Boston College Intellectual Property & Technology Forum January 20, 1998 SYSTEMS-ON-A-CHIP: INTELLECTUAL PROPERTY AND LICENSING ISSUES Fred M. Greguras ;* Copyright © 1998 Boston College Intellectual Property & Technology Forum, Fred M. Greguras I. INTRODUCTION There is an accelerating trend in the electronics industry toward implementing an entire electronic system on a single chip through the integration of multiple, reusable, virtual components including both digital and analog circuitry. These systems perform specific functions (i.e. digital signal processor graphics controllers) and are sometimes interchangeably referred to as intellectual property (“ip cores embedded” or “ip building blocks”). This trend toward such semiconductor systems has important licensing implications. Because of widely adopted industry standards, standardized physical components can be rapidly and easily mixed on a printed circuit board (“PCB”) to create an electronic system of multiple discrete integrated circuits. Each discrete device or component in the system can be purchased from separate vendors which own or have rights to the intellectual property in the device. The best technology, both digital and analog, can usually be implemented because discrete components can be easily purchased from different vendors. Current deep-submicron semiconductor technology makes dense integrated circuits feasible enabling the design and production of application-specific integrated circuits (“ASICs”) that combine components not on a circuit board, but within a single chip. Systems-on-a-chip are built of standard virtual components, combined with custom logic. Standardization occurs at the sub-chip level and customization at the device level. The combination of virtual components enables high quality electronic systems to be built faster and at a lower cost than products containing discrete physical components. Standardization is still developing and the goal of the growing Virtual Socket Interface Alliance is to establish technical standards to enable the mix and match of virtual components. II. STRATEGIC IMPORTANCE OF INTELLECTUAL PROPERTY RIGHTS The primary United States intellectual property rights in chips arise under patent law and the Semiconductor Chip Protection Act (the “Chip Act”). Copyright law is also applicable for software on a chip. Patent and copyright protection are available on a worldwide basis while Chip Act-type protection is not as widely available. Patent law protects elements of chip designs and production processes that are useful, novel and non-obvious. The Chip Act essentially protects the topology of original mask works fixed on a chip. Patent protection is the strongest type of protection, but is more difficult and expensive to implement. The party who owns or acquires, or has license rights for, worldwide Intellectual property rights related to a system-on-a-chip is in the strongest position in the marketplace, since that party should be able to offer the lowest cost device. The party with the most valuable intellectual property assets (including testing tools) is likely to be able to command the best terms in the licensing arrangements necessary to produce a system-on-a-chip. The more value the intellectual property has, the more protection it demands, because of the competitive advantage in a vendor's business strategy. Ownership of intellectual property rights in a system-on-a-chip can be complicated. A single chip is likely to require the intellectual property of multiple owners. A “system” will usually need both digital and analog circuitry, which may be owned by separate intellectual property owners. Traditionally, merchant chip companies have been in different functional or application market segments. Some produce microprocessors and RAM (random access memory), and others produce mass storage controllers, graphics and ASICs, or application specific standard products (“ASSPs”). Some companies offer only digital devices, while others offer only analog devices. Most companies will not have intellectual property ownership or licensing rights to incorporate all of these disparate functions. Intellectual property provides the best opportunity for differentiation in the marketplace because the owner may exclude others and use some rights to bargain for other needed rights. At some point, a vendor may not implement the next stage of integration on a chip because of the lack of such rights. Ninety percent of an integrated solution is not enough. An intellectual property portfolio provides leverage to obtain cross-licenses to obtain rights for such integration, and also encourages favorable settlements of disputes when the owner is allegedly infringing another party's rights. Obtaining such technology through outright acquisition of the company which owns the intellectual property or licensing of the many different intellectual property rights embodied in a highly integrated chip is difficult, particularly if each transaction must be negotiated separately. Variations in license terms and conditions among third party licensors increase the difficulty. The large chip manufacturers who own the intellectual property in the key component, the processor or controller, may have the economic power to achieve uniformity of such terms and conditions among their licensors or to acquire the owners of the intellectual property. License agreements must be carefully and precisely drafted in order to obtain all the required rights. Many existing license agreements may only authorize the licensee to produce a discrete chip with a single licensed function. Even if the intellectual property for a currently contemplated chip can be licensed from various third parties, it will be very difficult to expressly identify in a license grant all of the future reuses that may be needed. The scope of the license may not extend to reuse of the function as a virtual component in a system-on-a-chip. Under general principles of licensing law, any right not expressly granted to the licensee is reserved by the licensor. As a result, unless the licensee has been granted the express right to integrate a function into a system-on-a-chip, it probably does not have the right to do so under an existing license agreement. The right to sublicense, which is often only reluctantly granted, also must be expressly granted. Sublicense rights are important as a basis for cross-licensing to obtain technology from third parties needed to produce a system-on-a-chip. The express right to “have made,” found in many patent license agreements, is a right to subcontract for parts or components, and does not convey the right to sublicense. III. CONCLUSION Market forces will require the ever-increasing implementation of systems-on-a-chip. Price pressure and operating performance will likely continue to force even greater functional integration. The need for cheaper, smaller, and more powerful devices will continue. Technology advances enable the feasibility to increase the density of integrated circuits. Complex intellectual property and licensing issues must be addressed in order to meet market demands. Outright acquisitions, careful negotiation and drafting of license agreements and aggressive creation and protection of intellectual property are all crucial to staying competitive. * Fred M. Greguras is Senior Partner at Fenwick & West, LLP. 1998 B.C. Intell. Prop. & Tech. F. 12002