What Would Joe Do?
Peggy Aycinena is a freelance journalist and Editor of EDA Confidential at www.aycinena.com. She can be reached at peggy at aycinena dot com.
IP Update: Savage @ IPextreme
March 8th, 2012 by Peggy Aycinena
Even this deep into the era of IP and design reuse, it’s been my impression that things are not quite as far along as many in the industry would like you to believe. With that attitude in mind, I spoke with IPextreme CEO Warren Savage, who convinced me otherwise, although with several caveats.
Q: Are people still reluctant to buy IP?
Warren Savage: That’s last century, and no longer a problem.
Q: How long has that been the case? 2 Years? 10 years?
Savage: I would say the pressure cooker that the semiconductor industry has gone through over the last decade has made that shift fully complete
Q: And is that true for all categories of IP?
Savage: I think so. We work with the large semiconductor companies with their internal IP. They’re definitely buying IP, although I’ve seen them creating IP when they need to differentiate around that block. We’re also seeing a shift in our business. Today we’re selling more differentiated IP, not commoditized IP.
Q: Can you define differentiated IP?
Savage: Things that typically have a smaller market, but have more specialized features. Sometimes semiconductor companies will hear from their customers that a certain kind of feature is needed, something that’s too specialized.
Sometimes those are early indicators of where the market’s moving, but the IP companies haven’t yet read the signs. In that case, the semiconductor companies have no choice but to develop the IP in-house.
Q: Isn’t that a case where an IP vendor could act as a design consultant?
Savage: Actually, no. In my experience, in the semiconductor companies it’s still an argument for internal development.
Q: If a company is developing internal IP, how often do they turn around and market it externally?
Savage: The IP business is pretty specialized. Making IP that other companies can actually use is not a core competency for a lot of the semiconductor companies – the development, documentation, and support.
Q: Regarding documentation, even if IP is not being developed for external sale, doesn’t an engineer’s manager require adequate documentation?
Savage: Yes, that’s very common, but the [bigger] objective is to get the chip out. The semiconductor companies tend to be very short-termed focused. It’s rarely about getting an IP platform that will last for 10 to 15 years. It’s about getting the chip into the market as fast as possible. If documentation doesn’t exist, but the chip works perfectly, that’s good enough.
The way things actually work in these large semiconductor companies, it’s a “casual” reuse of IP. Your manager says, “You go to talk to that guy about his IP, even if it’s not well documented.”
That’s where IPextreme can come in. We clean up the IP and document it, so the company doesn’t have to, but can still sell it to the outside market.
Q: Are you privy to proprietary information about internally developed IP at your semiconductor customers?
Savage: Yes, we assist in the development of those kinds of IP, and become licensing agents for the IP if our customer chooses to market it externally.
Q: Are there tools available to help people do a compare-and-contrast between similar IP offerings?
Savage: I don’t think there’s so much out there with regards to that. There are some web portals – ChipEstimate and Design & Reuse – and there are forums where you can essentially get the rough cut of features and capabilities.
What we usually run into, however, is a direct engagement with a customer simultaneously comparing different IP, the literature, and the collateral. The IP vendor sometimes even lets the customer have access to the IP, to have hands-on experience with it prior to purchase.
Q: Is there a market for such a compare-and-contrast tool?
Savage: Companies pop up regularly in this evaluation space, but personally I’ve never seen one that the customers really like.
Q: How about the royalty versus the one-time, purchase-the-block business model?
Savage: It really depends quite a lot on the IP itself. For the most part, it’s related to the value [in relation] to the customer specifications, and what the market will bear.
If there are a lot of competitors, you can only compete on price if you remove the royalty aspect. Fifty percent of our products have royalties, and 50% do not. I chair the GSA’s IP working group, and this is one of the questions we pose to IP suppliers and buyers. We actually track this issue, and can get a sense of how much is purchased on royalty, and how much is purchased under other business models.
Q: Are there expenses related to tracking usage in order to extract royalties?
Savage: It’s not difficult at all to track royalties, and the benefits are so great, it’s well worth it to keep track.
Q: Why not always go with the royalty model?
Savage: If you’re selling into a heavily competitive IP market, there will be competitors without royalties. And once that market has established itself, if it’s to not have royalties then you’ll have to find some other mechanism [for revenue].
Q: If you were advising somebody on whether they should go into the IP business, what would you say?
Savage: One of greatest problems in the IP industry is the lack of recurring revenue stream. My number one advice to our customers [who want to commercialize IP] is to come up with a longer term royalty model that will sustain you over the long term, as well as other kinds of revenue streams.
Q: So, is selling IP a product or services business?
Savage: The idea of IP as a product business, versus a service business, is the way the [third-party vendors] have moved in the last 10 years. There were other companies moving toward IP plus consulting people, but those have faded away.
Q: Are there percentage changes that determine if an IP block can be marketed as something entirely new, like in the publishing industry, which is around 30%?
Savage: No. The percentage of [how much of] the whole thing [has been changed] is irrelevant.
Q: Do I have an obligation to reveal to my other customers who is using my IP products?
Savage: Typically, the percentage of a chip design that contains IP represents a tiny fraction of the overall functionality. When there are two companies using the same IP, there’s some other kind of [mechanism used for] differentiation.
Just look at how many ARM licenses out there, and note that ST Ericson competes with Qualcomm, even though [they’re both using ARM cores].
Q: Are the foundries competing in unhealthy way?
Savage: Yes. Some are creating their own foundational libraries, basic gates, etc., while there are also traditional IP companies providing that. Typically, the foundries supply only the most basic types of IP, while relying on their third-party [vendors] for specialized versions of things.
Q: How do engineering managers partition a design to decide which part is IP, and which part of the design is not?
Savage: First they look at the block diagrams, and then the project manger and architect have to decide, which can they buy and which can’t they buy.
For the ones they can buy, they’ll price those parts out. For instance, we’ll get a call from that project manager and a request for a quote. Or, they’ll send us a spread sheet where they’ve indicated which parts they plan to buy, and which parts they believe are unavailable and will have to design themselves.
Maybe it’s a UART, or something else, that their in-house engineers can crank out in a couple of weeks, which would be cheaper than buying it from a third-party. Then they go out and survey the market again, to see what’s available and build a plan around that.
Sometimes, we help with that, or perhaps if they’re talking to our competitors, they’ll ask us: Do you know anyone with this particular IP? Can you make a recommendation? It’s just depends.
That’s the problem with the IP industry – it still defies categorization.
One Response to “IP Update: Savage @ IPextreme”