NVM OTP NeoBit in Maxchip (180nm, 160nm, 150nm, 110nm, 90nm, 80nm)
Analyst seeks 'rethinking' of semiconductor IP
Analyst seeks 'rethinking' of semiconductor IP
By Richard Goering, EE Times
March 22, 2001 (5:12 a.m. EST)
URL: http://www.eetimes.com/story/OEG20010321S0038
SANTA CLARA, Calif. A "rethinking" of silicon intellectual property (IP) was presented at an IP/SoC conference business forum here by Eric Chen, senior analyst at J.P. Morgan H&Q. Chen said that commodity IP providers are doomed to failure, and that an emerging "domain IP" business model holds a lot of promise. A reaction panel of industry experts followed Chen's speech, and most agreed with the analyst's points. But some questioned Chen's gloomy view about commodity IP, and not all were comfortable with Chen's categorization of IP business models. Chen started his speech by stating that IP is a business model rather than an industry. "IP companies serve totally different markets with totally different growth paths," he said. "I would never upgrade or downgrade the entire sector." Wall Street is very skeptical about the IP business model, Chen said. However, he noted that IP company growth has far outpaced that of semiconductor vendor s since 1996. He said that the silicon IP business grew 36 percent between 1998 and 1999, and that "emerging companies" in the IP sector grew by as much as 150 percent. The IP market is attractive, Chen said, because it requires no inventory and no need for a lot of capital expenditure. And yet, he noted, IP companies exhibit high volatility, high return and high risk. In this respect, he said, the IP market bears many similarities to derivative securities. Because IP is an "information product," Chen said, it costs almost nothing to reproduce. If any information product becomes a commodity, its cost declines to zero or close to zero. Chen's conclusion: "If you're trying to be a semiconductor IP company, and you're selling a commodity product, you are going to fail." But there are some business models that can work, Chen said. One of these, "environmental" IP, addresses huge "cracks in the value chain" by solving urgent problems that can only be solved with IP. Chen identified Rambus Inc. and Num erical Technologies Inc. as environmental IP companies. It's a challenging business model, Chen said, because an extensive infrastructure is required. "It requires the entire industry to co-invest with the company," he said. But this creates value for those IP vendors who succeed, because the barrier to entry for potential competitors is very high. Chen's second category, "product" IP, addresses a more familiar design reuse market. Companies here include MIPS Technologies Inc. and ARM Ltd. Chen said there is a huge differential in valuation between programmable and nonprogrammable IP, with programmable providers enjoying five to ten times higher valuation. That's because an infrastructure with operating systems, compilers and debuggers must be created for programmable IP, creating a barrier to entry. "Domain" IP, the third category, is a new business model in which companies don't provide product-specific cores. Instead, they focus on a challenging technology area, like mixed-signal design, and p rovide customizable application-specific solutions in effect, platform-based SoCs that may comprise a number of IP blocks. The advantage, Chen said, is the ability to address many different, but related, markets. Chen identified Parthus Technologies plc as a promising domain IP provider. In the reaction panel, Bill McLean, president of Parthus' U.S. operations, welcomed Chen's comments. "Clearly there is a domain IP market," McLean said. "It's a very fast-growing marketplace, and we almost pioneered it. We elected to focus on the mobile Internet, and we take other IP and add more value." "I think application-specific IP is absolutely the new thing," said Jim Tully, chief analyst at Dataquest's semiconductor group. "I think we will see many commodity and star IP providers moving into the application-specific space." Tully agreed that it will be difficult for commodity IP providers to survive, but he said that it's still possible to succeed in that market if one gets there first. John Bourgoin, president and chief executive of MIPS Technologies, said that companies selling commodity IP can make decent livings, but won't have the returns of a "star" IP company. "Domain-specific IP is really just another way of talking about vertical IP," he added. Satya Simha, senior engineering manager in the information appliance group at Sun Microsystems Inc., noted that today's valuable IP is tomorrow's commodity IP. "I believe semiconductor vendors have to provide it [commodity IP] as part of their business models," he said. Tim O'Donnell, president of ARM Ltd. and also president of the Virtual Socket Interface Alliance (VSIA), took issue with Chen's comment that programmable IP always has a higher valuation than nonprogrammable IP. "The key for any IP vendor is to move into growing market areas," he said. "Programmability doesn't mean you will be successful." Michael Santarini and Rick Merritt contributed to this story
Related Articles
- Rethinking The Pursuit of Moore's Law
- Moore's Law, the bifurcation of the semiconductor industry and 3-D integration
- What's new in the Semiconductor IP market? The case of civilian Aerospace industry
- Easic's Zvi Or-Bach throws cold water on a semiconductor industry recovery
- The myth behind the semiconductor industry's growth-cycle guarantee
New Articles
- Quantum Readiness Considerations for Suppliers and Manufacturers
- A Rad Hard ASIC Design Approach: Triple Modular Redundancy (TMR)
- Early Interactive Short Isolation for Faster SoC Verification
- The Ideal Crypto Coprocessor with Root of Trust to Support Customer Complete Full Chip Evaluation: PUFcc gained SESIP and PSA Certified™ Level 3 RoT Component Certification
- Advanced Packaging and Chiplets Can Be for Everyone
Most Popular
- System Verilog Assertions Simplified
- System Verilog Macro: A Powerful Feature for Design Verification Projects
- UPF Constraint coding for SoC - A Case Study
- Dynamic Memory Allocation and Fragmentation in C and C++
- Enhancing VLSI Design Efficiency: Tackling Congestion and Shorts with Practical Approaches and PnR Tool (ICC2)
E-mail This Article | Printer-Friendly Page |