EDA Shapes Its Future

Battle lines are being drawn as the EDA industry attempts to grow outside of its traditional markets, but some say not so fast. Second of two parts.

popularity

In part one of this series, Semiconductor Engineering looked at growth within the EDA industry and the types of approaches being made to expand the scope of the markets that they serve. Scope expansion comes from the creation of new tools, the growth of companies in the IP space and the various ways in which opportunities can be found in new markets. Additional growth opportunities come from solving problems in vertical market segments such as automotive, or having an increased presence in markets where the lines are blurring. Examples of this include embedded software or the need to perform system analysis that may involve both electronics and mechanical.

In the past, the major EDA companies all followed similar paths, but with so many possibilities being presented to them, we are beginning to see a divergence in their strategies that may make the three companies look very different from each other in the future.

EDA growth opportunities are divided into several buckets. Selling new tools to existing customers is one technique that all EDA companies will continue to pursue even if the top line growth is small. “Core EDA remains the primary business for Cadence,” says Craig Cochran vice president of corporate marketing at  Cadence. “Our customers count on us to keep innovating in this area.”

The same story is told by all of the major EDA companies. However, that is the extent of the similarities in their strategies.

How Much More IP Growth Is There?
The most visible and largest growth opportunity so far has come from IP. IP now represents 25% of revenue at Synopsys and is growing at about 25% per year. But Synopsys is not the only one pursuing this strategy.

After initially failing to create an IP business in combination with its Tality consulting business, Cadence now has a successful and growing IP business. This was initially focused on verification IP with the purchase of Denali Software in 2010, but with the acquisition of Tensilica and other recent acquisitions it puts Cadence firmly in the design IP market as well. It is believed that IP now accounts for around $100M of Cadence’s revenue, with an expected 35% growth rate for Tensilica. Still, this is less than 10% of Cadence revenues.

“We are looking for ways to improve productivity,” says Cochran. “Designs are so large and people have to design a lot more, re-use a lot of what they have done before, retarget a lot of IP, and there are several approaches to doing this. One approach is by providing IP. This provides a faster growth than in core EDA.”

Mentor, meanwhile, has a small IP business, but so far it is not a significant source of revenue.

The big question is how much further can the IP market expand? Designs may contain more than 100 discrete IP blocks that cover 90% of the chip area today. While there may be room to extract some additional dollars from every chip, the pressure is on the semiconductor companies to keep chip prices low, and there will be even greater downward pressure in the IoT space. The IP companies already may be fighting to replace design wins.

IPchart
Worldwide semiconductor design IP revenue by company, 2011 and 2012 (in $ millions). Source: Gartner, March 2013

Looking Further Afield
Perhaps the biggest battle brewing is expanding the space in which EDA operates. Many of the algorithms developed to solve EDA problems are applicable to other fields, and some see this reuse of technology being more lucrative.

“Mentor’s focus is on system houses, and so we are not restricted,” explains Serge Leef, vice president of new ventures and general manager of the System-Level Engineering Division at Mentor Graphics. “If we run out of ideas about how we can help people design chips and have nothing left but to look at IP, then you are restricted. That is why we took a different path looking for markets that are adjacent—where we can leverage our core competences—and are beginning to mature and realize the value of advanced automation techniques.”

Mentor was clearly the pioneer in addressing markets outside of traditional EDA, but there are different ways in which the problem can be addressed.

“Cadence looks at markets in a more horizontal manner, rather than concentrating on vertical market segments,” says Cochran. “For example, in automotive, mixed-signal is very important and we are the leader in that space. So we are very engaged with all of the automotive providers in that area, but it is not very different from mixed-signal for other vertical markets. We will be looking at tools that solve the difficult problems that vertical markets, such as automotive, are facing. A couple of years ago we launched an automotive Ethernet solution as IP. We will continue to look at which pieces we can address in that space, how to add value.”

Synopsys also sees the automotive market as one of interest, but rather than develop tools specifically for it, the company has been developing IP that directly targets the market.

Not So Fast Says Adjacent Industry
“The primary competition is with a different ecosystem with different players,” says Leef. “There are people who are trying to solve the problems in different ways.”

But therein lies the difficulty. “Because the EDA companies don’t design chips themselves, they rarely understand the problems firsthand and, therefore, either don’t understand where the innovation should come from, or often get it wrong,” says Charlie Cheng, CEO of Kilopass Technology.

Michael Munsey, director of ENOVIA semiconductor strategy at Dassault Systèmes says not so fast. “Non-traditional EDA companies like Dassault Systèmes have been providing business applications and technology to other industries for years and now actively deliver solutions based on this technology to the semiconductor industry today.”

Munsey sees an opportunity to apply design data management and IP management software developed for other industries to similar emerging semiconductor IP management challenges.

“It all comes down to trust, and trust comes from proof of ability in the market,” continues Munsey. “We started with applicable core technology that has been proven and accepted in many different industries, which we then supplemented with targeted acquisitions.”

Another company making inroads into EDA is Ansys, which purchased Apache in 2011. “Ansys has been operating for 40 years,” says Aveek Sarkar, vice president of product engineering and customer support at Ansys-Apache. “It has managed to evolve and grow over that timespan by looking at the areas in which their customers are having the most problems.”

But that kind of acquisition has not been the norm in EDA. “Although some embedded systems suppliers are nibbling at the edges of EDA, it is more likely that the major EDA companies will be the acquirers,” believes Thomas Anderson, vice president of marketing at Breker Verification Systems. “They already span the complete SoC development process, so adding in more of the software components will be relatively painless on the technical side. The bigger challenge will be with the business models, since embedded programmers are used to working with free or relatively inexpensive tools compared to EDA products.”

Add’s Dassault’s Munsey: “It will be difficult for EDA companies to replicate Dassault Systemes’ offerings because its business experience solutions are different from typical core EDA tools. Further, we have thousands of customer examples in a broad range of industries, including aerospace, automotive and other truly high-tech industries that can provide valuable best-practices for the semiconductor industry.”

EDA Companies Placing Their Bets
There are two primary approaches to growth for the EDA industry – development of IP and expansion into adjacent fields. The three major EDA companies have each chosen a different combination of those paths. Synopsys and Cadence have jumped on the IP bandwagon and the market is abuzz with which IP company will be acquired next and by whom. Meanwhile, Mentor Graphics and Cadence have tried to migrate beyond the chip and are expanding into adjoining market spaces, although with slightly different strategies. Until a couple of weeks ago, Synopsys had made little progress in this area—until it shelled out $375M to buy Coverity, an application software verification company.