On behalf of the IP industry, I’m happy to paraphrase Mark Twain in that the news of our death as been greatly exaggerated.
Mark LaPedus of EETimes last week made a few provocative statements in his column last week Semi IP Sector is a Lost Cause calling the semiconductor IP sector a “non-profit business” and posted a follow-up challenge to companies to challenge his conclusion Letter to Semi IP vendors adding that it would be probably a better use of his time to improve his golf game than to write about the IP market.
Haughtiness, hyperbole, and selling magazines aside, Mark raises a number of good questions that are not obvious to people who loosely follow semiconductor design. Let’s take a look (Mark, grab a pencil):
1. Why are there so many IP companies, but only a few “winners”?
Fundamentally, the IP sector is extraordinarily fragmented and specialized. No one company can possess the engineering skills to develop every IP block needed for today’s SoC. Like many growing industries, acquisitions consolidate the winners leaving a variety of smaller companies either thriving in their niches or struggling to stay in business. Many of the companies are private and GAAP financials are not readily available for dissection by the media. Given that most IP companies are privately held, it’s a considerable disadvantage to those trying to figure out winners and losers.
2. Is IP quality an impossible nut to crack?
I personally think the “IP quality” issue that was a real problem 10 years ago is for the most part solved through the death of the companies that were contributing. Most of these companies crashed during the semiconductor downturn in the 1st half of this decade and only the horror stories remain. You just don’t hear about ARM, MIPS, and Synopsys having quality problems anymore. However, integration remains a key obstacle that many customers I talk to complain about.
3. What do shrinking geometries mean to the IP business?
It means that companies are going to need more IP to fill up those 45nm devices. At every process step since 0.35, IP use has increased. There are simply no indicators to the contrary that we are going to see a change. Soon, we are going to see a Moore’s Law like equation that predicts the amount of 3rd-party and internal IP that will be used in future designs. The data is out there.
I don’t think the IP market is particularly sexy or friendly for the media to write about. The boring truth is that it is the real growth area in semiconductor. I also think that we are in for some new IP business models and paradigm shifts in design as China and India start to become major players in semiconductor design (and not just manufacturing and software.)
If I’m wrong, I’ll probably be golfing with Mark and talking about whose chip is in the latest iPhone.
