EDA Moves Out Of The Shadows

This sector is again focusing on applications in much larger systems.


EDA has long harbored ambitions that are larger than a piece of silicon. The engineering challenges being solved on a nanometric scale are remarkably similar to ones being solved at a much higher level—architectural design, layout, validation, verification, debug, thermal mapping, and a lot more.

The problem, at least until recently, is that it has been difficult to gain a foothold in larger markets. Tools and supply chains are well established in those markets, and they’re working well enough. And as any startup will admit, it’s hard to displace an incumbent when business is thriving and there are no obvious threats on the horizon.

But in the past couple years, and particularly in the past 12 months, much has changed in the technology world. The rapid growth in mobility has flattened, and a number of new markets have cropped up, from automotive electronics to cloud computing, machine learning and IoT/IIoT. Those markets are considered “green-field” opportunities. Some of them, like IoT, are brand new. Others, such as automotive and IIoT, offer new opportunities in well-established markets. And EDA is in a good position to take advantage of all of them.

EDA’s transition beyond chips has been in the works for at least the past several years. Synopsys’ foray into software with the 2014 purchase of Coverity is a case in point. Cadence likewise has been investing heavily in neural networking design and embedded vision. And OneSpin has been edging its way into the machine learning market.

Siemens’ recent acquisition of Mentor Graphics is the most recent and largest test case of this approach to date.

“If you look at big mechanical systems—planes, trains and automobiles—there is no simulation platform and no multi-physics compute platform,” said Chuck Grindstaff, executive chairman of Siemens PLM Software, a business unit of the Siemens Digital Factory Division.

Siemens views software, hardware and electronics as the foundation pieces for industrial IoT and “connected intelligent devices.”

“Six or seven years ago, we were not thinking about the SoC and co-design,” said Grindstaff. “But we heard from the customer side, ‘We wished you worked together better.'”

He noted that systems companies already are moving in this direction, citing Bosch as an example of a systems vendor that is building its own foundries.

If this works—and so far there are not enough proof points—it opens the door to massive growth for the EDA industry. EDA has been running at 2% of semiconductor revenues for decades. While that’s a healthy market, it’s not the kind of growth opportunity that made EDA stocks so popular with investors back in the 1990s. It’s also offset by huge investments in R&D, because the problems being solved by EDA are some of the most difficult problems in the history of technology. Noise margins and thermal mapping at 7nm are difficult enough to solve, but it’s even harder to automate the solutions.

Perhaps even more compelling, this system-level approach takes EDA out of the lab and moves it much closer to the end customer, which in this case could be anything from automotive makers and industrial control equipment to their top-tier suppliers. After two decades of toiling behind the scenes, EDA once again may be headed for the spotlight.

Related Stories
Machine Learning Meets IC Design
There are multiple layers in which machine learning can help with the creation of semiconductors, but getting there is not as simple as for other application areas.
Verification In The Cloud
Is the semiconductor industry finally ready for EDA as a service?
Biz Talk: ASICs
eSilicon’s CEO discusses the future of scaling, the rollout of advanced packaging, and where the next big opportunities will be.