A Different Kind Of Sales Job

Why it’s no longer imperative to do a hard sell for advanced tools—and where to steer the conversation.

popularity

For years EDA companies have been pitching the benefits of ESL tools and high-level modeling, and for just as many years chipmakers have been throwing that idea back at tools vendors saying it’s too expensive to retool.

Talks between the two have waffled between polite bickering and contentious negotiations. Chipmakers have approached these “talks” from the stronger bargaining position, playing tools vendors off against one another by blaming them for not sufficiently reducing their time to market, for charging too much money, and for not investing in the most advanced problems.

The tenor of those negotiations is changing, though, because the terra firma beneath the feet of chipmakers is no longer so firm. There are several trends underway that bear watching individually, but they also have to be viewed together to get a complete picture. First, the trends:

Trend No. 1: Complexity is increasing and costs of developing chips is rising. You’d have to have been hiding in a cave to have missed this progression, which began back at 45nm and is exploding at the latest process node. It’s gotten so difficult, in fact, that some chipmakers have rolled back plans to migrate to finFETs and returned to 20nm. It takes more time and money to develop a complex chip at advanced nodes, and it’s only going to get more expensive.

Trend No. 2: IP is making it quicker to assemble an SoC, but it doesn’t always improve time on the back end for verifying that everything works together. This is compounded by the fact that most chipmakers want to mix internally developed IP that was created for a completely different process node or chip in with other commercially available IP. Adding software creates some programmability advantages, but it also makes it harder to bound the verification problem.

Trend No. 3: Internal budgets at chipmakers aren’t increasing—or at least not without a lot of screaming, and usually not for long enough—which means they have to accomplish much more with the same. The result is that they now have to rethink their entire operation to figure out where the fat is, where the holes are, and how to reapportion resources within their own internal design flow. If you think about the extreme pain that large corporations experienced prior to 2000 in shifting to enterprise resource management applications from SAP and Oracle, that pales in comparison to redefining a complex SoC flow.

Put these three factors together and you have the makings for a completely different kind of sales discussion. There will still be comparisons between one vendor’s tools and another’s, and there will still be some positioning statements to wade through that trumpet what the business is worth. But the reality is the most of the advanced chipmakers are in trouble. They’re staring at a progression to 10nm and beyond that would make anyone think about early retirement. They need help in the way of consulting, they need a better understanding about how and when to use tools, they need training on those tools, and they need to understand how they’re going to deliver on promises to their CFOs and investors in an increasingly difficult market.

For all the talk about squeezing tools vendors, this has suddenly become a seller’s market. But in order to make that work best, the conversation has to focus on the bigger picture—and it has to involve more people. The sale is no longer just about tools. It’s about economics. It’s about strategies for where to cut costs and how best to do that, and how to build flexibility into models. It’s also about raising the abstraction level where necessary, reusing testbenches, models, IP and strategies where it makes sense, and reducing effort and investment where it doesn’t.

For the first time in many years, this is a conversation between equals in an industry filled with brilliant minds and a renewed imperative to solve some very difficult problems together. It’s a very different conversation, and if it drifts back to the price of the tools it’s time to either change the direction or to talk to someone else within the organization.



Leave a Reply


(Note: This name will be displayed publicly)