Semiconductor Equipment and Materials Industry Forum Explores Options for Funding Large-Wafer Conversion; Preliminary Report on Findings of SEMI 300 mm Initiative Stresses Need for Greater Industry Cooperation on Costs and Timing Issues.

SAN JOSE, Calif.--(BUSINESS WIRE)--June 19, 1996--Both the circuit-manufacturing and the supply side of the global semiconductor industry are in agreement on the need to convert the production of computer chips to larger wafers to increase productivity.

They also agree that the conversion will be costly, with the original estimated price tag of $14 billion beginning to look conservative. Beyond that, however, the two sides appear to be divided on such fundamental issues as when the transition is needed and how development of the necessary technology will be funded.

This was the conclusion drawn yesterday when some 150 executives attended an all-day forum here called the SEMI Global 300 mm Initiative Executive Symposium, the 300 mm (about 12 inches) referring to the diameter of silicon wafers generally agreed to represent the industry's next increase from the 200 mm (about 8-inch) silicon wafers just now coming into full production. The meeting was called by Semiconductor Equipment and Materials International (SEMI), a global association representing suppliers to the industry that is attempting to facilitate the transition.

The main item on the agenda was a summary of the findings from extensive interviews the SEMI team has been conducting with executives from the major semiconductor companies, equipment manufacturers and silicon suppliers in most circuit-producing markets.

Typical of the mismatch in opinions on the transition came in response to the question, "What is the most critical element in the 300 mm migration?" Ron Horwath, 300 mm special program director for SEMI, said over 50 percent of the semiconductor producers identified "technology" as the most critical, with "equipment availability" coming in second at just over 40 percent. By contrast, the equipment manufacturers placed technology in a very low 5th position, calling "customer commitment" and "transition timing" the two most critical elements.

"Clearly, the supply side of the industry feels that the technology represents no stumbling blocks and they say the equipment can be available when commitments are forthcoming from the customer base," according to Horwath, who was selected by the other team members to present their findings. Generally, he said, the customer base seems to feel that many equipment suppliers are dragging their feet, "hesitant even to start a 300 mm program on their own" but the research to date suggests otherwise, since some 80 percent of the SEMI members interviewed said they have an active 300 mm equipment development team already in place. More than 50 percent said that they anticipate delivering 300 mm production-worthy equipment in 1997.

It is the extent to which equipment manufacturers will fund their own research and development (R&D) that represents a major stumbling block, Horwath said. "Our findings are that most of our members are willing to, or are, self-funding benchtop equipment development, at which point the product concept can be proven by customers. They feel, however, that they need a financial commitment from one or more customers to carry it further to the `Beta' development, which is close to the completed stage and thus far no semiconductor manufacturer, either independently or collectively through a consortium, has been willing to make such a commitment."

Another major cost factor is the price of 300 mm silicon wafers for the testing of new systems, which is now being quoted by materials producers at around $1,500 each, with anywhere from 100 up to 5,000 needed for test runs depending on the type of equipment or process under development. "Both of the major semiconductor consortia, SELETE in Japan and the U.S.-based counterpart, I300I (International 300 mm Initiative) are considering paying for these test wafers but no action has been taken," Horwath said.

SEMI acknowledged the need to develop an industry-wide strategy for dealing with the test wafer stumbling block which he said "is more threatening than it seems." He said the team's interviews indicated that the industry needs a minimum of 7,600 test wafers this year, but estimates that this will escalate to more than 42,000, a factor of six, in 1997. "Even if the per-wafer price were negotiated in volume down to $1,000, we are talking an industry-wide minimum investment of $42 million just to validate the equipment."

Some of the possible solutions to the problem raised during the meeting include: multiple use (sharing) of wafers on non-competing pieces of equipment; exploring with either equipment suppliers or the circuit manufacturers themselves the funding of a "wafer pool;" or jointly funding such a pool through the suppliers and the two primary consortia, I300I and SELETE. For now, Horwath said there should be continued dialogue on these options with both IC producers and suppliers.

As to the timing issues, Horwath said there continues to be some disagreement within the semiconductor community itself as to when the equipment and materials industry should be prepared for the transition.

"Early estimates had some producers saying they wanted to be in full production sometime in 1998. In our interviews with device manufacturers, up to six IC manufacturers are currently planning pilot line or low-volume factories in 1988, and several others indicated large-volume factory builds between 1999 and 2002."

He added, however, that answers to one interview question clearly reflected the current customer/supplier uncertainty over timing. The question was over when those being interviewed felt would be the right time for the transition. "One response from an IC manufacturer was, `When a complete tool set is available, factory builds will commence.' And from the equipment manufacturers we heard, `When a commitment or letter of intent to purchase a piece of equipment is obtained, development of a Beta-level tool will begin.'"

Horwath said the research phase of the SEMI 300 mm Initiative moves later this month to Japan for more in-depth interviews with producers in that market. Following that, he said, "SEMI hopes to facilitate the 300 mm conversion so that it can proceed at an affordable pace, with both sides assured of a reasonable and timely return on their investment."

Based in Mountain View, Calif., SEMI is an international trade association serving 1,700 companies participating in the $55 billion semiconductor and flat panel display equipment and materials markets. SEMI maintains offices in Austin, Beijing, Brussels, Hsinchu, Moscow, Tokyo, Seoul, Singapore, Boston and Washington D.C.

CONTACT: SEMI

Lisa Anderson, 415/940-6918

[e-mail]

or

Mathews & Clark

Walter Mathews, 408/736-1120

[e-mail]