Yields on joint ventures

Electronic News, July 27, 1992 by Jack Robertson

The spate of global semiconductor joint ventures is both good news and bad news.

First, the bad news: the alliances to develop, and possibly produce, next-generation devices could mean fewer customers for new semiconductor gear.

The high cost of that sophisticated equipment is a major factor in driving the makers into consolidating their development and possible future production. In turn that ratchets the equipment industry even harder: facing possibly even lower economies of scale for machinery that already is at minuscule quantity levels, it could be in a tighter price/profit squeeze.

The National Academy of Science's National Research Council also just released a report warning that the spread of U.S. semiconductor alliances with Japan could be bad news for the country--hardly words one would expect from the U.S. government. After reviewing a decade of deals, the report claimed it "documents that the prevailing flow of semiconductor technology through alliances is from the U.S. to Japan. A largely one-sided outflow of technology from the U.S. to Japan, if continued over the 1990s, could have the cumulative effect of eroding the foundations of America's capacity to innovate in this industry."

The U.S. government body cited the experience of TriQuint Semiconductor, which explored a joint venture in Japan to help market its gallium arsenide integrated circuits. "Potential partners insisted on an extensive license (to make the semiconductors)," the report stated. "TriQuint regarded the price as being too high if technology is to be granted in exchange for capital, so no alliances have been formed to date."

The "Japan That Can Say No" might have a different perspective on the semiconductor deals. Regardless of viewpoint, however, it is axiomatic that any joint venture should be a win-win arrangement for both sides. In the real world of global competition, however, that ideal isn't always achieved--as many companies on both sides of the Pacific may try to take advantage of each other.

The U.S. semiconductor firms themselves, basking in their recent high profile deals, still face plenty of uncertainties ahead.

The DRAM partners aren't sure at this point how fast new devices will repay the projected gigabuck investment costs. The DRAM joint developments especially are a sporty bet--as the market for each new generation seems to be far slower coming than anyone ever expected. If the 64M and 256M markets are as tardy arriving as their 4M predecessor, the payback for huge investments that must start in the near term could be long in coming.

The rush into flash EpROM joint ventures has nearly as much unpredictability. The newcomers seem to be betting on a potential massive market when flash devices might replace hard disk computer storage units. But continuing hard disk state-of-the-art advances could also push the hoped-for flash bonanza into a more distant future.

The good news may be that the new digital Advanced Television (formerly HDTV) technology and new generation computer graphics/multimedia systems--all potentially big memory hogs--could hit the market concurrently with the multimegabit memory chips.

All this impacts commodity semiconductors, always the most volatile market segment. The outlook for logic, ASIC, custom and specialized devices still includes some high risks, but possibly better profit niches.

But then the rush into global partnerships has largely bypassed these semiconductor types--further evidence that makers may not feel impelled to share the grief and risks by banding together on these devices.

Japan is hoping that the bevy of new joint ventures signed with American firms will be good news in easing semiconductor trade friction with the U.S. Now that it is almost certain Japan will fail to reach the 20 percent foreign market share target by the end of the year, that country is hoping that the influx of cooperative semiconductor deals will ward off some of the growing criticism.

Any such deals could be misleading, however, if eventual production ends up largely in Japan and with captive U.S. producers. More details must be spelled out on these transpacific joint ventures before they vent the pressure on Japan for greater foreign semiconductor market share.

One also wonders how much the transpacific joint venture bandwagon is being propelled by the bad market news out of Japan and the better business news on the U.S. front. Of course, longterm business alliances aren't spawned by temporary market conditions. Still, the latest plunge in the once vaunted Japanese semiconductor market may have started people thinking.

DRAMs, the linchpin of Japanese producers, offer about the lowest profit return on silicon of any semiconductor device. Facing staggering new development and equipment costs with no immediate big market upturn, it may be prudent to relax the "go-it-alone" strategy to join hands with safe American partners.

Finally, the burgeoning alliances and their high price tag may be bad news for the struggling U.S. efforts to forge national industry-government initiatives, such as Micro Tech 2000 or the similar Semiconductor Research Corp. technology road map. If the major players are already lined up with Big Money commitments with Japanese partners in similar quarter-micron line width developments, is there money and energy left for American initiatives?