EDA execs prognosticate on 2011

Whether seen as continued growth or as a soft landing, the group of EDA leaders saw 2011 neither as a continuation of the explosive return of 2010 nor as a looming calamity, but as something in between.

By Ron Wilson, Editorial director -- EDN, January 7, 2011

Executives from EDA vendors gathered at the annual EDA Consortium CEO Forecast event this week to gaze into the always-fraught future of the EDA industry. Rather than seeing the impenetrable gloom of 2009 or the explosive rebound of 2010, the four pundits present saw a middle road: moderating growth and, behind the scenes, profound change.

Leading off the discussion was Ravi Subramanian, president and CEO of analog simulation vendor Berkeley Design Automation. While not offering a numerical forecast, Subramanian did point out some important changes in the EDA landscape.

Last year was game-changing, Subramanian said. Not only did 2010 see incredible growth rates as EDA companies snapped back from the depths of 2009, but the structure of the market changed, as well. "We saw some companies fade, and others, such as Marvell, come to the fore," Subramanian observed. "And we now have to acknowledge the rise of a fabless semiconductor industry in Taiwan."

Subramanian said we must bear in mind that this is the first capacity-constrained recovery in anyone's memory. Always before capacity has led demand into a recovery. This time, we are going into the growth phase with tight foundry capacity. No one really knows the implications of this change. But the CEO clearly believes the demand growth is there and will continue strong.

He attributes this not to some expected rebound in North America, Japan, or Western Europe, but to developing economies. "There will be 460 million people joining the middle class from developing markets by 2015," Subramanian said. "Their price expectations are forcing chip vendors into a race to deep nanometer processes." These new customers will create demand for cloud computing, smart automobiles, tablets, and 4G communications. From these needs will grow demand for new, less expensive semiconductor solutions, which in turn will create new design starts, leading to new technical challenges and demand for new tools.

Subramanian said that as design teams race to advanced technologies to meet their customers' system cost needs, they will encounter challenges with parasitics, noise, circuit mismatches, and process variations. Somewhat self-servingly, he suggested the greatest challenges would be in getting the analog/mixed-signal portions of the design right.

Speaking from a large-company point of view, Cadence Senior VP and Chief Strategy Officer Charlie Huang first apologized that his company was in a quiet period-"a very quiet period"-and so he would have to speak in general terms. That said, Huang suggested that he expected a return to more modest growth, perhaps in the 5% range. He said that green technology, cloud computing, mobility, and increased video traffic would all contribute to new design starts, but that in particular verification could be a much more dynamic market this year than it has been in the past.

Continuing down the path of non-quantitative forecasting, Synopsys chairman and CEO Aart de Geus offered his familiar, and quite abstract, definition of EDA: providing algorithmically deep tools in support of engineering problems. Within that framework, de Geus cited three areas that represented growing engineering problem sets. One was the growth in Internet traffic, for which de Geus cited a recent claim that 20% of all Internet traffic now came from Netflix. A second was genetic engineering, for which the CEO cited Craig Venter's project to create an entirely synthetic C0₂-eating microbe. Finally, he noted that the Internet of things was destined to include virtually every inanimate object that performs a function, and that its implications for human life were not yet apparent.

Mentor Graphics Chairman and CEO Walden Rhines took the discussion in a more quantitative direction. Under the projection of two separate safe-harbor foils, Rhines suggested two different forecasting models for 2011. One model, based on an amalgamation of analysts' forecasts, suggested 8% revenue growth for this year. The other, based on Rhines's tried-and-true model of projecting EDA industry revenue from the previous year's semiconductor-industry R/D spending, came up with 14% growth. Neither of these, the safe-harbor statements warned, represented Mentor's official guidance.

In the process of explaining the two models, Rhines observed that EDA growth was far from uniform, either across regions or product lines. In recent years all of the growth in EDA-industry revenue has come from the Asia-Pacific region, he claimed. Europe and North America have remained flat to down. Further, he pointed out that consistently, revenue from existing tools for the big three players has been flat, and that virtually all growth has come from new capabilities. De Geus reinforced this point with the related observation that designers will not replace an existing tool for a marginal improvement. "Customers only move when something breaks," he said.

On the whole, the picture was one of optimism. Whether seen as continued growth, as Rhines's numbers suggested, or as a soft landing, as Berkeley Design's Subramanian characterized it, the group saw 2011 neither as a continuation of the explosive return of 2010 nor as a looming calamity, but as something in between. That land in between will be driven by new technical problems at advanced process nodes, and by the emergence of strong chip-design expertise in new geographic regions. The world is the same, and yet it continues to change.