May 15, 2008 – The acquisition of DFM firm Ponte Solutions by EDA firm Mentor Graphics not only takes one of the last pure DFM companies off the table, it also answers the question about what side of the chipmaking wall DFM belongs, and terminates an inflection point that might not be seen again for a decade or more.

“In a way this is the end of the DFM dream,” according to analyst Gary Smith, in an e-mail exchange. “When DFM first came on the scene there was a lot of uncertainty over where it belonged: Was this a semiconductor equipment play? Would the foundries control its destiny? Or did it belong in EDA?” And if EDA was the right answer, was DFM a “once every 10-15 year inflection point” that could change the EDA playing field? The latter seems to have come true, Smith noted, but the impact wasn’t enough to hoist a pure DFM firm into the top ranks of EDA companies. “DFM tools had to be integrated into the design flow which tied it back to the router, so it became a big company game,” he said.

Along with Blaze, Ponte was one of the last substandial DFM vendors standing, and “it was a matter of time before Ponte would be acquired,” Smith told WaferNEWS. [Ed.: This statement has been corrected to list Blaze as still a standalone DFM firm, not “bought out” through its merger with Aprio. We apologize for the error.] He cited the Mentor-Ponte combination as “a good fit,” though its technology is only “additive” to Mentor’s in-house capabilities.

More broadly from a strategic market perspective, buying Ponte cements Mentor’s claim as the top EDA/DFM firm, with access to both “keystone” DFM tools, Smith noted: its own DRC engine, and the IC Router which Synopsys/Magma/Cadence have (and Mentor obtained through last summer’s purchase of Sierra Design Automation, which debuted the first DFM router). Now, “chances are good that one of them will drop out,” Smith said of the three other EDA firms, with the survivors left to try to take marketshare in DRC. “I would say Mentor would have to screw up pretty badly to lose this one, and they haven’t been screwing up much this century,” he quipped.

This might be the end of consolidation in DFM/EDA (Blaze is now the “lone ranger” in the segment, noted Smith), but hardly the end of DFM technology innovation. “Each node brings new challenges, so we have a lot of work ahead of us” until the next projected inflection point ~2020, he pointed out — at which point bigger issues loom: whether CMOS, and possibly silicon itself, may have run out of steam. — J.M.