CEO Perspective: IP to shrink and stretch (By Michael Gulett, President and CEO, ARC International)

The uncertain economy is creating both opportunity and challenge for intellectual-property providers. While designers will increasingly incorporate third-party silicon IP into their chips, the number of IP providers will decline. This apparent contradiction comes from the interaction of two trends: increasing chip integration and shrinking design cycles.

The trend toward higher levels of integration is driving the growth in demand for IP. To integrate what was once a board with a dozen chips into a monolithic piece of silicon, designers need the silicon IP for the technology those chips embodied. But that technology used to come from seven or eight suppliers. No single company today has all the IP needed for every chip design, nor can com panies invent or create all the missing IP themselves and still hit their time-to-market. Even the biggest company must acquire IP, and if the biggest does, so does everyone else.

The difficult economy is further increasing the need for IP. Companies are trimming their design teams, and those teams need to focus on the final product, not the individual pieces. As a result, they have more need to buy than to build those pieces. Silicon IP for standards like USB and wireless LAN, which are too complex for customers to create easily yet do not add uniqueness to the customer's design, are prime areas of opportunity for IP providers.

As development cycle times become shorter and shorter, however, design teams are finding they can't even spend the extra time it takes to find sources of IP. They are streamlining their list of suppliers to work with the vendors that provide the most help in reducing the design effort. The shrinking design cycles, then, are causing customers to demand more than IP fro m their suppliers; they want a complete system solution.

To meet this demand, the IP industry will shift. IP providers will need to become complete embedded-system solutions providers, offering a complete suite that includes not only silicon IP, but also software, operating systems, communications protocol stacks, development tools and the like if they are to survive. Further, IP providers will need to offer a breadth of IP. The more supported cores they can offer that are already proven to work together, the more valuable their cores are to customers who have no time to spare for integration. The industry needs to evolve from simply supplying IP to becoming complete embedded-system solutions providers.

This will force a consolidation of IP providers even as the demand for IP grows. Providers that offer point solutions aren't going to survive long-term. The market for an individual core is relatively small, so, unless their IP is the next, best rocket engine, the point-solution providers won' t be able to grow to offer more complete solutions. They will either be acquired or will fold and the industry will end up with fewer, larger providers.