Experts At The Table: IP Business Models
System-Level Design sat down to talk about what makes a company successful in the IP market with John Koeter, vice president of marketing for IP and systems at Synopsys; Art Swift, vice president of marketing and business development at MIPS Technologies; Vishal Kapoor, vice president of product management for SoC realization at Cadence; and Charlie Cheng, CEO of Kilopass. What follows are excerpts of that discussion.
SLD: How effective is the ecosystem in dealing with IP issues such as context and integration?
Cheng: It’s up to the IP companies to build the ecosystem, or at least a large part of it. We all believe what we are doing is very important to our customers and this is a piece of it. Our job is to add enough value so that instead of just getting back a third of one percent we get 1.5%. That’s a huge difference if it’s a $300 billion market. That’s a $4.5 billion business.
Koeter: Our best estimate is that only about 25% of the potential IP market is realized through outsourcing. So for every $1 you capture through the merchant IP market there is $3 from companies doing it internally. If the IP market was fully outsourced, like the EDA market is largely today, it could be a $4 billion market. It’s very common today to have $1 million in merchant IP on a chip today.
Cheng: So the big question is how you get the industry to go from $2 billion to $4 billion. Quality outsourcing of standard IP is one approach. Having unique, patentable IP is another. We’re all trying to add enough value to say a piece of IP commands the price.
Swift: I see a strengths and weaknesses in the ecosystem model. On the strength side, the ecosystem is incredibly adept at going after volume. If you look at the Android platform there are thousands of applications, hundreds of developers, and everything associated with it. It’s in tables, cell phones, set-top boxes, TVs—everywhere. People saw the potential to re-use that technology. Where the ecosystem is less adept is when it comes to things that aren’t so clear. Although many-core is here, there are not many tools to aid the use of multicore. It’s been an underserved market. If there’s a new trend that’s surfacing the ecosystem might not react quickly enough and we have to fund the ecosystem. Another weakness is that while we all work together on a friendly basis, we don’t really work that closely together. We’re not pre-integrating solutions. We don’t have joint teams on accounts. There isn’t incredibly close cooperation between IP vendors. We could do a better job of that as an industry.
Kapoor: I have never like the word ecosystem. It’s a loosely defined term that includes anyone in a market. What we’re looking at is a set of alliances that bring value to the customer, and there are certain things you can do to drive that. You can add standards, for example. It has to result in a better offer for the customer, and that doesn’t happen proactively. With Android it comes down to which distribution you’re using and how tightly coupled that is to hardware. Standardization could help bring those pieces together. There are two types of standards. One is the standard around a particular vertical. PCI Express is an example of that. When you’re outsourcing you’re lowering the risk. The challenge is to make sure you’re not lowering the value, as well. The other is when you try to bring the hardware guys, the tool developers and the distribution guys together with a de facto standard and work toward a more complete standard. If you don’t have open dialog and there’s skepticism about who gets what part of the market, that’s not going to happen.
Koeter: We couldn’t run an IP business without an ecosystem. On a daily basis I’m talking to our foundry partners and software partners about codecs for our ARC processor, other companies for our USB stacks, design services companies and even other IP companies where we are helping them model their technology and port it into our system-level solutions. But sometimes you get to a different value proposition with your ecosystem. The foundries care about volume. Most IP companies care about design starts. Processors may be the exception of that because there’s such a substantial royalty component. Occasionally we run counter to our partners, but it’s not something we haven’t been able to work out.
Kapoor: When it gets down to complementary IP or broader ecosystem partners, that’s where the opportunity is.
Swift: There’s also a contrast between standards and community. It’s great to put standards in place but equally important is the community driving to get products into the marketplace and working collaboratively to do that. In our current business we rely on standards, but we also depend on that community. The virtuous cycle is more developers, more software, more developers and more software.
Koeter: Functional standards are well expressed and really important. That includes USB and H.264. Those are well established and well known. It’s the same for IEEE 1500 in the test space. What’s not well expressed is any kind of integration standard. All the time at industry events we hear about the need for IP integration standards. Right now there aren’t any.
SLD: Is the money in the IP or the tools that go around that IP?
Koeter: Our IP business is completely separated from the tools business. Each needs to stand on its own merits.
Kapoor: Absolutely.
SLD: If you have a whole suite of things you can price some parts differently to provide a whole solution, right?
Kapoor: These are independent. But integration is also an opportunity for the EDA industry. The way customers have managed that is either through over-design. Beyond that there isn’t a way that says this is how two pieces of IP will work together. I don’t know how you do that without agreeing to a functional standard.
SLD: How will the IP supplier market change in the future?
Koeter: There will be two camps of IP suppliers. One will be the broad IP suppliers that can provide a broad line of IP. These companies will have IP revenues of $80 million to $100 million and up. MIPS is in that category. There also will be another class at $10 million or below that has a niche in which they’re very successful. True Circuits is an example of that. Kilopass is another example. It’s that ‘no-man’s land’ between $15 million and $100 million that isn’t sustainable. You’re not big enough to have the scale to invest in quality, but you’re too big to have a specialty.
Cheng: I disagree with that. If you look at the cost of running an IP business, there’s not a disproportionate increase in what you call ‘no man’s land.’ The cost is prohibitive in the start, and it gets progressively easier as the business gets bigger. Your operations team to collect and enforce royalty is a pretty big part of your operating expenses at $5 million. It becomes manageable when you’re $50 million and miniscule when you’re $500 million. Quality is proportional to the business that you develop. Where it gets hard is that at $50 million you have to find an exit strategy for your investors. You’re big enough to go public but not big enough to sustain it. One great example is Virage Logic, which had a single focus on SRAMs. They went public on that, but they couldn’t grow and diversify for awhile. If you have a single product line, at $50 million you’re very susceptible to back-ordering. That’s why companies are adding second product lines.
Swift: But if you’re in the business for awhile and you’re shipping 600 million or 700 million units and you’re collecting royalties, that gives you a fantastic revenue base on which to build. The licensing revenue can fund day-to-day R&D. And then you have this funding engine underneath that which is almost 100% profit. It would be really difficult to start a processor IP company from scratch today. I also think that the business model being so highly leveraged and profitable is subject to consolidation. We’ve seen it with what Synopsys has been doing of late. It’s inevitable there will be consolidation in this industry.
Cheng: I agree about the consolidation. MIPS has been going for two decades, so that’s a different story. But very few companies with expertise in one area can diversify with enough expertise to escape consolidation. It’s very hard.
Kapoor: There always will be specialty players. There also will be large companies that look to consolidate because it adds efficiency. But can you get to that escape velocity as a one-trick pony? Most companies are one-trick ponies, and that’s not a bad thing. But to get to a big enough business is a challenge.
Tags: Cadence, Kilopass, MIPS Technologies, Synopsys