Between The Lines

It’s been quite a while since a product release has caused such a big reaction in the EDA industry. I’m sure all the big EDA players offering tools in the place and route world have busy on their phones today speaking with financial analysts, whom are wondering what EDA place and route startup’s ATopTech’s announcement means?

Early this morning ATopTech announced a triple whammy: they officially announced their company, their floorplanning to GDSII tool suite, and to top it off that they won a contract at big customer Broadcom.

This story started to get legs a few months ago, when John Cooley interviewed then Cadence executive Eric Filseth. The interview essentially claimed that there were some shenanigans going on in place and route benchmarks but didn’t say at which customer. The interview led to a string of speculations on Cooley’s site where at first ATopTech was claimed the winner of a benchmark at Broadcom, followed by other postings suggesting that Magma and Synopsys won the Broadcom benchmark. Wall Street analysts had been following the chatter and questioned EDA executives on their respective earnings calls last quarter.

But it wasn’t until today’s release by ATopTech that at least some light had been shed on who really won the Broadcom benchmark and who didn't?

ATopTech officials refuse to put a real dollar amount on the win other than to say it is a multi-year, multi-million per year deal. This of course opens the terms of the deal up to rampant speculation from industry watchers. The ATopTech tool lists at $750,000 so that could mean the deal could be worth as little as three licenses for each year of a 2-year subscription or TAM, or it could mean its hundreds of licenses for tens or hundreds of millions. ATopTech is a privately held company so they don’t have to disclose financials. Certainly, ATopTech officials expected competitors would downplay the deal. To ATopTech’s credit, its officials were very forward in admitting that the deal does not displace the incumbent tool. The incumbent is Magma.

It may be coincidence (the timing sure is interesting), but both Magma and Mentor also issued press releases on the P&R front today. At 6 AM Eastern, Magma announced its new Talus is being used at Broadcom, seemingly to squelch speculation that it has been displaced by ATopTech, meanwhile, at Noon Pacific, Mentor announced upgrades to the Olympus P&R tool suite, which it gained in its acquisition of Sierra Design earlier this year. Maybe the releases are coincidence but maybe not. Either way, it illustrates that P&R competition is heating up.

Of all the tool categories in EDA, digtial IC place and route is probably proven the most lucrative. P&R tools have a high price tag and, if you have the best one, you can sell a lot of them. Cadence, Avanti, Magma are all companies that came to prominence largely on the backs of their place and route offerings. P&R has also proven to be a strong foundation for a company trying to go public and build the next great EDA empire. And of course, today, there are more choices than ever when it comes to P&R tools, as Cadence, Synopsys, Magma, Mentor in addition to startups ATopTech and Pixis all battle it out for customers moving to 65nm and 45nm processes.

But certainly as the competition heats up, there’s several things to keep in mind. Customers tend to use multiple tools. It wouldn’t be surprising if Broadcom uses several different implementation tool suites. It wouldn’t be surprising even if individual groups at Broadcom use several different implementation tool suites (if they have the budget, why not use em?). Some tools are faster in some functions, some are more accurate in some functions. The big questions are what are the winning tools being used for exactly, by how many people, who didn’t make the cut, and more importantly, WHY didn’t they make the cut? Also, keep in mind, some of these tools are works in progress so vendors that didn't make the cut last week, can to some extent make it next week. Ultimately, though, the competition is great for users.

Whether it will be good for EDA vendors, remains a question?