Hideo Inayoshi, a top engineer with Hitachi Ltd. (HIT ), helped turn his company into a world leader in the production of microcontrollers, those tiny computers embedded in everything from cell phones to microwaves to automobiles. Throughout his 30-year career, Inayoshi's most important job was to beat competitors like Mitsubishi Electric Corp. to the next innovation. So his last assignment was full of irony. Late last year, he took charge of negotiations that landed a deal in March to merge Hitachi's system-chip operations--centered on microcontrollers--with Mitsubishi's by April, 2003. Such a deal would have been unthinkable just a few short years ago. But that's the price of survival. Says Inayoshi: "We all feel a sense of crisis."
Some would go so far as to say a sense of panic. The Japanese chip industry, the terror of Western rivals and the undisputed master of the global business a decade ago, is in deep trouble. Sure, the Japanese retain areas of excellence, like system-on-chips for consumer electronics. But they've lost key markets to rivals in the U.S., South Korea, and Taiwan as they've fallen behind in the high-stakes race. Collectively, the Big Five--Hitachi, Fujitsu (FJTSY ), Toshiba (TOSBF ), NEC (NIPNY ), and Mitsubishi Electric--lost $4 billion on an operating basis in 2001 just on their chip operations. While most chipmakers around the globe are predicting an upturn in business, the semiconductor operations of the Big Five will remain in the red. Hitoshi Shin, first vice-president of Merrill Lynch Japan Inc., is forecasting a combined operating loss for this year of $1.2 billion for the group.
In good times, these industrial combines, which manufacture everything from computers and plasma displays to telecommunication equipment, could offset losses in chips with earnings from other divisions. But today no one has any cash to spare. That's why the only way out is to merge or die. By the time the dust settles, the Big Five may look more like the Big Two, centered on NEC and Toshiba Corp., the strongest players. "There are just too many chipmakers in Japan today making the same products," says Shigeyuki Umezawa, who oversees the semiconductor sector from the Ministry of Economy, Trade & Industry. "They have to regroup or merge."
Indeed, the industry is abuzz with reports of complicated new alliances. After signing off on the Hitachi merger, Mitsubishi Electric's new president, Tamotsu Nomakuchi, said he is also open to a chip tie-up with NEC. Hitachi has already spun off part of its memory chip, or DRAM, operations into a joint venture with NEC, and Inayoshi is now shopping around for other partners for Hitachi's remaining memory business. Fujitsu is reportedly interested in folding its system-on-chip operations into those of Toshiba.
While the Japanese ponder their next move, their situation deteriorates almost by the day. At its peak in 1988, Japan controlled 67% of the global chip business in terms of production value; by 2001, its share had declined to 29%, according to Merrill Lynch. South Korea displaced Japan as the champion of memory chips in the first half of the 1990s. Then the second half of the decade saw the rise of chip-design firms in the U.S. that farmed out their manufacturing to Taiwan's fast and efficient contract manufacturers. And before long, China will emerge as a major manufacturer and consumer of all kinds of semiconductors.
Why did Japan fall behind? For years, the Big Five focused on mastering the capital intensive technology needed for producing chips. The Koreans and Taiwanese, by contrast, simply bought state-of-the-art chipmaking equipment from Japan and the U.S., and then won over customers with their speedy service. "Sure, the Koreans and Taiwanese spent more on equipment, but they hit on the winning formula," says Merrill's Shin.
The U.S., as well, stole a march on Japan. American companies like Intel (INTC ) and Texas Instruments (TXN ) abandoned memory chips years ago to focus on niche areas--Intel on microprocessors for PCs and TI on communications chips for mobile phones. Japan also lags behind the U.S. in the important field of chip design, whose leaders determine the technology road map for future development. "If we hope to set standards in next-generation chips, we need to boost our design capabilities," says Toyoki Takemoto, chief operating officer of the Semiconductor Technology Academic Research Center (STARC), an industry body promoting design research.
Concerned about Japan's declining competitiveness in a key industry, the government has sprung into action. Lawmakers approved a supplementary budget earlier this year that provides $240 million toward the cost of building a test wafer plant making chips with a circuit width of 100 nanometers (or 0.10 micron). Considering that 130 nm is now state of the art, the project is aimed at helping Japanese companies jump to the next level. That's just the beginning. The government has boosted its budget for spending on science and technology to $185 billion from 2001 to 2005, up by 40% over the previous five years, with a large chunk slated for industrial research.
Industry and academia have also banded together to launch the most aggressive chip-related research effort since the mid-'70s. Last year researchers from both sides started working together on two projects that aim to push the edge in system-on-chip design into the 70-50 nm range over the next several years. The market for system-on-chips that will power sophisticated multimedia devices is expected to pass $200 billion by 2005, and Japanese industry hopes to capture a 25% market share.
Japan's chip industry also wants the country's universities to contribute more in the way of technology development. The private sector used to account for 90% of Japan's research and development, but over the past decade the ratio has declined to 78%. NEC Chairman Hajime Sasaki, a chip industry veteran, heads a government committee that is helping draft legislation to foster more collaboration between industry and academia. One reform would enable universities to own and license intellectual property developed in university labs. "We're trying to accomplish what the U.S. did in the early '80s, when it enacted reforms that encouraged tie-ups between industry and universities," says Sasaki.
Another survival strategy for chipmakers is to join with consumer electronics companies that can provide marketing savvy and expertise in applications. NEC and Matsushita Electric (MC ), which both helped Nintendo Co. (NTDOY ) develop its new GameCube video-game machine, may soon be ready to co-sponsor new projects. Toshiba's prospects have improved thanks to its collaboration with Sony Corp. (SNE ) in developing chips for Sony's hugely popular PlayStation 2 game console, released two years ago. Last month, Toshiba, Sony, and IBM (IBM ) announced they would co-develop an advanced, multimedia system-on-chip that would power not only the next PlayStation but also a wide range of network devices of the future, including home servers--computers that will process entertainment and other data flowing in and out of the home. "This is the kind of product that will set us apart from the rest," says Yoshihide Fujii, vice-president of Toshiba's semiconductor operations.
Many obstacles still remain for Japan's beleaguered chip industry. For starters, there's a shortage of engineers capable of designing the kind of system-on-chips required for next-generation electronics. Another problem is the slow pace of decision-making. "Japanese companies are typically too slow and bureaucratic," says Akira Minamikawa, senior analyst at WestLB Securities Pacific Ltd. in Tokyo. "In this industry, you need to be fast and decisive." If Japanese chipmakers fail to make the grade, they know the consequences: a humiliating slide to the bottom of an industry they helped get off the ground 20 years ago.
By Irene M. Kunii in Tokyo