Echoing findings contained in his firm's 2005 worldwide EDA market trends report, issued late last year, Gary Smith, chief EDA analyst at Garter Dataquest, said he expects two or three new companies — possibly including Magma Design Automation Inc. — to emerge as leaders in the EDA space over the next three years as customers adopt electronic system level (ESL) tools. He likened it to the shift that occurred in the 1980s, when a move to register-transfer level (RTL) methodologies resulted in a new set of EDA leaders.

Over the next three years, Smith said he expects new companies to emerge and substantial acquisitions to take place, including mergers between significant companies of equal size.

According to Jim Hogan, who was a general partner at Telos Venture partners until it was dissolved by partner Cadence Design Systems Inc. last year, EDA is on the verge of a new wave of startup innovation by small companies that formed when the economy began to improve in 2003. With EDA startups typically requiring a three- to four-year incubation period, Hogan argued, promising startups formed during this period will begin to emerge in the next year or two.

"There is going to be a ton of stuff happening," Hogan said.

Though EDA revenue has been stagnant for the past four years, panelists argued that the pending move to ESL methodologies presents great opportunity for innovation among EDA and embedded software tool vendors. But panelists noted a number of challenges and criticized companies in both spaces for not fully embracing new opportunities.

Citing the results of a Gartner Dataquest study issued last year, Smith said that 27 percent of engineers were using tools developed internally. Smith said this figure, up from a typical 12 percent, reflects that companies were developing their own ESL tools because those tools are not yet commercially available. He said this situation exists because there have been no ESL "evangelists" — no major companies have had a financial stake in pushing the industry to adopt ESL methodologies.

In a normal environment, Smith said, only "power users" make their own tools because they are running three to four years ahead of the rest of the market. Big EDA companies, Smith said, have not developed these tools because there are not yet enough companies using them.

"We are [developing tools] internally because we want to be the best at what we do, and can't buy it from someone else because it's not available," said Richard Tobias, chief technology officer and vice president of engineering at Pixelworks.

"There is an opportunity for the EDA companies to recognize that a whole new family of products is needed to solve these problems," said Len Perham, chairman of Optimal Corp. "Customers need new [system level] products. If they don't come up with the new products, someone is going to step around them." EDA getting smaller chunk
According to Jim Kupec, chief operating officer at eSilicon Corp., while semiconductor industry growth over the past four years has ranged from 4 to 18 percent per year, EDA spending as a percentage of the total available market has declined from 2.5 to 1.7 percent.

"That tells me the relevancy and the ability to extract value has gone down for EDA vendors as a whole," Kupec said. He blamed a decline in average selling prices, changed in worldwide consumption, piracy and other issues.

In a strongly worded attack on EDA vendors' pricing policies, Perham, former president and CEO of Integrated Device Technology Inc., suggested that EDA companies undercut the industry by not charging enough for tools.

"The reason I paid what I paid for the software [while at IDT] is that that is what you were willing to sell it to me for," Perham said. "If you are not smart enough to get value for your product, if you want to give it to me, I'll take it."

Hogan said that as the industry matures, EDA is becoming increasingly commoditized. There is less differentiation between products in the core EDA flow, he said, and vendors are fighting for market share. Growth will occur outside the "core" EDA flow, he said, in areas of ESL and design-for-manufacturing.

Hogan and other panelists argued that EDA is on the verge of some disruptive events. Hogan suggested that one of these events will occur in the place-and-route flow, with placement technology remaining highly value but routing technology becoming commoditized.

Calling it an "absolute disaster," Smith said the embedded software tool space is far worse off than EDA.

"If you want to become the next No. 1 in EDA, develop a hardware-software compiler," Smith said, adding that this would probably create an automatic $2 billion market.

Though the common belief is that ASIC design starts have been on the decline for the past several years, panelists said this situation has begun to change.

"In the consumer space, we don't see that," Tobias said. "I see us doing a lot of designs."

Kupec said the number of ASIC design starts was on the decline in 1999 through 2002, but have been relatively stable for the past three years.

Smith said design starts have begun to rise and that he expects a big upturn in design starts in the near future.

"The ASIC world is not going away," Kupec said.