As companies struggle to get to the market sooner with a new generation of products, other companies are emerging that support them in a "faster, cheaper, better" approach to design. These design-oriented companies provide already-designed subsystems as intellectual property. While this is familiar for software, which is licensed to users, and is gaining familiarity for FPGA firmware, the idea has been extended to circuits and subsystems more generally.

Designers are faced with three design alternatives for realizing a design:

1. Design and build the subsystem in-house, from scratch. For the parts of a product that contribute significantly to its competitiveness, this is the preferred approach. Companies are in a market segment because they are good at what is distinctive to that segment.
   
2.   Buy the subsystem. Power supplies are the leading example. There is no point in re-inventing a supporting subsystem, especially if the product design team has insufficient expertise in designing such a subsystem. Unless a company is building highly complex products such as jet aircraft, buying subsystems has profitability limits; manufacturing profit margins are reduced by conceding it to subsystem suppliers.
   

3. A new alternative is to buy the design. This has the advantage of retaining manufacturing profit margins while reducing development risk and time-to-market. (However, saving development time, of itself, is not a sufficient competitive advantage as a long-term business strategy. See the article, "Idea-Driven Versus Time-Driven Competition")

IP companies exist to support product development by providing software, but hardware "IP" in the form of "designware" can also be supported, by providing designers a complete, manufacturing-ready documentation package for their subsystems. In effect, such companies become an extension of a product design group's engineering resource base. As companies mature, their design activity becomes focused on those aspects of a product design that carry their distinctive competencies to the market.

For instance, in 1950s oscilloscopes, Tektronix designed and built the whole instrument, down to the component level (plus unobtainable components, such as matched timing capacitors for sweep generators, CRTs, and also those distinctive-looking front-panel knobs). Nowadays, it is cheaper to let a tool-and-die house make the knobs and a high-volume power-supply manufacturer build the supplies. Building them in-house in lower volume incurs a disproportionate capital expenditure for machinery, factory floor space, and employees with skills not directly related to the company's business focus.

When this out-sourcing trend is extended to engineering, the return-on-investment is better using "farmed out" subsystem design, especially if it already exists in exact or similar form to what is needed. By acquiring parts of the design through purchase of a one-time license (as software is), the subsystem technology is copied from the design-supplier's technology base to the design group's base.

In-house technology-base building through licensed transfer (or even buyout, if it adds significantly to a product's competitiveness) is an emerging trend in the evolution of design activity in the electronics industry. When confronted with a build-or-buy decision, don't forget the new alternative: buy the design.

Ó Dennis L. Feucht, 2000