Renesas merges East and West

New chip firm's management finds that creating a global culture is tougher than blending two Japanese companies

By Bill Roberts, photography by David Toerge -- EDN, November 1, 2004

Sections: A last-ditch effort 1 plus 1 equals what? Consolidating fab processes

Daniel Mahoney, CEO of Renesas' U.S. subsidiary (left) is working with Satoru Ito, president of Renesas (right), to help speed up the decision-making process, which can be notoriously bureaucratic in Japanese companies.

Midway through Renesas Technology's first year, the Japanese company's executives determined that the flash memory market would grow faster than they had expected. Because it produces two types of flash, Renesas was uniquely positioned to cash in on the opportunity if it had enough production capacity. The executives took just 30 days to decide to nearly double its annual capital spending to add equipment to take advantage of the situation. In Japan's semiconductor industry, that's turning on a dime.

"Renesas shortened the decision-making process to at least half of what it used to be," says Satoru Ito, president of Renesas, who was CEO of Hitachi's semiconductor business before Hitachi and Mitsubishi merged most of their non-DRAM chip assets in April 2003 to create what is now the world's third-largest semiconductor company, with nearly $8 billion in 2003 revenue (see the chart "Two Companies Make No. 3 in Semiconductors," below). "It would have been extremely difficult to act this quickly at Hitachi or Mitsubishi."

Ito touts streamlined decision-making as one of the merger's advantages, right up there with the combined fab capacity and complementary product lines in microcontrollers, flash memory and other chips for autos, cell phones and consumer electronics. Streamlined decisions also illustrate the management style pursued by Ito and Koichi Nagasawa, Renesas' CEO and chairman, who came from Mitsubishi.

They're trying to create a culture that can compete in the fast-changing chip industry, a culture with fewer management layers, quicker decisions, a more global outlook and a more Western leadership style than is typical in Japan. "To create an entirely new culture like this is more difficult than to merge the cultures of two Japanese companies," Ito says.

That's because the management style he's trying to achieve is still foreign to many Japanese. "Japanese electronics executives would like to adopt a more aggressive approach to the market," says Greg Sheppard, an iSuppli analyst, "but they can't change their corporate culture." With a blank slate, Renesas has a chance to create a culture in sync with today's chip industry, he believes. "These companies now have to reinvent themselves every year, not every three to five years as in the past."

As Renesas weaves together product lines, marketing, sales and manufacturing, Ito is using training, communication and persuasion to create a new way of doing business. In building the culture, he is striving to meld the best aspects of East and West (see "Consensus Meets Executive Decisions," below ). "The Japanese are still consensus-oriented, and they do things not because they were told to but because they have agreed that they are the right thing to do. The good thing about Americans is that they take charge and own their decisions," says Ito, who spent part of his 30-year Hitachi career in the U.S.

A last-ditch effort

By the mid-1990s, the Japanese were losing semiconductor market share; their way of doing business was failing them. "They were slow to react, slow to make changes," says Trevor Yancey, vice president of technology at IC Insights. Like their Japanese rivals, Hitachi and Mitsubishi were vertically integrated conglomerates, weighed down by the volatile, capital-intensive chip business. In 2000 Hitachi and NEC merged their DRAM units into a new company, Elpida Memory. In 2002 Elpida acquired Mitsubishi's DRAM business.

Neither Hitachi nor Mitsubishi wanted to keep most of what remained of their chip operations. "The companies that made up Renesas were orphans," says Mike Kelly, vice chairman of M&A investment bank Broadview. "They weren't making any money. They were eating up capital. Hitachi and Mitsubishi put them together because they had no place else to go. Renesas was a last-ditch effort."

Vertical integration once worked for Japan, but the cost of competing in semiconductors changed the formula. "Chips are viewed as less strategic now," says Sheppard. One exception is Sony, which sees the semiconductor as core technology in game players, he says. "In most cases, service, software and marketing are becoming the distinguishing characteristics for these conglomerates."

Structurally, the conglomerates have never been able to act quickly. Decisions pass through layers of executives. Boards include people who do not understand the chip business and have to be continuously reeducated. "Some conglomerates have 75 people on their boards," says Kelly. "The vice president of marketing can say no, just because he's on the board."

Ito says a driving factor behind Renesas was the parent companies' recognition that conglomerates cannot make speedy decisions. "The size of these companies was an inherent barrier to doing things quickly enough for the semiconductor market."

Hitachi owns 55 percent and Mitsubishi 45 percent of privately held Renesas, which has a small board of directors and a management team with considerable autonomy. "The purse strings are ultimately controlled by the partners," says Sheppard, "but Renesas is fairly independent and has to stand on its own two feet. It can't depend on getting money from the parents now."

"To create an entirely new culture like this is more difficult than to merge the cultures of two Japanese companies." —Satoru Ito, Renesas

Ito's prime example of how much things have changed under Renesas is its decision to nearly double its capital expenditure that first year, from 70 billion yen to 120 billion yen (about $1.1 billion). Daniel Mahoney, CEO of Renesas' U.S. subsidiary, thinks Ito is too modest when he says the decision took less than half the time it would have previously. "In the past, it would have taken at least six months, and maybe a year, for that kind of decision."

Yancey agrees: "In the past, when you saw planned expenditure levels at the beginning of the year, they never changed much."

1 plus 1 equals what?

The merger also made sense because the products complemented each other. "There wasn't the kind of product overlap that would have forced them to figure out which half of the company to can," says Jim Handy, a Semico analyst. Head-count reductions have been minimal. In March 2004, Renesas had 26,500 employees, 700 fewer than a year earlier. "It moved forward without a lot of restructuring."

Merging the Hitachi and Mitsubishi semiconductor units, which were second and fourth in microcontrollers in 2002, automatically made Renesas No. 1, ahead of Freescale (formerly the semiconductor division of Motorola ) and NEC (see "Two Companies Make No. 1 in Microcontrollers," below). Tom Starnes, an analyst at Gartner, says 1 plus 1 ought to equal more than 2. "It is a bigger company with more resources, and surely it can do more. I just haven't seen anything definitive yet."

In microcontrollers Renesas offers a broad spectrum, from 8-bit to 32-bit technology. Yet it supports two architectures with proprietary instruction sets. Starnes believes that Renesas will need to slowly and judiciously deal with some redundancies, if only to reduce confusion for new customers. "The best thing in the world would have been if those two architectures were the same. Now you have one company supporting both," he says.

Others don't see that as negative. "Once they are established as a customer's architecture, microcontrollers are like an annuity," says Sheppard. "Renesas can migrate customers over time if it chooses."

Ito vows to continue both microcontroller families, because customers want to stick with the one they've adopted and the applications they've built with them. Renesas will not integrate the two or standardize on one instruction set. Ito says that it is integrating the peripheral IP from both companies to work on either architecture.

There are also integration opportunities for microcontrollers and flash memory. Hitachi was one of the earliest proponents of integrating the two. Although Renesas is fifth in the world in flash (see the chart "1 And 1 Make 5 in Flash," below), it is well positioned to capture new market share. Hitachi made AND flash, whose performance is similar to or better than NAND's but with a different architecture; Mitsubishi made NOR flash. Renesas is rare, if not unique, in offering both. "It ended up being the most balanced flash maker in the industry," says Handy. Few current products use both types of flash, but analysts expect to see more in the future.

Ito talks about becoming market-focused, not product-focused. On its first anniversary, Renesas restructured from a strictly product-line organization to having three business divisions: standard products, memory and systems solutions. The latter has units for automotive; cellular; and consumer electronics, mainly PCs and audiovisual devices. Analysts don't see significant results yet.

Ito says Renesas plans to integrate microcontrollers, flash memory and peripherals to offer customers complete systems for devices in all three markets. It is also developing partnerships with software developers, middleware developers and systems integrators. Microcontrollers, especially for the automotive market, are a slow-but-steady cash cow. The other two markets have greater growth potential.

Consolidating fab processes

The fab lineup is another advantage. By Yancey's count, Renesas has 17 production fabs at seven sites in Japan and two sites in Germany. Five were Hitachi; four were Mitsubishi. Renesas would be near the top of the list of semiconductor companies, just behind Samsung and Intel, if they were ranked on capacity, says Yancey.

One site, called Trecenti, previously owned by Hitachi, is a 300-millimeter fab. Yancey says most of the extra capital spending in 2003 went there, doubling capacity to 13,000 wafers per month. That's about half the capacity of the world's biggest 300-mm fabs, but Yancey believes that Renesas has enough for now. The company is now building products at sites from both parents. For example, it makes Hitachi LCD drivers at a Mitsubishi fab. All new designs on advanced technologies—130 nanometers and smaller—will be built through a unified process at Trecenti, Ito says.

 

"In the past, it would have taken at least six months, and maybe a year, for that kind of decision."
—Daniel Mahoney, CEO, Renesas U.S. subsidiary

 

Yancey says both parents were respected for their R&D and that there's no reason to believe that that will change. "The Hitachi conglomerate was especially well known for a long time as an innovator and had one of the largest R&D labs in the word."

Although Renesas is privately held, it has disclosed some first-year financials. It had net income of 8.6 billion yen on revenue of 986.6 billion yen. That put it ahead of plan, says Mahoney. Because of cost reductions achieved in its first year, he adds, Renesas is well on its way to lowering its break-even point by 100 billion yen at the end of 2005.

Analysts say the merger has gone relatively smoothly. Then again, there were few expectations. "You haven't heard about a lot of this going well or that going poorly, because no one knew what to expect," says Kelly.

Others are skeptical. Because Renesas is privately held, management can say whatever it wants, notes Tim Wang, an iSuppli analyst based in Japan. "There's no obligation to disclose anything." Wang believes that Renesas has more duplication than it lets on and will eventually have to deal with it, probably through layoffs, which are difficult in Japan, where most workers still expect lifetime employment. Ito says he plans no layoffs.

Most analysts think Renesas has a chance to become profitable in its markets if it continues to break away from the constraints of the traditional Japanese go-slow mentality and to develop a global outlook. There are early signs it is succeeding. Although 60 percent of revenue comes from Japan, Renesas is stronger globally than other Japanese chip companies, with 10 percent coming from sales in the United States, 10 percent from Europe and 20 percent from Asia outside Japan. It recently had important design wins in the European cell phone business. "It is particularly strong in Europe with the handset guys," says iSuppli's Sheppard. He adds that the Japanese executives at Renesas actually listen to their executives in the U.S. and Europe, which is rare in a Japanese company.

So if Renesas doesn't become more global, it won't be because Ito hasn't tried. Even the skeptic believes that Ito's heart is in the right place. "He has great intentions. I think he theoretically understands what to do," says Wang, "but the time and amount of money it will take to implement this is far more than Americans would allow. I hope I'm wrong."

Would you rather merge two companies or two companies' divisions? Why? Send your thoughts tofeedback@eb-mag.com.

Bill Roberts, a contributing writer at Electronic Business, believes that East is East and West is West but in the 21st century the twain shall meet. 

TWO COMPANIES MAKE NO. 1 IN MICROCONTROLLERS

Rank	Company	Market share Q1-04
SOURCE: ISUPPLI
1	Renesas	21.4%
2	Freescale	15.0%
3	NEC	10.7%
4	Infineon	7.0%
5	Matsushita	5.5%
	Others	40.4%
	Total market in Q1 '04	$3.368 billion

TWO COMPANIES MAKE NO. 3 IN SEMICONDUCTORS

Rank	Company	Sales, in billions of $	Market share
SOURCE: ISUPPLI
1	Intel	27.036	14.9%
2	Samsung	9.675	5.3%
3	Renesas	7.971	4.4%
4	Texas Instruments	7.850	4.3%
5	Toshiba	7.571	4.2%
6	STMicroelectronics	7.238	4.0%
7	Infineon	7.109	3.9%
8	NEC	5.705	3.1%
9	Freescale (Motorola)	4.629	2.5%
10	Philips	4.512	2.5%