Semiconductor IP houses struggle to survive as ASIC design starts continue to dwindle
www.cmpnet.com The Technology Network |
Semiconductor IP houses struggle to survive as ASIC design starts continue to dwindle
Crista Souza
EBN
05/26/2003 10:00 AM EST
URL: http://www.my-esm.com/showArticle?articleID=10100072
The semiconductor IP business model has always been shaky at best. But having struggled for years to prove the value of IP, the question vendors now face is who's going to buy it if no one's building chips?
ASIC design starts have dwindled to a paltry number. iSuppli estimates that only 1,450 new commercial ASIC designs were undertaken last year, down from roughly 5,000 in 1998-99. And the trend is expected to continue well after the semiconductor industry recovers from its recession.
The reason? Design costs are becoming more prohibitive with each process generation, and companies are making hard choices about where to put their development dollars.
The result? For IP vendors, it's less opportunity.
MIPS Technologies suggested this recently when it decided to kill off its flagship 64-bit custom processor-core business to focus on synthesizable cores. The custom cores had more performance, more cachet, and a higher price tag, but MIPS couldn't sell enough of them to turn a profit.
The company characterized the move as a recognition that ASIC economics are changing. But the underlying message was that it's getting tougher for IP vendors to make a buck.
Despite being the second-largest processor-core supplier and one of the top IP vendors overall, MIPS is struggling to find the right formula, and you can bet it gets worse as you go down the line.
To be sure, MIPS and others, including rival ARM, can and do make a living licensing synthesizable IP cores. They cost less to support and are easier to design with.
But at the end of the day, the big payoff for IP is in high-volume chips that will pay royalties for many years. Such lucrative designs are becoming scarce and can take years to produce a steady revenue stream for an IP supplier.
MIPS, one analyst noted, hasn't scored a million-seller since the Nintendo 64 game console in the late '90s, and royalties from that design are tapering off.
ARM has remained dominant--and consistently profitable--by virtue of being the oldest player in the game. But with more than 75 semiconductor partners at last count, ARM is close to saturating its market. Once that happens, where will it go to find new business?
Maybe China's budding chip design community holds the key. Stay tuned on that front.
For now, consolidation of the supply base is certain to continue. Some of the top IP vendors could merge or be acquired. Then all the survivors need to do is figure out how to stay alive until the royalties start rolling in.
E-mail comments to Crista Souza at csouza@cmp.com.
www.cmpnet.com The Technology Network |
Copyright 2004 © CMP Media, LLC
Related Articles
New Articles
Most Popular
- System Verilog Assertions Simplified
- System Verilog Macro: A Powerful Feature for Design Verification Projects
- CANsec: Security for the Third Generation of the CAN Bus
- Memory Testing - An Insight into Algorithms and Self Repair Mechanism
- Last-Time Buy Notifications For Your ASICs? How To Make the Most of It
E-mail This Article | Printer-Friendly Page |