BOSTON In the midst of nearly universal upbeat forecasts for the EDA industry in 2001, Erach Desai, an analyst at Credit Suisse First Boston, has issued a controversial report predicting that EDA growth will be hampered by a semiconductor "inventory correction." Desai has also lowered his revenue targets for Cadence Design Systems and Mentor Graphics and is particularly critical of Mentor, whose stock recently hit an all-time high.
Desai may prove to be a visionary, but he's a lone voice in the wilderness right now. Vendors and analysts alike are taking issue with his conclusion that slower-than-expected semiconductor growth will hurt EDA sales in 2001.
Most observers, in fact, believe that the EDA industry is in the cusp of a huge retooling cycle for 0.18-micron designs and is heading for something like 20 percent growth in 2001. If there's a semiconductor slowdown, the prevailing view is that semiconductor vendors will still increase their EDA purchases in order to stay competitive.
Desai is hardly predicting disaster. He's lowered his EDA "macro" growth forecast for 2001 from 22 percent to a still respectable 18 percent. He's lowered revenue targets for Cadence and Mentor by around 5 percent, with a corresponding decrease in earnings. But even this modest retreat is drawing plenty of fire.
Desai argues that EDA revenues have historically been impacted three to six months after the semiconductor industry weakens. "Our guys on the semiconductor side believe there is at least a six-month inventory correction," Desai said. "This will affect spending on the EDA side."
Desai accepts Synopsys' recently lowered revenue estimates of around $680 million for the company's fiscal year ending Oct. 2001. But he's lowered his Cadence estimate from $1.475 billion to $1.410 billion, and his Mentor estimate from $640 million to $610 million, for calendar year 2001.
Desai's central thesis is "dead wrong," said John Barr, an analyst at Robertson Stephens. In the past, Barr said, the EDA industry has enjoyed as much as a year and a half of strong growth after a semiconductor industry downturn and Barr argues that there's no semiconductor downturn in sight for 2001, just lower-than-expected growth.
Still, Barr's forecast of 15 to 20 percent EDA growth for 2001 is in line with Desai's amended number. Barr believes Desai started with a number "too high to begin with" and used some questionable reasoning to reduce it.
Barr is expecting $1.5 billion revenues from Cadence and $660 million from Mentor in 2001, slightly higher than Desai's original figures before he reduced them.
Jay Vleeschhouwer, an analyst at Merrill Lynch, has roughly the same Cadence and Mentor figures as Barr. He said it's possible that semiconductor weakness will impact EDA in 2001, but not likely. "Would major semiconductor vendors react to short-term concerns like this in such a way as to impact their long-term competitiveness?" he asked. "Customers have not suggested they're planning to do anything of the kind."
Vleeschhouwer, in fact, thinks it likely that semiconductor product development spending the source of most EDA revenues will accelerate from its recent 14-16 percent compound growth rate in 2000 and 2001.
Desai's conclusions aren't drawing any support from the vendor side, even from Synopsys, whose revenue guidance Desai accepts. "I think it's inconsistent with what we've said, with what we've heard competitors are saying, and with other analysts' views," said Brad Henske, Synopsys' chief financial officer.
"They semiconductor vendors don't go fire their design engineers when there's a little overcapacity for a few quarters," Henske said. Noting Synopsys' success with Physical Compiler, Henske also disputed Desai's contention that significant revenue from the 0.18-micron retooling won't come until the second half of 2001.
"We haven't seen any slowdown. We expect the industry to be growing at 20 percent, and we expect to grow at or above the industry," said Bill Porter, chief financial officer at Cadence.
Desai said he downgraded Cadence's revenues primarily because of the lowered overall EDA growth rate. But in Mentor's case, he said, there are other factors at play.
"They are strategically not strongly positioned," Desai said. "With the exception of the physical verification line, they are at best number two or three where they have multiple competitors. They have absolutely no position in design creation. I believe that for a major vendor to remain major, they have to be in the design creation part."
Desai also questioned Mentor's claim that the company is building backlog and charged that Mentor is the only EDA vendor not moving to a subscription-based model.
A Mentor spokesman declined to comment, noting that "we haven't had any real contact with Erach for some time." But Barr came to Mentor's defense. "I think they've had a great year," he said. "They've got good backlog built and good momentum with some of their products."
"Vendors have not seen any indication that anything is slowing down," Desai acknowledged. "They're saying I'm wrong, and the stock market is telling me I'm wrong. I think the jury is out for Q1 and Q2 2001. It's hard to say who's right and who's wrong."