• Skip to main content
  • Skip to search
  • Skip to footer
Cadence Home
  • This search text may be transcribed, used, stored, or accessed by our third-party service providers per our Cookie Policy and Privacy Policy.

  1. Blogs
  2. Breakfast Bytes
  3. NXP: Can Silicon Valley Really Crack the Automakers' Code…
Paul McLellan
Paul McLellan

Community Member

Blog Activity
Options
  • Subscribe by email
  • More
  • Cancel
Automotive
NXP
ADAS

NXP: Can Silicon Valley Really Crack the Automakers' Code?

11 Jul 2019 • 7 minute read

 breakfast bytes logo The second panel at the recent NXPConnect was about Silicon Valley versus traditional carmakers (OEMs in car industry terminology). The first panel I covered recently in NXP: Self-Driving Cars: What's the Payoff for Carmakers?

Can Silicon Valley Really Crack the Automakers’ Code?

The panel was moderated by Mario Herger, consultant and author. In particular, author of the book The Last Driver's License Holder...Has Already Been Born. The panel itself was

  • Ben Landen of DeepScale
  • Sethu Gopal of Rivian
  • Anil Rachkonda of Continental
  • Leland Key of NXP
  • Helen Pan of Baidu

Mario started with a story about an earlier ear when a new industry did crack the code and pushed all the incumbents out:

 I happened to be in Mannheim recently, where Karl Benz had his first factory. In the 1880s, a woman could not manage her own money, Bertha was a widow who met a dashing engineer who wanted to put an engine on a horse-drawn carriage. He was Karl Benz. She married him. She was the first VC in the automotive industry, so to speak. But the city of Mannheim wouldn’t give a permit to test the car on public roads. So without telling her husband or the city, she took the prototype car to her mother, around 70 miles through the countryside. She was a nerdy engineering girl and had to fix a number of problems on the road. Her sons had to push it up the hill at one point. There were no gas stations, but there was a pharmacy where she bought half a gallon of benzene. The pharmacy is still there with a sign saying “first gas station in the world.” Somehow a few years later they sold 600 cars per year. But none of the carriage guys got into cars. They never cracked the code of automobiles.

 "So first question: what is the code today?"

Anil said "One word, plastics". No, wait, he said "One word, sensors". They can understand well and we have to make sense of the data. And nobody knows how to handle data better than the Valley.

Leland said that in his history of being in automotive, it was always about horsepower and cupholders. But now we are entering a transformative era.

Ben said that we are in a transition from hardware to software. The automotive industry has never been strong about developing software, yet early Teslas had more code than F-16 jet fighters. "Clearly this is not a task within the core competence of every car company. So on top of the mechatronics, we have software, and need outside capabilities that have never been germane to the industry."

 Sethu looked at the structure of the industry.

It took 100 years, or maybe just 80 to get to JIT [just-in-time]. But that came with an over-reliance on tier-1s and tier-2s in order to get anything done. Infotainment is where everyone found out because everyone knows what's happening there due to mobile. You can just put up requirements that are stale three years later with the car comes out. So OEMs need to crack that sort of practice. Consumer and mobile are much faster and more agile, but you have to bring it in-house. But how much.

Some numbers: Google has 2B lines of code. Human genome 300B, then a mouse, then car software. Se we are fourth in line. How much do you want to bring in-house?

Helen says we need collaboration. Now that automotive is high-tech AI technology, people feel that the automotive world is sexy again.

"So is it collaboration?" Manu asked. "Or should people pursue their own technology? What does Silicon Valley have going for it?"

 Helen told her own story. She moved to Silicon Valley in 2011 after 10 years in Oregon (having grown up in China).

At CES in January, we announced Apollo Enterprise, we want deep collaboration with OEMs and tier-1s. It is open to all developers and platforms, and some other large enterprises. It's on Github, so it’s an open platform. You don’t need to bake a brick to build a house.

Sethu pointed out the camaraderie in Silicon Valley:

It's like going to the gym, it's more motivational if there are other people around. Some things can’t be achieved by 20-30-50 people. Open Street Map, for example. There is power in numbers—power in the sense of belonging. No matter how we dice the talent, people are here despite the traffic. Plus the weather is good!

Ben reckoned it comes down to different philosophies on how technology should be built.

Automotive is a global industry. But also not global in the sense that each region thinks differently and has different companies. Silicon Valley has to globally please all these different entities. The fundamental difference is that Silicon Valley thinks big and figures out details later. Traditional automotive companies are is the other way: they figure out every little detail of the BOM [bill of materials] and then do the ideas later. There's probably some happy medium. Silicon Valley has “move fast and break things” versus three to four years before a car delivered to the specs, and things like charging a phone have changed.

Leland said that is changing already. "They used to think they knew everything in Detroit but there is a realization that they can't do everything on their own." But he also pointed out that there is a completely different mindset on reliability from consumer and mobile.

It's hard to build a service organization for millions of vehicles, even though it's easy for a few thousand. People don’t want to wait a week and Joe’s Garage Shop is not going to be able to fix this sort of stuff.

 Anil admitted that Silicon Valley has not yet broken the code although they are trying to do so.

The main thing about R&D is that there is a lot of risk but potentially a huge win. That fits the venture ecosystem mentality, too. Even Google operates differently, as a young company with lots of innovation.

I’ve been a chip designer for 10 years and since then on the management side. This is one place where all the greatest minds in the world get together on a regular basis. All the experts are coming here. That’s what makes the Valley so special.

Mario went back to another story:

I visited a company in France that builds autonomous chariots. Only 20% of employees even have a drivers license. But BMW is people who all love driving. So are we even talking about a single code? Or are we just creating new rules? Automotive didn’t crack the code for horse-drawn carts, it cracked the code for building engines.

Sethu pointed out that nobody wanted an iPhone. "They could never imagine what they wanted next." Which reminded me of the (probably apocryphal) quote from Henry Ford that if he'd asked his customers what they wanted they'd have asked for faster horses.

 Ben said we're not going to converge on a single solution everyone is happy with. "People argue about CDs versus MP3 versus vinyl versus streaming. There will always be a market for non-autonomous vehicles if we are willing to offer it. So we can’t oversimplify." Today, Ford has over 5,000 SKUs just for radios.

Anil warned us not to be too US-centric. "China is the biggest market today. For the next 10 years it is expected to grow much more slowly." [In fact it is on track to shrink for a second year running measured in vehicles]. "Maybe we should be wondering who will break the code first, Silicon Valley or China."

Helen echoed this, talking about a study that showed that from 2007 to 2017, over 90% of global automotive growth was in China. Many of the people there are buying a car are buying their first car. "China only started driving 40 years ago, so major cities have unified traffic signs, whereas the US grew up over a century and every state has different signs."

A final question from the audience. "Can the automotive industry crack Tesla?"

 Leland admitted that he's old and Detroit, but clearly they have brought the spirit of innovation. But:

I’m waiting to see them sell and maintain millions of vehicles. They’ve accelerated development cycles, they have over-the-air update, they have legitimized electric vehicles. Where they have not succeeded they have yet to reinvent a business model. They are an OEM who sells cars to people and as far as we can see they do it at a loss. It's a unit economics business model. If they were truly to crack it that would mean Tesla has 30% gross margin and everyone else 10%, or they have a lot of different add-ons and everyone else doesn’t. They're not there, at least for now.

 

Sign up for Sunday Brunch, the weekly Breakfast Bytes email.