Integration of DFM functions into place and route tools will disrupt the size and dynamics of the DFM market segment.
Development in EDA normally follows a sequential path that starts with research followed by invention. The next step, implementation is, in most cases, done by small startup companies who have the organizational flexibility to address a new market. In some cases, as in the case of DFM tools, new tool offerings create a new market segment. The last step in the evolution of a new application is integration. This is accomplished in one of two methods: either the tool class is integrated in the design flow, like what happened with synthesis for example, or is incorporated in existing tools.
Occurrences in the last few days indicate that DFM capabilities are being incorporated into place and route tools, thus eroding the size of the total available market (TAM) for pure DFM products.
ATopTech has just officially announced its existence and has, at the same time, released a place and route (P&R) product, Aprisa, that incorporates many of the functionalities found, until now, only in specialized DFM tools. The following day, Mentor announced that its Olympus-SoC P&R product also had incorporated traditional DFM capabilities.
The reason for such integration is obvious: it is far more efficient to deal with DFM issues while performing the P&R function, than to fix the problems crated by a less sophisticated tools afterwards using DFM tools. For sure neither of these two new products can handle all of the manufacturing issues found in the 65nm or more advanced processes. For example neither product deals with issues related to lithography. But, in this instance, how much longer do designers need to wait until established tools like Calibre will address those problems internally?
Given the low level of acquisition activity in EDA, what does this mean for the future financial health of DFM specific companies? it is almost assured that the size of their TAM is shrinking, just when competition among them is increasing. I suspect that anxiety levels are rising among the venture capital firms that have backed those promising DFM startups.
At the same time, not all is well for companies like ATopTech either. Their announcement had all the required pieces to instill confidence in their future health as a company: strong backers, a product with a track record of working silicon, and a corporate deal with Broadcom for multiple, multi-year licenses. It took only one day for Magma to announce a corporate deal with Broadcom for its own P&R product Talus. As Eric Thune, VP of Sales and Marketing at AtopTech, shared with me, Aprisa has been used in conjunction with established P&R tools, like Talus, to deal with some specific logic blocks, not necessarily the entire design. This is a limiting factor to the size of TAM for the company.
It is well known that about 80% of the EDA market is controlled by four companies. If their growth continues to erode the size of the remaining 20%, the number of venture capitalist firms willing to invest in new EDA startups will diminish even more. Is it time to consider a new model for research and invention in EDA?