The EDA Consortium annual "CEO Forecast and Industry Vision" Jan. 5 brought forth a number of provocative questions about the future of EDA. A quick summary: Yes, the EDA industry is recovering nicely from the recession, but changes are needed so the industry can grow faster and realize more value.
The panel was moderated by blogger Dan Nenni of Silicon Valley Blog and the newly formed SemiWiki site (below, right). Panelists (left to right) were:
Jan. 5 also brought some good news from the EDA Consortium. According to their Market Statistics Survey, EDA revenue for Q3 of 2010 was up 11.9% on a year-to-year basis. We'll have to wait another quarter for the full-year 2010 report, but it looks like it will show substantial growth.
A quick summary of opening presentations follows, and then I'll move on to the questions.
The semiconductor industry grew over 30% in 2010. Current forecasts call for a "soft landing" with 6-10% semiconductor growth in 2011. "Getting mixed-signal right is the long pole in the tent as we enter the deep nanometer era. New EDA spending will be targeted at breaking that long pole."
We're out of the "abyss" of the recession and are returning to "predictable" EDA growth. Now we can turn our attention to new market opportunities opened up by consumer electronics, cloud computing, video, mobility, and low power. "We are very excited to witness the emergence of a new era of EDA we call EDA360. We will not only deal with silicon issues, but also the explosion of computing power and applications that will drive the success or failure of a product."
Result: "A new era, a new Renaissance, and new growth for the entire EDA industry."
Aart de Geus
EDA vendors are exploring complementary domains - including mechanical, hardware/software, and optical. Meanwhile, EDA has supported a growth in complexity that is "completely unmatched in the history of the world...going forward, EDA will drive an enormous growth in compute capabilities."
"Anything of value will get smarts. Your car, your body, your children, every appliance in your kitchen. It will be interesting when we start to compete with the IQ level of our toasters."
As he has in previous years, Rhines came armed with detailed charts showing how one year's semiconductor R&D spending correlates with the next year's EDA spending. This year he hedged his bets a bit by also considering analyst forecasts for Cadence, Mentor and Synopsys, who collectively hold around 72% of the EDA market. Conclusion: EDA growth between 8 and 14 percent. "This is not a forecast for Mentor, but I think it's a very good feeling for all of us."
We're too happy about 8-14% growth. The industry we serve is rebounding much faster. We're getting a declining piece of the pie with every node. How do we get a bigger piece?
Huang: Growth has to do with what value you bring. And, EDA didn't fall as much as the semiconductor industry in the recession. De Geus: Semi guys were down 30 percent, and equipment was down 65 percent. We've traded off some of the ups and downs for stability. The question is, is that a good tradeoff?
Does EDA grow by doing more of the same or doing something new? Do you see any new market opportunities?
Rhines: All EDA growth has come from new capabilities. Recent growth is due to design for manufacturing (DFM), electronic system level (ESL), new analysis tools, and formal verification.
We grow inorganically. We acquire. We're going to run out of companies to acquire. Emerging companies are a dying breed. How does EDA grow?
Huang: Yes, investment is down, but emerging companies can still succeed. Rhines: The EDA industry has gone from around 20 acquisitions per year to 15 - not zero. What has diminished is the number of venture-funded startups.
2010 was a year of acquisitions of IP companies (Denali, Virage). The EDA business model is different from the IP business model. Will the transition to IP ease the transition to adjacent industries?
De Geus: There are a lot of tradeoffs with royalty models. Moving to a different business model is "not a panacea."
How can semiconductor companies get more help with 3D IC challenges?
Rhines: All EDA companies are working to extend their tools - the question is whether customers are willing to pay the price. Huang: It's questionable whether 3D ICs will be a huge business for EDA, but it's a way to extend Moore's Law, and thus extend the growth of EDA.
How can startups get value for their products if big vendors are selling discounted tools?
Huang: Companies like Silicon Perspective Corp., Plato, and Denali [all acquired by Cadence] grew rapidly, showing that "it can be done." Big companies can't chase every idea - go for one we can't. Subramanian: "You need to have patience and a strong business strategy."
Questions raised at this event reflected the kind of discussion we should be having. How can we all work together to make the EDA industry stronger, larger, and more effective? How can we derive more value and recognition? That's what companies come to the EDA Consortium to discover. Then - off to compete for the benefit of our customers.
Photo by Joe Hupcey III
Thomas -- in my view there is much to be said for relative stability. EDA should grow by expanding its focus (as suggested by EDA360) rather than taking on a boom-and-bust business model.
>> De Geus: Semi guys were down 30 percent, and equipment was down 65 percent. We've traded off some of the ups and downs for stability. The question is, is that a good tradeoff?
Aart is absolutely right. The current EDA business and accounting models keep you very stable (especially no big downturns) compared to the IC industry.
Stability and slow growth vs. flexibility and chance for big growth (also big variations).
The EDA industry would have to provide more flexible models in order to have a chance to take a higher share in the IC industries fast recovery. But then only if you would also share some of the risks in the downturns.
There is little hear that has not been said several times in the past two decades. The industry is shrinking steadily, and in greater speed than the economic factors. There is less spent on marketing, less on development because there is less investment and less ROI in both investment and development. We are watching the neutrification of an industry.
That's all well and good and a little bit rosy... but I predict that EDA company headcount in the US will continue to decline as more EDA development jobs are sent overseas. EDA is definitely not a growth industry if you are a US-based engineer. Folks in the industry starting to look elsewhere as the handwriting is on the wall, especially as they see management&bean counters making more offshoring decisions - and having to deal with the aftermath of these "penny-wise and pound-foolish" actions.
Also, where's the innovation in EDA at this point? It's been in the small startups for years now, but as Rhines points out, there are less venture-funded EDA startups these days. The venture money is going to web, energy and healthcare related startups. EDA is a tough sell to VCs - they can get a much quicker return on their investment in a web startup. It can take many years of development for an EDA startup to get any revenue, let alone make a profit.