Diversification, anyone will tell you, is a smart move. Excelling in one area is no guarantee against a changing market. Common sense dictates that you hedge your bets. But even the most successful companies have stumbled by straying too far from core competencies that fueled their original success. And, despite its allure, diversification may be as risky as sticking it out in a flattening core market.
Case in point: Last week’s much-anticipated announcement that Intel Corp.—the company that had been “printing money” making the Pentium chip at the heart of a gazillion personal computers—would sell its communications and application processor businesses to its Santa Clara neighbor, Marvell Technology Group Ltd., for $600 million.
The deal transfers ownership of Intel’s XScale technologies, which include chips for Research In Motion Ltd.’s BlackBerry 8700, the Palm Treo and the Motorola Inc. Q, among other devices. Marvell intends to use the XScale technology to become a player in the cellular phone and consumer electronics market.
After the deal closes, which is expected in four to five months, Intel plans to continue to manufacture products that were sold to Marvell, while Marvell—a “fabless” company arranges for other fabrication facilities.
“Intel is soul-searching to determine how to make money,” said Len Jelinek, analyst with iSuppli. “They spent several billion dollars to enter the mobile communications business and couldn’t make a go of it.”
The $64,000 question, looking back, is why? How can a large company full of incredibly bright people fail at a seemingly reachable, sensible goal?
Jelinek paused, and cautioned that his view was not based on inside knowledge of Intel’s workings, but on more mundane, outsider perceptions of brand identity.
“People in the wireless industry may well have looked at them and said, `They’re not really in this industry, in terms of development and structure.’ It may or may not be true, but that may be the perception. That and the fact that they were always second fiddle to Texas Instruments [Inc.].”
Purchasing agents in the wireless space may well have always overlooked Intel’s specific wireless offerings and simply viewed Intel as more of a PC chip maker, asking themselves `What is the particular strength and strategy of this company?,’ Jelinek speculated. “Intel’s strength and strategy was not in this space.”
The trend among many companies, Jelinek said, is to focus on core competencies and, in Intel’s case going forward, likely includes improving margins by migrating its fabrication facilities to next-generation technology.
Intel’s manufacturing strategy is to move its ongoing PC chip fabrication from 200 millimeter wafers to 300 millimeter wafers, Jelinek said. (Chips are made on wafers. Larger wafers offer more cost-effective manufacturing, which improves margins.)
As Intel continues to manufacture its XScale chips for Marvell under the terms of the deal, and as Marvell seeks another fabrication partner, Intel might use the transition to explore upgrading its existing plants to 300 millimeter wafers, rather than the capital-intensive option of building new facilities from scratch.
“Over time, as Intel takes stock of its manufacturing, it has to face the question, `What do we do with our older facilities?’,” Jelinek said. “Refurbishing older facilities is less capital intensive than building new ones from scratch.”
Marvell is not left without its own challenges.
“They’ve absolutely got to advance the technology,” Jelinek said. “The manufacturing of the XScale technology has been done with mature processes, second-generation manufacturing. They need to take it to advanced processes with their new foundry partners so that they can reduce costs and improve overall performance.”
Will Strauss, principal at Forward Concepts, said Marvell stands a good chance of success in the wireless space.
It’s not well-known, Strauss said, but Marvell is the market leader in the 300-million-unit-per annum hard disk controller market, second only in size to the mobile handset’s 800 million to 900 million unit annual sales, but larger than the PC market’s 200 million unit annual sales. Strauss said that Marvell understands digital signal processing technology—required in both disk controllers and wireless handsets—and they already have a wireless presence in wireless local area networks.
“Marvell is better focused and more aggressive than Intel could ever be in communications, and I predict that they will do well in the cell-phone chip market—if they can cope with personnel transitions,” Strauss said. “The biggest problem for Marvell that I see is culture shock for the incoming Intel folks and many won’t survive. I hear Marvell’s management is very aggressive.” About 1,400 Intel employees are said to be transferring to Marvell with the deal.
As for Intel?
“I estimate that they invested $3 billion to $5 billion in its cellular baseband and application processor operations since 1999,” Strauss said. “They wrote off $600 million and the deal with Marvell brought another $600 million, so in that sense, this deal was a fire sale. Now they can stop bleeding and do what they do best—print money with Pentium chips for PCs.”
Marvell itself has said that in terms of market strategy it will move into the market for chips for smart phones and PDAs, while Intel will refocus on its core competencies, including low-power processors for PCs and emerging technologies for mobile computing, including Wi-Fi and WiMAX broadband wireless technologies.
Marvell’s current efforts include development of storage, communications and consumer silicon solutions, and its product portfolio includes switching, transceiver, communications controller, wireless and storage solutions for communications infrastructure for enterprise, metropolitan, home and storage networking.
