Since the week after DAC in San Diego, the top three EDA vendors have displayed an appetite for acquisitions. Mentor graphics opened the proceedings by purchasing Sierra Design Automation. Last week Cadence acquired Invarium and this week Synopsys acquired all of the semiconductor IP of MOSAID. It is not just that if you do not spend it you will eventually pay it in taxes. The reason for these purchases are to be found in the requirements of designers working on project targeting 65nm or smaller processes.
As I have said many times in the past, and to be sure I have not been alone in this position, the industry is moving from a tool chain problem to a supply chain problem. It takes a team that includes the designers, the suppliers of EDA tools, the foundry, the mask making shop, and the manufacturing test companies to work together seamlessly to minimize the probability of a re-spin in the ultra-deep-submicron sector.
Minimizing the number of EDA vendors on a team increases the quality of communication, at least in theory. So it is no surprise that first Mentor moved to close a major hole in its IC design flow. And last week Cadence purchased Invarium in order to significantly improve its interactions with the mask making shops by obtaining tools that allow it to closely control, and when possible optimize, the mask data generated from the design. Joe Sawicki, Vice President and General Manager, Mentor Design-to-Silicon division remarked: "For nearly a decade, starting with the 130nm node, lithography simulation has been a critical part of manufacturing flows at the fabs. Starting with the 65nm node, this technology has also begun to move upstream into the design cycle as part of a new DFM enabling flow. The Cadence acquisition of Invarium validates the importance of the strategy their competitors have had in place for a number of years."
Synopsys purchase of the MOSAID semiconductor IP is interesting because it gives it memory controller blocks for DesignWare. Since the amount of software, or firmware if you prefer, in advanced SoC continue to rise, memories are using an increasing percentage of the die area. So now, Synopsys customers can obtain the controllers from the same vendor that provides most of the other design tools.
Sanjay Srivastava, founder and CEO of Denali Software, is not worried about the purchase. When I asked him for his reaction, he told me that: "I believe it is further validation of our products and our strategy. DDR is clearly become immensely important for SOC designs. At the high-end customers need an integrated verification IP, configurable controller, and PHY, something which Denali has been promoting for quite some time. We have a large customers base of about 250 design wins to validate that
our products and our strategy is working well in serving our customers.
As far as increased competition with Synopsys goes, I believe it will only be better for the customers."
In fact I have more problems understanding the Synopsys acquisition than the others. Synopsys paid $15 million for the full rights to the SIP portfolio. Traditional business logic would say that they expect at generate at least $45million in revenue over the life of the product to recoup their investment when you take into account the cost of sales, the present value of the money spent, and the depreciation of the IP itself which can be rendered obsolete by a new memory architecture or a better implementation. I do not see them being successful against Denali outside of the Synopsys customer base, since Denali has a strong reputation and a fuller portfolio.