Now that automakers are racing to add infotainment systems, collision avoidance systems and even self-adjusting seats, it seems like a good time to be in the computer business.
Freescale Semiconductor Inc. in Austin, Texas, is the global auto industry's second-largest producer of microprocessors, according to Strategy Analytics.
Bob Conrad, 54, senior vice president of Freescale's automotive microcontroller division, spelled out his company's strategy in an interview with Special Correspondent David Sedgwick.
Q: How big is the market for Freescale's microcontroller division?
A: For us, it was a good year. We grew about 8 percent year on year in automotive microcontrollers. There are two standout segments: infotainment and [collision avoidance]. Both of those segments are growing from the midteens to 20 percent a year.
How long do you expect that kind of growth?
I'd say for the next five years or so. Demand for infotainment and [collision avoidance] seems to be insatiable. We upgrade the performance of our chips by a factor of two to 10 times in each generation.
Is it true that some vehicles have 100-plus processors?
Absolutely. Some vehicles have 150. The latest Mercedes S class and the high-end BMWs, they're up in that range. The global average is 25 to 27 processors.
Aside from infotainment and collision avoidance, how are microcontroller sales for other vehicle equipment?
[Microcontroller sales for] powertrains, brakes, active suspension and traction control are fairly saturated. They grow either with the market or a little behind. But we're still seeing growth for all of them. We are benefiting from North America's return to health, and last year, Europe finally stopped getting worse.
How much does Freescale spend on r&d?
The company averages around 18 percent of sales. Some businesses are higher and some are lower.
What is your global market share for microcontrollers?
We're No. 2, with a 22 percent market share. We gained a point last year. Renesas [the global leader] lost a couple of points; we picked up half of that.
Are you the top supplier of automotive microcontrollers in North America?
Yes. We're No. 1 in the United States.
Do you have a strong presence in China?
We have an assembly plant in Tianjin. For microcontrollers, the vast majority of our business there is still with the worldwide Tier 1 suppliers -- companies like Continental, Visteon, Bosch, Delphi. We also serve China's local electronics market as well. That market is fragmented, with many, many suppliers.
The 2011 earthquake in Japan destroyed a Renesas Electronics plant and your plant in Sendai. That caused a worldwide shortage of microcontrollers. What have you done to guard against future catastrophes?
You want to have two suppliers for wafers, and also for testing and assembly. You can do it if you plan ahead. We can't claim we've got everything covered, but we've made a lot of progress.
Are you doing any dual sourcing?
Yes. We can pretty quickly replicate any part in either factory [in Taiwan and Texas]. That's an example of what we're doing.
Has the global shortage of automotive semiconductors finally eased?
Yes, I believe so. In the last 18 to 24 months, capacity has not been an issue, overall. We grew pretty strongly last year and this year as well, so we have longer lead times. We have to watch a number of items quite closely, but we've increased the capacity of both fabrication plants.
Renesas is the big supplier in Japan. Has Freescale gained market share in Japan in the wake of the earthquake?
Yes, we've had a strong increase of business there over the last two or three years. It's pretty clear that the auto industry is looking for some diversity of supply. But it does take quite a while. We're growing from a small base. Renesas is the biggest in that market, and we are second.
Getting back to collision avoidance: You just introduced a microcontroller embedded in cameras. Will that help provide 360-degree vision?
Yes. In collision avoidance, we are well-established in microcontrollers for radar used for adaptive cruise control. We are a newer entrant into the vision side of it.
What is your assessment of autonomous vehicles? Is the technology equal to the task?
The technology is on a good learning curve. Our next-generation microcontroller is three times better than the last generation. But when you talk about autonomous driving, you get questions like: When do you hand over control of the car? Who is held liable by the legal system? The driver? The carmaker? The suppliers? That's the bigger challenge.
Automakers are steadily introducing features such as intelligent cruise control, lane keeping and parking assist. Is this a good indication that we can develop autonomous cars?
Stitching those functions together is a high-level problem. I don't know if the software is solved. There may be different approaches to that. But I think those are normal problems that we can grind our way through over the next few years.
Do you think the hardware -- that is, the microcontrollers -- is powerful enough for autonomous vehicles?
Adequate or nearly adequate, within a generation or so. To me, the hardware is more readily solvable. You can see how it will be resolved.
How quickly have prices been falling?
When you've got brand-new technology -- like processors for radar or vision -- you are increasing performance by two to 10 times in each generation. So the cost of performance is going down a lot, but the actual cost of the chip is fairly static. Once a technology gets more mature, at first the chip's price goes down 10 percent a year, then 5 percent, and then 2 or 3 percent a year. That may take place over five to 10 years.
What is your product cycle?
It depends on the application. With airbags, for example, there is not much change going on. The processor might be used for five to 10 years. That's pretty long. Whereas in advanced safety or infotainment, you may refresh the product every year, and you may make generational changes every couple of years.
How often do you design an all-new chip?
For infotainment, it might be three to five years. They are really quite large investments. For [collision avoidance], we're going a little faster than that right now.