?’”

Shifting value
Regardless, changes are underway throughout the automotive supply chain. And whenever value chains go through a lot of change, new entrants come in.

Indeed, changes in the thinking of automotive OEMs have been evident for at least the past five years, when it was clear that the OEMs intended to start making their own chips — ironically to avoid becoming “the Foxconn of cars,” said Kurt Shuler, vice president of marketing at Arteris IP. “Foxconn does things to spec. But the ideas, the intellectual property, the value-added content reside outside. The ODM lives off of very tiny margins. For the Tier 1s, they’ve always been the specialist in mechanical or hydraulic, traditional automotive electronics, things like that. And now they’re seeing that they’re getting attacked from below from their own suppliers, as the chip guys — Infineon, NXP, and others — start to create reference design systems with their own silicon that can be adopted by a Tier 1. But an OEM also could buy that directly and do their own software. So the Tier 1s also are getting attacked from above by the OEMs.”

Companies such as Tesla and GM’s Cruise have publicly stated they have their own internal silicon design teams creating their own vehicle brains, as well as ADAS and ECU devices, among other things. Other new entrants include Apple, Lyft, Uber, and Google/Waymo.

“Now there are companies from outside the well-established automotive ecosystem that’s been in place forever,” Shuler said. “A company like Foxconn may be looking at it thinking, ‘What’s our core competence? It’s in making electronic stuff. Maybe there’s a business here for us to be able to do this.”

This also speaks to the disruptive nature of computers, and of the ability to do computing pretty much anywhere and everywhere. “It destroys someone’s value to create value for someone else,” Kouthon said. “While the engine is gone, it’s replaced by a motor, and a lot of things are shifting completely. Not everybody can build engines, but pretty much anyone can build the motor.”

Ramp up the R&D
Still, the ante to play this hand is huge, and it’s complicated by the fact that much of this is brand new technology.

“I’m sure Foxconn has the money to dive in and hire a ton of engineers, but that doesn’t mean they will be successful to the degree they are accustomed,” said Fritz. “Only time will tell. Every OEM is hiring like crazy and shoring up relationships with a small group of key suppliers, so I wonder if/when an ODM supplier will have a window of opportunity to provide their solution to any of the top OEMs with enough volume to make the investment worthwhile.”

Fritz noted that a few years ago it was fashionable to claim that future cars will be like a smartphone on wheels. “That has been proven completely false. That isn’t to say that similar technology to smartphones won’t be in cars. In fact, it is today. But that technology represents perhaps 10% of the new technology that needs to be developed to meet consumer expectations after years of over-hyping by the industry. While getting a jump start on the market by applying major subsystems from an ODM, there is potentially significant danger that you will not be able to break free from the ODM while costs remain high, and your margins shrink due to the lack of significant differentiation. In 10 to 15 years, when more efficient and cost-effective technical solutions are developed, ODMs will begin to capture a large portion of the value stream in automotive. Until then, there are many technical, business and consumer-driven challenges to be overcome.”

Security concerns
Another important reason for OEMs to keep a handle on the entire system is the growing threat of cyber attacks, which takes on new importance in safety-critical systems. The more connected and sophisticated the electronics in cars, the greater the attack surface.

When it comes to applying hardware safety and security verification, it’s a complex task given that automotive systems are very complex electronic systems, noted Marc Serughetti, senior director product marketing and business development, embedded software solutions at Synopsys. “The computing part, the electronics part is becoming a huge part of automotive. We all know that. It’s not news. But what’s interesting is to see the evolution of what needs to be developed, what needs to be verified. From the perspective of several years ago, people were looking at functionality and functional verification of the design. We know that over the past four or five years, safety has become the big thing with ISO 26262, and other standards and certifications, a lot of technology is developing in this area. Then, naturally, because it’s electronic, because it’s a computer, because it’s open to the network and all sorts of things, security is the thing that’s coming together and you can’t really detach safety and security. They are tied together. There’s overlap that is important.”

“In fact,” he said, “you can think about safety and security having the same type of concept in terms of how you approach development, how you look at the design. In safety, you look at failure modes, and how to have safety mechanisms for those failure modes. Then you go through the design to the verification, all of the traditional activities. Security is parallel to this; you have to look at potential vulnerabilities, and you have to think about from the start where the design is going to be vulnerable? What do I need to do to protect against this? How do I go about verifying? The two tie together. Those are parallel flows in the design.”

Further, in automotive you want to be able to analyze as much of the whole system as you can, as early as possible, which goes with the notion of “Shift Left,” said Jason Oberg, CEO of Tortuga Logic. “The more you can analyze it in the beginning, the more accurately you can design the system, from all aspects — performance, usability and everything else. This is really important and interesting from the security side. Security is a system-level problem. If people try to think of security in a siloed mindset of just building their own Root of Trust, or adding a little security feature here or there, thinking it is good enough, because this concerns the entire system, the whole notion of virtualization and hardware virtualization, and being able to look at the security of the whole system early on can help you reason about attack patterns, what could go wrong, and help find potential weaknesses and vulnerabilities early on. When you’re going about the process of designing the final system, you’ve taken that into account so it helps drive the security strategy. It can also help drive the features you need to add, and how you should react if something unforeseen comes up.”

Oberg said from an automotive security perspective, simulation- and emulation-based verification is really important because there are so many layers, and the supply chain is fragmented, unless you’re someone like Tesla that owns the whole thing. With the exception of Tesla, most automotive companies have a fragmented supply chain model, so having ways of modeling the whole system can allow the OEM to get a really good understanding of some of the security implications of the system. If there’s a malfunction in the ADAS system, does that introduce a vulnerability? Or if someone breaks in to the infotainment system, what can they actually get access to? This information can be used as the platform for building out the rest of the key components. You can know that there are some fundamentals that you’ve designed in.”

Formal verification plays an important role here, as well, tracing through systems that are incredibly complex to find potential security holes. And that challenge will only grow as the software and firmware is updated throughout a vehicle’s projected lifetime.

“We’ve had a completeness checker that has been around probably for more than a decade, but it’s something that requires discipline, it’s iterative, it takes more expertise,” said John Hallman, product manager for trust and security at OneSpin Solutions. “But with some of the advances in automation, we’re breaking that barrier of entry much further down. So we’re applying it to items like processors where we can extract automatically extract a lot of information that’s needed, and then be able to perform these automated, complete types of checks through the formal tools very quickly. That’s a big part of where we’re going to grow here.”

He noted that it comes down to a number of tradeoffs, particularly cost. “To change the way people think about security, and that security has to be built in and not certified with a 10X price tag,” Hallman said. “What’s something that can be done with maybe a 1X or 2X type of effort?”

That will be critical for ODMs to build security into vehicles. “Especially with these EVs, they will start trimming the margins even more,” Kouthon said. “Hopefully at a certain point there is an economy of scale that happens, and the low prices can be sustained. But usually, this is what goes into the attack on the investment in security. People have a tendency to invest in features more than in security. The way it manifests in the automotive world is that often they choose security solutions that are adequate, but not necessarily the best. In our case, we drive to have software security based on hardware security. Also, you can do security only by providing secure firmware, which basically means that if your microcontroller is compromised, your security breaks.”

Conclusion
Electrification and increased levels of intelligence are transforming mechanical devices into complex electronic systems, up-ending a supply chain that was finely tuned for mechanical devices and some rather basic third-party electronics. But as those electronics increasingly define and control the vehicle, it’s not entirely clear what will differentiate one vehicle, what makes one better than another, or even who will else take responsibility if something goes wrong.

What is clear is the increasing electrification and intelligence being designed into vehicles is creating uncertainty throughout the automotive industry, and so far it’s difficult to tell who ultimately will control this market.