February 16, 2010
A look at EDA in 2010
Please note that contributed articles, blog entries, and comments posted on EDACafe.com are the views and opinion of the author and do not necessarily represent the views and opinions of the management and staff of Internet Business Systems and its subsidiary web-sites.
| by Gabe Moretti - Contributing Editor
Posted anew every four weeks or so, the EDA WEEKLY delivers to its readers information concerning the latest happenings in the EDA industry, covering vendors, products, finances and new developments. Frequently, feature articles on selected public or private EDA companies are presented. Brought to you by EDACafe.com. If we miss a story or subject that you feel deserves to be included, or you just want to suggest a future topic, please contact us! Questions? Feedback? Click here. Thank you!
The first thing I extracted from the responses I received to my request for a 2010 forecast from EDA companies is that no one tried to provide a financial forecast. Although I expected such an outcome, I was hoping to be wrong. The problem is that our planet is going through some very challenging times, both in the developed and the developing countries, although for different reasons. This makes forecasting financial results quite difficult since our small EDA industry can be significantly impacted by factors outside the electronics industry.
The Financial aspect
In the last two years the EDA industry has been severely impacted by the financial turmoil that hit the entire developed countries sector. According to data provided by Jay Vleeschhouwer, arguably the most senior financial analyst covering our industry, the total revenue of the industry has shrunk by about 20% from an actual in 2007 of $5.15 billion to an estimated $4.18 billion in the just concluded 2009 calendar year.
Another interesting number to look at is the growth experienced by EDA from 2005 to 2007. In that period the industry went from a total revenue of $4.09 billion in 2005 to the afore mentioned $5.15 billion in 2007. The previous growth and the later decline are almost symmetrical, and we are back to the 2005 revenue level but with a dollar that is worth less. Thus in the last four years our industry has seen no growth. Can we start growing again in 2010?
Probably a bit, but not by much, especially if one takes the projected inflation into account. The inflation should not be high, around 1.5%, and EDA vendors should be able to modify their prices to reflect that. On the positive side, as the contents of the responses show, some markets should show growth in revenue. On the negative side, if further consolidation in the semiconductor manufacturing sector occurs, revenue will be negatively impacted. Above all this, the availability of consumers discretionary funds, could cast a depressing shadow, lowering demand for electronics products.
The significant erosion of net worth, or as in the USA, perceived net worth based on past inflated real estate evaluations, has increased the savings rate of consumers who are now more likely to postpone discretionary spending than to respond to the marketing efforts of distributors of electronics gadgets. I personally, and I am sure most readers also do, know well-to-do persons who are "making do" with cell phone that are two years old, for example, and do not plan to replace them this year. The initial positive indicators following the Christmas shopping period indicated a willingness of consumers to spend. Unfortunately as the real final numbers become available, reality is
actually the opposite. Retailers did not realize the income they expected, and more consolidation in this sector is likely.
We cannot take Silicon Valley as an example of consumers appetites. What goes on in the Valley does not go on in the rest of the consuming world. The Chinese market is the major hope of electronics companies. Not just for the appetite of Chinese consumers for electronics gadgets, but also for the demand of products, like automobiles, that have a high electronics content. Of course on the positive side, if a Chinese person purchases a cell phone for the first time, he is likely to purchase the latest model. To a lesser extent this is also true for India, Brazil, and Russia. But China is now moving to curb inflation, by increasing bank reserve requirements and thus making loans,
especially consumers loans, harder to get.
My prediction for 2010 is that the EDA industry will end the year with revenue between $4.25 and $4.30 billion. On January 28 Jay Vleeschhouwer, a leading EDA analyst now at Ticonderoga Securities, forecasted a growth to between $4.27 and $4.33 billion. My forecast results were arrived at totally independently from Jay's, but the results are quite similar. This is either good news or an indicator that Jay and myself really want this industry to succeed.
Although modest, this growth will be significant in showing that the sector still has vitality and creativity, and that its management is competent. Should EDA fail to reach the lower of the estimate, it will result in significant consolidation in the industry. In order to stay in business companies will seek to merge or at least find new business alliances. The goal will be not just to shrink overhead expenses, but to leverage the value of individual products by packaging offerings from more than one company into a total flow that takes the user from architectural design to manufacturable silicon.
Potential EDA Growth Sectors
My analysis of the responses shows three areas of concentration: implementation choices in semiconductors, architectures, and methods. The use of FPGA devices continues to grow and shows no signs of flattening, the high cost of verification facilitates the growth of IP use and reuse. Power concerns aid in the use of multicore and thus architectural design, communication and graphics requirements demand investments in analog and RF, and the growing cost of development requires better methods and standards.
Semiconductors industry forecasts predict a 7% growth in revenue this year, driven mostly by computers and cell phone purchases. I believe that a major factor in the growth of computer sales, mainly laptops, will be the introduction of Windows 7 by Microsoft. This operating systems is much more user friendly than Vista and many people will want to switch and at the same time take the opportunity to upgrade their hardware platform.
Not much is new in cell phones, as the majority of users do not take full advantage of the G4 connectivity characteristics. But there is still an enormous growth potential for new customers in developing countries and certainly some users will upgrade older phones to newer ones, if for nothing else, peer pressure. These two market segments will require new designs and thus continual as well as additional investments in EDA tools.
The positive aspects are balanced by two negative factors in the semiconductors industry: consolidation and costs. A new 300 millimeters fab capable of supporting 45 or 32 nm processes costs a minimum of $6 billion, an enormous investment even for the largest of semiconductors companies. These costs have generated a movement toward consolidation in the industry. Globalfoundries is the best example of the consolidation. This consolidation impacts EDA vendors revenue in two ways: consolidation of engineering department that lowers the headcount, and less business units negotiating license agreements.
FPGA is a sector that has proven sustained growth in the past few years, and that growth will continue in 2010. Mentor Graphics is the company that has been the leader in providing independent EDA tools to the segment, although both Cadence and especially Synopsys are making inroads in this market. All three companies, in addition to much smaller specialty vendors, will continue to benefit from this growth. It would not be prudent, though, to consider the growth in this sector as a net growth of total EDA revenue, since a significant number of FPGA designs are direct replacement for designs that were done as custom ASIC in the past. In addition, may be unfortunately, FPGA tools
still have a lower license price than ASIC tools. This in spite of the complexity of designs now implemented in FPGA devices.
With larger chips comes increased architectural complexity. The use of multiple execution engines on one chip has now become a regular approach to increase execution speed, to partition power use, and to facilitate design reuse. The focus on consumers driven markets require vendors to both modify existing features, and add new ones without having to redesign an entire chip. Hitting a market window as it opens versus being even only a couple of months late, means making a profit versus loosing a significant amount of money. Thus a schedule slip due to design complexity means failure.
What has been labels the ESL market, and the IP market are therefore good candidates for growth. In the IP sector the growth will come from an increase in the use of existing cores, as well as the introduction and use of new IP products.
To foster real growth EDA vendors need to move away from just providing hardware modules for reuse, and begin to develop and sell both hardware/software modules and software only "cores", like full Real Time Operating Systems (RTOS). For too long has this industry been fixated with hardware only implementations: growth cannot be fueled by Moore's law anymore.
The traditional model of one central processing unit executing one software program stored in a separate memory moved from the IBM 360 and similar computers to silicon. Venerable computing engines like the Intel 8080, 8048, and 8086, or the Motorola 68000, underscored the separation between hardware and software while proving that roomful of electronics could be replaced by a chip.
But the world started changing as early as Intel's introduction of the 8051, which was no longer called a CPU but a microcontroller (MCU). The term firmware became popular, as computer scientists and software engineers learned to separate persistent executable code from code and data that could be modified during execution. A manufacturing industry building Read Only Memory (ROM) devices flourished in the seventies. Now, of course, one hardly finds a ROM device in a product, as that type of memory is fully contained in the die together with the execution engines and a good portion of traditional memory.
EDA vendors, thus, must adjust to provide all of the components that go into a chip, not just the hardware, or they will shortly become the providers of the smallest portion of the system. Firmware is not just as important as hardware, the system architecture must be designed by making hardware/software tradeoff at the early stages of architectural design.
The increase use of IP, especially third party cores, is fostering the growth of Virtual Prototyping, a market segment that Synopsys is taking aggressive steps in dominating not just with internal development, but also with its acquisitions of Virtio, VaST, and CoWare. Virtual Prototyping aids in the verification of not only the third party hardware, but also in the development and debug of firmware and the verification of the hardware/software integration.
In order to diminish design costs and improve simulation performance, almost all architectural design is being done using the C language or one of its derivatives, like C++. Eventually the designs must be translated into implementable RTL code, and this is also a market segment whose growth will continue.
Unfortunately licenses for tools in this segment are not as lucrative as those for physical implementation, an approach that EDA vendors are going to rue in the future. There have been, and unfortunately still are, too many "silicon heads" in charge of EDA companies who cannot let go of the glamour of fabricating extremely small devices on silicon. There is a reason for the birth of "silicon foundries" and the fact that most electronics products vendors have abandoned their internal fabs, and it is not only the cost of a new fab. It is, instead, the shift in importance from the capability of producing silicon to the ability of designing competitive systems. The time is not that
far away when Architectural Verification will be the last step in the process of passing a design from a system house (the eventual product vendor) to the implementation chain. The signs are already here. If you do not believe it just ask Jack Harding of eSilicon.
The World Is Not Digital
Design was relatively easy in the simplified world of digital design. Analog was a strange addition one needed only once in a while and was left to beings one would suspect of originating from another world. Then came mobile devices, computer controlled engines and other car and avionics systems, and we met the real world.
Messy, approximate yet precise, requiring solving physics equations in order to design a circuit. If this was not enough, enter the world of RF design. From cell phones to ubiquitous computing, to home networks, we live in the middle of a true electronic ocean. In this world design reuse is still in its infancy. Errors are more difficult to find, simulation of even a circuit that would be very small in digital terms take hours, and costs tend to grow rapidly. Although circuit simulation tools have made significant progress in the last fifteen years, this market segment offers many opportunities for expansion. From aids in circuit design to reuse, to better yet simulation. We
have had digital emulator and accelerators for many years, and we had analog computers in the past. Can we get analog accelerators and emulators?
Although the 2010 fiscal year does look promising for increasing the prices of licenses of existing products, it holds plenty of opportunities to expand the EDA industry with new products targeting markets not fully served or that are yet to be served. Consolidation, both within vendors and among customers will continue in an effort to manage costs, or simply to avoid going out of business. It is unlikely that a year from now we will hear from EDAC that the cumulative percentage of market share of the biggest three EDA vendors has dropped below 75%. Smaller companies will have to be even more creative in order to compete. Yet technological and business opportunities still make
this an exciting industry to be associated with.
You can find the full EDACafe event calendar here
To read more news, click here
-- Gabe Moretti, EDACafe.com Contributing Editor.