During 2000 most of the semiconductor industry saw spectacular growth. Looking forward, however, many raise the question of how much "juice" is left in this growth cycle. Will it slow, like other elements of our economy recently? As a foundry serving almost every segment of the semiconductor industry, UMC certainly sees 2001 as an "up" year, with continued advancements in the communications and consumer segments.
Driving much of the past growth for semiconductors was a highly successful and significant business model-fabless semiconductors. Today, we expect growth to come from the birth of a new business model-fabless ASICs-whereby systems companies will turn to fabless design houses for ASIC design through fabrication and packaging, as well as the continued expansion of the fabless semiconductor model.
These fabless companies are forming alliances with semiconductor foundries to ensure they get the required capacity to guarantee the production of a vast portfolio of devices. These agreements with foundries, along with rising market demand for production of ASICs, will drive continued growth of the semiconductor industry further into this year and beyond.
With an increasing number of companies employing this new approach, the fabless ASIC model is quickly becoming a significant trend in the semiconductor business. Not only is it highly viable in today's marketplace, but it will contribute significantly to meeting the increasing market demand for highly sophisticated ASICs that are quickly designed and manufactured.
A look back
To understand the evolution of the fabless ASIC company, we need to take a look at some history. In the 1980s many in our industry envisioned a highly integrated business model in which a company could build semiconductor fabrication facilities to obtain maximum profits. But not everyone found an automatic correlation between building fabs and profits. This realization led a small pool of companies to focus their early lives on design and marketing while leveraging outsourced fabrication.
Over time several fabless semiconductor companies flourished, due to their nimble approach and finely honed design skills, to become admired both for their product port- folios and the market capitalization they achieved.
In the 1990s the fabless model grew to become the preferred business model for many chip developers. The market capitalization of these companies now exceeds $190 billion. This fabless model, combined with the growth of strong, independent foundries, has created a vitality in the industry in which the laser-like focus of the design companies, along with the technology offered by the foundries, allows a "best-in-class" selection process to build a strong innovation engine for both players in the semiconductor industry.
Birth of a new model
This brings us to the recent birth of the fabless ASIC model. During 2000 we saw the beginnings of this model with the introduction of a number of promising businesses calling themselves fabless ASIC companies. As with the fabless semiconductors model, this fabless ASIC model relies on the best-of-class selection process-the best designers, the best foundry technology, the best tools, and the best intellectual property.
The difference, however, is that these fabless ASIC companies are not producing branded products. Instead, they are applying their skills to implement designs for system companies or smaller electronics companies that do not have the deep-submicron silicon engineering infrastructure.
While the current $25 billion ASIC market has been primarily driven by its highly integrated manufacturing and design infrastructure, these new fabless ASIC companies allow the industry to deliver a best-in-class choice directly to the system house.
Already, there are two noticeable trends in the fabless ASIC model. First, many of the early entrants have been successful design houses that have adopted this new business model and are now actually delivering completed units. The evolution to this model has allowed these newly positioned companies to scale their revenues, achieve higher levels of valuation, and maintain consistent growth.
Another trend is the use of pure-play foundries by traditional ASIC companies (or integrated device manufacturers). The largest ASIC companies are looking to foundries to improve their competitivenesswith advanced technology, while small and midsize companies are finding it harder to justify continued funding of advanced manufacturing facilities.
Our industry faces increasing competitiveness, shrinking market windows, and evaporating talent pools. This best-of-class approach to ASIC design and manufacturing results in more first-time-right silicon, faster time-to-market, and ultimately higher profitability.
Because fabless ASIC companies are not required to fund or dedicate precious resources to fabrication facilities, they can devote more time, attention, and resources to creating high-caliber talent pools with deeper system knowledge and acquiring world-class third-party EDA tools
The tables have turned from previous years where traditional ASIC companies were known for high-end foundry technology. This has now
become the pure-play semiconductor foundry's concern. Today's foundries are keenly focused on advanced process technology-conquering smaller process geometries including 0.15- and 0.13-micron (and soon 0.10-micron), as well as the creation of 300-mm fabs-and improving manufacturing efficiencies.
And foundries are aligning with key fabless ASIC companies to provide best-of-class turnkey design services and manufacturing for system companies.
Together, the best ASIC designers are teaming with the most advanced foundries and making this a promising year for the fabless ASIC manufacturers and the pure-play semiconductor foundries.
Jim Kupec is senior vice president of worldwide sales and marketing at Hsinchu, Taiwan-based UMC, and president of UMC (USA) in Sunnyvale, Calif.