Automotive electronics in general and autonomous vehicles in particular represent enormous potential for the EDA market, but not in the ways you might be thinking. That was the consensus of a panel of experts who spoke at Cadence on July 16. 

Venture capitalist Jim Hogan, Frank Schirrmeister, group director, Cadence product marketing, and Ken Potts, Cadence group director for strategic planning, spent a fast-paced hour walking through the rationale for the growth of autonomous vehicles in the coming years—as well as the challenges and opportunities.

But they cautioned that for EDA companies in general and Cadence specifically to exploit the growth in design, we must think differently about the opportunity.

"Automotive is going to be great for us but we've got to think beyond ICs," Potts said.

Drivers for New Driving

What benefits will smarter and autonomous vehicles bring?

  • Safety: Statistics show that while safety has improved remarkably since 1950, still 1.2 million people are killed and more than 20 million people are injured in car accidents each year in North America. "More people die from car accidents than anything else," Hogan said.
  • Productivity: Hogan showed a slide with data from the Federal Highway Administration and Morgan Stanley Research that suggested a $422 billion gain in annual productivity if people could work in their autonomous cars.
  • Pollution reduction: More efficient automated driving and optimized routing are likely to cut carbon emissions.

This sounds like a scenic drive down profitability lane in the coming years, but, as with most things, it's not that easy.

"There's a huge increase in development challenges, especially in software," Hogan, founding partner with Vista Ventures and a Cadence co-founder, said. This requires multiple integration across systems and partners and "that's going to be tough," he added.

Helping the cause is the fact that electronics represents the majority of a car's BOM cost today—and has since it overtook the cost of metal 20 years ago, Hogan said.

In addition, there's an established but expanding ecosystem for development. Consider all the research labs from Detroit automakers that have sprung up in Silicon Valley in recent years, he added.

Also helping the cause is the explosion of software complexity. Lines of code are forecast to increase around 10 percent by 2020 for both smart phones and automobiles, yet the code gap is widening between the two. Forecasts suggest the average car will have 35 million lines of code in six years compared with about 25 million for smart phones. That gap will be roughly double what it was in 2010, according to automotive-industry figures Hogan shared.

In addition, said Hogan:

"There's a huge productivity issue: They can't create enough lines of code. So we have to figure out automated ways to do that. So we, meaning EDA, understand that sort of thing. We built compilers for a long time. We'll see the sort of EDA tricks we used on semiconductors to build code."

Security Concerns

One of the key opportunities for EDA vendors will revolve around safety and security within automotive electronics systems and specifically autonomous vehicles.

"The first time something visible happens that's tied back to that system, all hell is going to break lose," Potts said.

Schirrmeister added:

"For me, it's the applications which really give enough value to the customer that they're willing to deal with the security aspects. And to the degree the safety and security aspects are solved, they will drive this first."

Schirrmeister referenced this month's recall in Japan, involving 175,000 vehicles said to be affected by a software glitch in the engine control unit causing unintended acceleration.  These types of issues will be solved from the systems level and represent an enormous opportunity for verification, he added.

"I agree there need to be (regulations) but 170,000 cars multiplied by whatever cost, that's a budget that will drive these safety aspects," he said. Hogan added, "To the extent automotive guys do this on their own, they're better off."

So the opportunity, potential, and challenges are plain, but approaching the market requires a new way of thinking, panelists agreed. The automotive end market is estimated at $2 trillion, of which semiconductor sales represent $26 billion, with single-digit growth ahead of it, Potts said. EDA's share is even smaller of course: Automakers spend $300 million on all EDA tools to get those products to market, Potts said, adding they spend another $1.7 billion on design-productivity tools.

Potts noted, however:

"Automotive is interesting not because they spend $26 billion on semiconductors. They—over the space of the last 10 years—have gone from a technology laggard to a technology leader."

But the real opportunity for EDA is embracing a bigger piece of the pie—solving a bigger problem within the automotive electronics market.

"It's not about verifying ICs anymore; it's about verifying these systems," Potts said. "Who is the only industry that even has a clue to go about doing that? It's us."

Hogan added: "Automation, virtualization, integration—all these things EDA does. This is ripe for us to exploit."

Brian Fuller

Related stories:

- Ethernet in Cars—The Next Big Thing for Ethernet

- Google Driverless Car's Sensor, Vision, and Computing Future 

Anonymous