EDA’s Big Challenge

By Ann Steffora Mutschler
It is not news to anyone that the growth rate of the EDA industry has been less than impressive, to put it politely. Traditional EDA implementation tools have hit commodity status and something’s got to change. Thankfully, there are a host of challenges coming in the form of system-level (and higher) design, not to mention what will be required for true 3D chips.

Another harbinger of change are the cold, hard facts: EDA venture capital funding has dropped off to a startling low. According to Mentor Graphics, which tracks EDA venture capital funding activity on a quarterly basis, funding fell to $29 million in 2010 from $169 million in 2007.

EDA Venture Capital Funding Activity

2007 $169M
2008 $79M
2009 $51M
2010 $29M

(Source: Partial capture of market activity by Mentor Graphics)

When it comes to fabless companies and semiconductor suppliers, the Global Semiconductor Alliance (GSA) reported last week that in March, seven semiconductor companies raised $191.7 million, a dollar amount increase of 534.8% from February but 9.6% lower year-over-year. The trend is the same for the number of deals closed in March.

On a quarterly basis, in Q1, 21 semiconductor companies raised $407.1 million, 105.4% over Q4 but 10.3% lower than Q1 2010.

So what does all this mean? It depends on whom you ask. Aart de Geus, chairman and CEO of Synopsys, said at the company’s user’s group conference last month, “There is actually always much more innovation than one sees. The challenge is always, how do you economically make it viable because the initial phase of ideas tends to be relatively cheap from an investment point of view. Once you start investing in a sales and support channel that’s where the costs go up very radically…for all of our industry and the semiconductor industry as well, the challenge of making it is fairly tough against the growth rates.”

The economic downturn was very tough for many companies that barely made it, he continued. “On the other hand,” de Geus added, “you can say the economic downturn is where you get a new wave of startups that has not been seen yet or is barely visible. Great ideas happen wherever they’re going to happen – there are some ideas that you can develop in isolation, there are some ideas where you really need the rest of the environment.”

Expanding EDA’s reach
With this backdrop of modest growth, EDA vendors are looking to expand the traditional chip market in a number of ways. One way is to look out to nearby adjacencies, particularly as the cost to design and manufacture big, complicated chips limits the number of players.

One strategy readily apparent is acquiring companies in closely related adjacencies, which all of the top EDA vendors regularly do. But another intriguing strategy is to identify technologies that were developed for solving EDA problems that have relevance in adjacent markets and repurpose them in some way. In other words, inexpensively explore peripheral opportunities without reinventing the science, explained Serge Leef, VP of new ventures and GM of the system-level engineering division at Mentor Graphics, who is tasked with exploring opportunities for Mentor’s technology outside of traditional EDA boundaries.

Leef said the merchant semiconductor players seem to be doing a relatively good job in developing vertical platforms by creating silicon foundations for vertical spaces ranging from computational, control and memory perspectives. That makes it less desirable or effective for anybody other than merchant semiconductor companies to really want to design big chips.

“In the old days of the mid-’90s, you pretty much had to be a Nokia or Sony Ericsson or Motorola to create a smartphone. They all had to design big, complicated system chips and they were basically on their own with all this stuff. Most of the effort went into the silicon design and relatively little went into the software usability, industrial design and so on,” he said.

Leef also observed that the design community has evolved into two groups. One group of people is designing big, complicated, vertically-oriented silicon platforms. The other group are system designers that use those sophisticated platforms as a building block to construct interesting solutions for particular domains, niches and innovative use models.

Applying EDA technology outside of chip design
The EDA industry has spent 25 years on tools that solve problems by combinatorial optimization. Floorplanning and placement/routing tools are based on optimization techniques, Leef noted. “Optimization opportunities exist in a variety of system design areas. For instance, if you were to look at a modern vehicle that has 50 to 80 onboard computers and think about it from the standpoint of optimization, the problem is not terribly dissimilar. What’s different here is what you are optimizing. You are probably optimizing (1) cost so if you can eliminate a number of those computers that reduces the cost per vehicle or (2) weight by reducing overall cabling used for interconnect. The way you can manipulate the system is by moving functions from one computer to another, by rearranging those computers in the vehicle so that you can get some kind of enhanced cost, weight, performance or power management benefits. You can take sophisticated, distributed systems—like cars or airplanes—and apply optimization techniques that have matured in the EDA world for 25 years and gain substantial benefit.”

Similarly, Mike Gianfagna, vice president of marketing at Atrenta, believes that finding new markets for EDA will happen in incremental steps. “From our point of view, the market in the past few years has been miserable. If you’re a Series A or Series B EDA company, you’re the walking dead. You got $8 million of funding and you’ve got $1 million of revenue. On a good day you can sell the company for $4 million and you owe your investors $8 million—and they’re not giving you any more money. What in the world do you do? That becomes a write-off for somebody, unfortunately. There are a bunch of companies like that and some of them will do okay because they’ll be in the right place at the right time. Some of them won’t.”

However, he does see a glimmer of hope. “Everything is cyclical and I think EDA is going through another interesting part of the cycle.” Specifically, Merrill Lynch re-instituted coverage of the sector, and EDA companies are seeing a steady stream of bankers that are interested in the sector and re-instituting coverage on EDA companies.

Also, Apache Design Solutions filed its S-1, which is the first EDA IPO in more than a decade.

“I think where the opportunity lies is not in the same old world. Honestly, are we going to see a lot more place and route companies and extraction companies? I don’t think so. That part of the market is consolidating. If we define the back-end implementation flow to be synthesis, place and route, extraction, verification and tape-out, that flow has to be really fine-tuned for the target foundry. With the number of foundries in the world able to be counted on two hands, and down to one soon, why do we have an EDA market? Why don’t the five foundries simply each buy their own back-end tool set from some EDA company and be done with it. While the foundries have always maintained they didn’t want to become a tool provider, do they have a choice any more? If one does it, everyone will have to do it,” Gianfagna said.

Irrespective of if that happens, that’s not where the growth is. “The growth is at the hardware virtual prototype level and even more at the software virtual prototype level—that’s where you’re going to see the opportunity for more design starts, more creative application of SoCs, 3D makes it even more interesting because now you’ve got different ways that the supply chain can collaborate with each other with vertical stacks,” he added.

Echoing these sentiments, U.C. Berkeley professor Jan Rabaey, in a recent interview by Synopsys’ Karen Bartleson and Yvett Huygen, said the EDA industry needs to begin thinking beyond chips. Many of the tools and methodologies developed in the semiconductor world can work on a macro scale, he said, contributing to the evolution of complex systems, an understanding of the challenges that come along with them, as well as the positive and dark side of these new technologies.

With new innovation in the virtual prototyping market, VC funding is sure to ramp up. But it may have to be above the area that is now the focus of traditional EDA.