Mentor Graphics has purchased Sierra Design Automation and finally closed the "hole" that everyone pointed to when describing Mentor as a "point tools" company. In fact Mentor has not been a "point tools" company for quite some time with the exception of the IC design market where, admittedly, it did not have any place and route offerings.
The fact the Mentor found it necessary to spend $90 million dollars to bring such capability in house strongly reinforces a point of view that was initially mentioned by Aart De Geus at DATE three years ago. I may be paraphrasing it a bit but its substance is as follows:
The level of integration required to successfully develop and manufacture a design at 65nm or smaller geometries is so demanding that point tools companies, no matter how good their products, will find hard to achieve.
This remark, coming almost concurrently with pronouncements by all three big EDA companies stating that the level of acquisitions would decline significantly, lost most of its intended meaning and was understood just as another way Synopsys chose to signal its decreased appetite for startups. Instead Aart was making a profound technical observation, one that has found in me an enthusiastic supporter. I have repeated the same message since then many times in various environments, from panels to columns to technical articles.
Although in the case of Mentor it is possible for engineers to use Calibre even with the Synopsys, Cadence, or Magma flows, tools integration is only one part of the team effort required for success. Design teams need to integrate with teams from EDA companies, foundries, mask makers, and even test companies in order to understand the capabilities and shortcoming of the various suppliers in the flow. So, if you are an IC design house and use Calibre for your 65 or 45 nm designs, you are likely to need to bring Mentor people at your Cadence (or Synopsys or Magma) meeting. I am sure you know how this would not be a welcome outcome, especially to Synopsys and Cadence that compete with Mentor in many other markets.
The bad news for the point tools companies is that using their tools requires a greater effort on the part of the design team, since they will need to integrate both the tool and the support engineers into the process. This makes the team larger and thus complicates communication. Again this is not a welcome outcome on the part of the large EDA vendors because the team from the point tool company can take back information from one vendor and involuntarily communicate competitive information to another vendor while working hard to make the different customer design teams successful.
So, Mentor has finally entered the IC place and route business. I expect greater efforts on the part of at least Synopsys and Cadence to address the RET problem so that they can improve their account management before Mentor's sales force integrates effectively with Sierra's. The result will be a significant reshaping of the IC back end market. The final shape of the slices of the pie are difficult to predict today, but it is sure that the slices will look different, even at the 2008 DAC.