Will Japan's semiconductor industry recover?*
06/01/2002
Engineers at Japan's big chipmakers — including Hitachi, NEC, and Toshiba — are worried their companies might not make it through another year. While foreign competitors with higher margins, ample capital, and focused strategies are investing for the next generation, Japan's integrated electronics companies are hemorrhaging red ink, and some are dismantling their semiconductor divisions.
If these new businesses are actually given independent authority to make their own management decisions, to capitalize on their technology, and invest strategically as needed, then Japan's semiconductor industry will recover. But semiconductor divisions that do not escape from their parent companies' control will not survive.
We hear more and more depressing reports from engineers at Japan's big chipmakers. "It looks like two Tokyo chipmakers plan to spin off their chip development centers," notes one source. "Our company doesn't know what we should make," says another. "A year from now, our company may no longer exist," others tell us. Semiconductor divisions are piling up such huge losses, they're undermining the future of their corporate parents. Now China looks like it's becoming the world's low-cost semiconductor production base. Japan's semiconductor industry no longer seems to have much of a future.
Though the world has changed, Japanese management has not kept up. Most corporate groups got into the semiconductor business just because it was a growth industry they wanted to have covered within their family. But Japan's struggling banks can no longer keep handing over money to fund every group with its own entry in every field.
Japan hasn't been able to keep up with the competition in the current round of investment in next-generation production facilities. Thus, even if the semiconductor market recovers strongly later this year, Japan's big semiconductor makers will only continue to fall further behind.
The best hope for recovery for Japan's semiconductor industry is an emerging crop of more independent chipmakers with management control of their own business, whether within a larger corporation, spun off from the parent company, or established as a new venture. Obvious as it sounds, very few Japanese semiconductor producers have actually been able to decide their own independent strategy and investment plans.
Japan's big chipmakers have been talking about getting out of the commodity memory business for the last ten years, with new plans every time the market dives. In the early 1990s, the strategy was to develop the more profitable microcontroller business. In 1995, the plan was to target logic devices like media processors and graphics chips. In 1998, the answer was to focus on systems-on-a-chip.
Whenever the DRAM market picked back up, however, it looked like the best way to get immediate returns was to pour resources back into DRAMs again. New strategies competing for resources got pushed aside midstream. Dominated by parent company managers who don't understand the IC business, and continually competing with the DRAM business for funds, the integrated electronics companies' semiconductor units haven't been able to execute strategies of their own (see figure).
The problem is compounded by the low margins in Japan's semiconductor sector, even in the best of times. Profits are typically more than 20 points below those of leading chipmakers in other countries. In good times, semiconductors may generate modest earnings that make up for losses in the consumer electronics business at the Japanese parent, but in bad times, they generate massive losses. Foreign rivals make more money in good times and lose far less in bad times, so they can still come up with the money to invest strategically for the next cycle. In the down cycles in Japan, in contrast, when a Japanese chipmaker racks up big losses, its corporate parent steps in to fix the problems at the poorly performing unit and cuts back spending.
When top managers who don't understand the industry make capital investment decisions, they tend to invest when semiconductor sales are the strongest at the peak of the market. When all those new plants come on line 12-18 months later, there's excess capacity, prices fall, and then everyone cuts back capital spending.
"When the Japanese semiconductor makers start investing," says one analyst, "it's an indicator the market is going to turn down."
There is hope for the industry's recovery in the current crop of spin-offs, which may provide the opportunity to at last break free from the DRAM business and from parent company control. However, as sources note, "A partial spin-off like Elpida Memory may be fatal." Though this spin-off is nominally an independent company, no one there actually has power to make decisions. Many think such spin-offs will have to bring in outside or foreign managers to become independent.
Groups of engineering pros have broken the bonds and aim to run their own companies. These spin-offs can make decisions faster than the big companies, but unlike venture companies, they still have access to the big company's resources, especially its people. NEC Compound Semiconductor Devices, for example, spun off from NEC, and is developing its own non-Japanese-style management to capitalize on its technology. Fujitsu Digital Technologies, a group of chip design specialists, is also seeing some success.
Revival will also require making better use of the country's engineering talent. The Ministry of Economics, Trade, and Industry (METI) is working on ways to make it easier for people to move from one company to another. — Yosuke Mochizuki, editor in chief, Nikkei Microdevices
*This article was translated for Solid State Technology by Paula Doe from the January 2002 issue of Nikkei Microdevices, our partner in Japan.