Fabless Flexibility: The Truth Be Told
Ron Norris -- Electronic News, 8/5/2002
As the semiconductor industry moves into a recovery mode, debate has risen over whether there will be sufficient capacity, particularly in the dedicated foundry industry, to accommodate the anticipated growth that will accompany a broad-based recovery.
A recent article in Electronic News (see "Fabless Foes Froth For Fight," June 10) postulated a skewed position that only integrated device manufactures (IDMs) can provide the level of supply assurance needed by OEMs and their customers so that they can meet their growing system-level market demand.
Will the fabless model make fabs such as Motorola’s MOS18 wafer fab in Dunfermline, Scotland obsolete?
While this may have been true a decade ago, the current state of the industry infrastructure suggests it is certainly not the case today.
The underlying premise that only IDMs who own 100 percent of their manufacturing capacity have the flexibility to accommodate sudden upturns ignores the even more fundamental truth that supply/demand flexibility is not inherent to one class of semiconductor company, be it fabless or IDM; rather it is the result of management planning and total corporate resources.
It instead postulates that some well-known major players can better leverage their considerable manufacturing clout to skirt the offerings of fabless companies which, they naively comment, may be stalled at the doors of soon-to-be overutilized foundries.
Amazingly, the sources for such unsupported predictions have conveniently overlooked a major shift in industry business models. For the past 10 years it has been the fabless semiconductor companies who have proven most nimble in meeting demand ups and downs—along with, more recently, many farsighted IDMs, such as ADI, Agere, LSI, Motorola, Philips Semiconductor and others who have begun capitalizing on the superior flexibility of the foundry model and are strategically outsourcing more and more of their manufacturing.
Third-party supply chain experts who have studied the semiconductor industry have found that it is not unique. In fact, they have found the manufacturing capacity requirements of the industry reflect a set of dynamics similar to those that also appear to drive other highly capital-intense industries, such as the airline industry. Those dynamics are high capital expenditure requirements and the ability to change on the fly.
The same experts have identified three characteristics that allow successful companies to capitalize on these dynamics. These characteristics are capacity magnitude and flexibility, a large customer base and a diversity of markets that permits a sustained and acceptable level of capital spending—even when one or more markets experience rapid change.
Given these three characteristics, the conclusion is self-evident. This is a perfect description of a well-run and profitable dedicated foundry; the primary example of which is Taiwan Semiconductor Manufacturing Co. (TSMC) At $2 billion-plus per fab, state-of-the-art-manufacturing facilities are now out of the reach of all but the very largest companies. Fewer still are the companies whose prudent management is willing to take on the risk, even if they can afford today's price tag.
From a manufacturing perspective, the 100 percent in-house IDM has but one customer—itself. But if it is spreading the risk by also dipping its toe into the foundry pond and offering capacity to others, economic self-interest dictates that when capacity gets tight, it is the other party who suffers.
It is this economic self-interest that drove the beginnings of the dedicated foundry industry in the first place.
Dedicated foundries, on the other hand, have become highly skilled at simultaneously balancing the production demands of a large diverse customer base. While it is recognized that the 2000 boom was partly the product of overestimated demand, foundry customers consistently received the product they needed, if not the capacity they said they wanted.
In this day of specialization, semiconductor companies large and small tend to focus on a few market segments. A well-run foundry, on the other hand, spreads it risks over a broad number of markets.
TSMC was the only dedicated foundry to show a profit throughout 2001. Its management often cites its customers' market breadth as one of the primary reasons.
The other increasingly important contribution that the leading foundries are bringing to the semiconductor market today is advanced process technology development.
At today's nanometer technology levels, leading edge process technology development requires significant and very focused resources. Commensurately, the levels of integration enabled by these technologies require a similarly increased and focused level of resources in the product development for the semiconductor product companies—the IDMs and the fabless product companies.
The product innovation and focus achieved by fabless semiconductor companies continues to pressure the IDMs to focus more on their product investments. This, in turn, will continue to pressure the ability of all but a very few (one, maybe two) to apply adequate levels of resource to their independent process technology developments.
As a result of all of these important changes in our industry, fabless and fab-lite companies have flourished, will continue to grow and will do so because they can meet their customers' demands better than the 100 percent in-house manufacturing-centric IDMs.
Had the capacity naysayers consulted systems-level customers, those customers would certainly have stepped forward and counted off the advantages of working with nimble fabless and fab-lite visionaries. From the systems companies' perspective, fabless and fab-lite IDMs would, in the final analysis, seem to be the preferred choice. Not only can they meet overheated demand, but they also offer innovative IC designs that meet their specialized product needs.
Without the overhead of fabs, and with their manufacturing partners providing industry-leading process technology and related services, these companies hold the distinct advantage of being able to concentrate their core competencies on proprietary designs. The ultimate net result is faster time-to-revenue and greater potential market-share gain for their customers.
The bottom-line is that there are no silver bullets. It is the ability of any given company to meet their customers' demands in a timely manner that will determine success or failure. The advantage, therefore, goes to the most customer-centric, flexible organization.
e-insite.net |
