What Would Joe Do?
Peggy Aycinena is a freelance journalist and Editor of EDA Confidential at www.aycinena.com. She can be reached at peggy at aycinena dot com.
EDAC CEO Panel: Practically perfection
March 14th, 2013 by Peggy Aycinena
From the podium in San Jose’s DoubleTree Hotel, Jasper Design Automation President & CEO Kathryn Kranen introduced tonight’s EDAC CEO Forecast Event as being “practically perfection” and she was right. With 97 people in the room, the event ran for 97 minutes and the audience [undoubtedly] gave the panel discussion a 97% approval rating. Kudos to all involved, including EDAC for hosting, and OCP-IP, Mod Marketing, and the DoubleTree for sponsoring the event.
Kranen started off the evening by bragging on good news out of EDA: It’s up and to the right for revenue in the industry, with a 4.9 percent increase between 3Q11 and 3Q12. She cited increased stock valuations over the last year for ARM [37%], Cadence [30%], Mentor [26%], PDF Solutions [98%], and Synopsys [17%] as an indication of the viability of EDA as an investment vehicle: If you’d put $100 into each of these companies a year ago, she said, you would have netted a 41% increase in a portfolio today worth $706.90, beating out other investment indices such as the NASDAQ and S&P 100 over the same time period.
[Kranen didn’t mention, however, that if you’d put all $500 into PDF Solutions a year ago, your portfolio would be almost $1000 at this point. Perhaps no one in EDA would be so imprudent as to put all of their eggs in one basket?]
In her capacity as EDAC Chair, Kranen celebrated five new members of the consortium: Blackcomb Design Automation, Sigasi, Vayavya Labs, VWORKS/ASTC, and Tech-X Corp. Considering the rate of acquisitions of smaller EDA companies by bigger EDA companies these days, five new members of EDAC is most certainly something to celebrate.
Kranen talked about the new EDAC website, as well as upcoming industry events: DATE in March, the Jim Hogan/Joe Costello interview in May, DAC in June, and an EDA Reunion Event at the Computer History Museum in October. Reflecting on the EDA industry’s renewed sense of community and optimism, Kranen then introduced Needham & Co’s Rich Valera, who in turn introduced the evening’s panelists:
* ARM President Simon Segars
Valera’s discussion with the five panelist over the next hour was based on their responses to a recent poll of EDAC members regarding various issues swirling about in the industry today. The panelists and tonight’s audience saw the slides showcasing the results of the poll at the same time.
Of the poll respondents, 75% said consolidation has not helped innovation, although 50% believe it has helped pricing. Approximately 50% believe the three big companies will consolidate down to two in the next several years.
Rich Valera asked for a response from the panelists.
Aart de Geus – Unfortunately, the question doesn’t include a definition of innovation. We distinguish between point innovation and systemic innovation, and in the area of the latter, Synopsys is the most innovative company in the industry. Consolidation has not hurt us at all.
Wally Rhines – For once, we agree with Aart [laughter]. Startups developing a new technology and then being acquired is the source of innovation. Consolidation, meaning acquisition, is the way innovation is implemented in EDA.
Lip-Bu Tan – We provide real value to our customers as we continue to innovate, and last year we raised out prices which is very appropriate. We allow our customers be more successful.
Wally Rhines – We are on the same learning curve in EDA as the semiconductor industry. As we grow, the unit cost of tools per transistor changes in exactly the same way as the unit cost per transistor. When the transistor unit costs go down, so do the tools.
Rich Valera – And the three big companies consolidating down to two?
Wally Rhines – For the past 15 years, Mentor, Cadence, and Synopsys have provided structure in EDA, while the overall market share for these three companies has remained the same. The industry’s very stable with the current situation.
Lip-Bu Tan – As an industry matures, it consolidates. But it is also good for the customers to offer choices to them.
Raul Camposano – The real question should be: Will we become part of a bigger industry? For instance, should we be part of the fabs? Or perhaps become part of a company like ANSYS, a company with revenues higher than the EDA companies, but lower than ARM. We definitely need to think bigger than just the die.
Of the poll respondents, approximately 30% believe that Moore’s law can be extended through FinFET technology, 50% are not sure what would help, 20% believe “other” technologies would solve the conundrum, while a resounding 0% believe EUV lithography will provide any help at all.
Rich Valera asked for responses.
Aart de Geus – Moore’s law is not broken. Nonetheless, FinFET is a fabulous discontinuity, with 14-nanometer FinFET’s not long in coming out of the fabs.
Wally Rhines – Moore’s law is exponential, and even Gordon Moore acknowledged that no exponential is forever. Mathematically, however, Moore’s law of scaling is becoming irrelevant as the cost-per-transistor learning curve stays the same with either FinFET or 3D packaging. Meanwhile, we have so much capacity in place, we can cruise through 28 and 20 nanometers. Most importantly, whether stacked dies or FinFETs, it all provides opportunity for innovation and growth in EDA.
Of the respondents, 29% believe it’s a net positive, 38% believe it’s a net negative, and 33% believe it’s a net neutral.
Rich Valera noted that companies like Apple, Samsung, and Google have benefited from verticalization in product areas such as smart phones and tablets. He asked if those companies are buying more tools, or if not, does it mean a net negative for EDA.
Simon Segars – Verticalization makes no difference to EDA, because there will always be people approaching design from many different ways, with lots of different performance points. Just because a few companies are doing things a certain way – companies like Apple, Google or Samsung – doesn’t mean there aren’t still lots of choices about how to do things, whether the industry is integrated or disaggregated.
Wally Rhines – When in doubt, I refer back to the data. The concentration in cell phone providers peaked in 2007; today that sector is not as consolidated as it was 6 years ago. Consolidation is never a permanent state of affairs.
Lip-Bu Tan – There is a net gain with verticalization, because growth then is clearly application-driven.
Per the respondents, today’s food chain ranking includes: 1) IP companies, 2) System companies, 3) Fabless companies, 4) Software Companies, 5) IDMs, 6) Foundries.
In 5 years, however, the respondents have re-arranged the rankings: 1) IP companies, 2) Software companies, 3) System companies, 4) Fabless companies, 5) IDMs, 6) Foundries.
Rich Valera noted that the rankings indicate IP companies will continue to reside at the top of the food chain, while other sectors will rise or fall with a re-ordering over the next 5 years.
Aart de Geus – These changes will allow us to have further systematic collaboration, and to realize further integration through cooperation.
Per poll respondents, the challenges include: System-level software/hardware design, C-to-RTL synthesis, FinFET design automation at the physical level, Analog design automation, Software for power management, Low-power design, and Verification/Simulation.
Rich Valera asked panelists to pick one over-arching concern from the list.
Aart de Geus – Synopsys works on all of this!
Wally Rhines – The real question is, which area of concern would benefit most from a discontinuity in the technology? That might well be analog and mixed-signal design automation.
Of the respondents, 81% said there would be no IPOs in the next 12-18 months, with only 19% believing there was a possibility. Comments regarding the outlook for IPOs in EDA included: “Bleak”, “Extremely light”, “Very poor”, and “Not so good, as exits have been disappointing to the VCs”.
Rich Valera asked for feedback.
Raul Camposano – This may be true in North America, but don’t forget about opportunities for IPO-like business investments and growth in countries such as China and Chile. There you will find VCs still willing to invest in EDA.
Q1 – Why did ARM endorse the Cadence acquisition of Tensilica?
Simon Segars – The cost of innovation keeps going up. Tensilica is a good small company that’s been around for a long time. Their processors sit next to ARM processors on-chip, so the Cadence acquisition will benefit our mutual customers.
Q2 – Does Mentor get the last laugh over critics in the past who have been skeptical about Mentor’s investments in software for designing the harness for automotive electronics?
Wally Rhines – It did seem like a strange investment 20 years ago when we introduced the product. People asked, is it really EDA? However, we accelerated the investment 15 years ago which has paid off. We saw a 40-percent increase in revenues over the last quarter. But it is the type of investment you will only see if you live long enough. [laughter]
Q3 – What about the hypothesis that Intel will purchase Synopsys one of these days?
Aart de Geus – The economics of such an acquisition would not work. However, if anybody was willing to propose the kind of multiples we’ve seen in the Tensilica offer, we might be willing to entertain an offer. [laughter]
Q4 – What about IPO versus M&A?
Lip-Bu Tan – I’ve invested in over 100 companies that have gone IPO. But you need to remember, you celebrate an IPO for only a nano-second and then the pressure beings to builds. It’s perhaps better to build a company and sell it. Then you can relax and enjoy your success. [laughter]
Q5 – Will the Cloud change things for EDA?
Aart de Geus – In EDA, the answer is no, because there is no such thing as secure data.
Raul Camposano – However, the Cloud and the Internet of Things will change the world!
A final note of interest from tonight’s EDAC Panel. One of the slides displayed the companies mentioned by poll respondents when asked to name the “Hottest EDA Startup”.
* Most mentioned: Oasys Design Systems
* Other Notables: OneSpin Solutions, Berkeley Design Automation
* At least one Mention: ICScape, Jasper Design Automation, Calypto Design Systems, ATopTech, Forte Design Systems, Breker Verification Systems, DeFacTo Technologies, Blue Pearl Software
Tags: 14nm FinFET, Aart de Geus, Ansys, Apple, ARM, Cadence, Computer History Museum, DAC, DATE, DoubleTree, EDA Reunion, EUV, Google, Gordon Moore, Jasper Design Automation, Jim Hogan, Joe Costello, Kathryn Kranen, Lip-bu Tan, Mentor Graphics, MOD Marketing, Nimbic, OCP-IP, Raul Camposano, Rich Valera, Samsung, Simon Segars, Synopsys, Wally Rhines
2 Responses to “EDAC CEO Panel: Practically perfection”