Value creation while maximizing return on capital
01/01/2007
The semiconductor industry represents one of the miracles of the capitalist system, requiring steel-nerved risk-takers. Huge infusions of capital are needed to build the world’s most intricate and complex factories to make devices that will soon be outmoded in markets that often fluctuate wildly. Early juicy margins can be quickly eroded, particularly if challenged by copycat products at lower costs or even higher performance chips from competitors. Market windows are short, and holding or gaining share means continuously upgrading or even supplanting successful designs. Yet, in spite of the big gambles required, new fabs keep getting built, and even superfabs may be on the horizon.
The new fabs, now costing more than US$2 billion, either get subsidies and tax breaks from governments, or are built by huge chipmakers like Intel, Samsung, IBM, or Texas Instruments with stockpiles of cash. Risks of this magnitude are unsettling to bankers. Even the Chinese government backed off its grandiose plans for conquering the chip market once it became clear that while the thirst for cash is seemingly endless, making profits isn’t easy in this business.
Yet the money keeps rolling in as governments continue to pour major resources into supporting chipmaking or bringing it to their countries, as is now happening in India. The Information Age has been launched by the products of our industry, and a profusion of cool gadgets and systems-from iPods, pocket camera phones, and GPS locators to desktop supercomputers and the wireless internet-continue to emerge. Governments know that many jobs and infrastructure industries follow success in a major growth industry like chip manufacturing or electronics and they want a piece of the pie.
Still, in spite of a half-century of breakneck progress, the whole industry can’t quite keep up with Wal-Mart. While very successful at value creation, the industry has found it difficult to achieve high efficiency for invested capital, especially for tool and materials suppliers. One reason is rapid technological obsolescence. Steadily broadening the markets for electronics depends on lowering price/performance points at an exponential pace, and that’s really hard to do. Aside from rapid technological change, part of the problem appears to have been the insular approach of semiconductor fabs, with a tradition of solving problems secretly and keeping some process steps proprietary. Even such trivial areas as wafer marking and indexing schemes have boosted total industry costs unnecessarily. In spite of the tradition of secrecy, eventually most of the industry seems to migrate to a common optimal approach as secrets seem to seep out around the edges.
Complexity has grown so great that collaboration is steadily replacing the isolation of the past. National consortia, once dedicated to advancing only domestic agendas, now collaborate in many pre-competitive and standardization areas. In Japan, once-fierce rivals have joined together on major fab projects. A Fab Owners Association has recently emerged (Cupertino, CA) to help solve joint manufacturing problems and to work more cooperatively with tool and process developers. Groups of companies are coordinating process dynamics so that work can be transferred among fabs more easily, led particularly by IBM. Intel now can switch work from its busiest to less busy fabs, and Samsung can switch processes back and forth between DRAMs and flash memories, depending on demand. Toolmakers are making it much easier to get tools set up and on-line, with standardized ports and uniform gas and electrical connections.
While tough competition will remain, the trend toward cooperation on standards and uniformity in noncompetitive areas will continue in semiconductor manufacturing. Collaboration, cooperation, and more information sharing may replace much of the not-invented-here attitude of the past. Eliminating waste wherever possible is a key to maximizing return on capital, and the industry needs to work together to do that. That will mean even lower costs for electronic products, and we know how much value that can create! Watch out, Wal-Mart.
 |
Robert Haavind
Editorial Director