After weeks of rumors,
sold off the first significant chunk of its business as part of its plan to refocus the company.
But while the $600 million Intel reaped for its line of handheld communications and application processors didn't excite its own shareholders, it went over like rotten fish among investors of the firm on the other side of the table --
Shares of Marvell plunged 15%, or $7.76, to close at $44.14. Intel's stock slipped 23 cents, or 1.3%, to $18.05 amid Tuesday's broader market downturn.
The deal gives Marvell an entry into the cell-phone handset business, one of the hottest semiconductor markets in recent months. But the near-term dent it will put in Marvell's margins, along with the fact that the business is not best of breed, seemed to be foremost on investors' minds.
"This isn't the first time the company's made a significant acquisition," says Doug Whitman, of Whitman Capital, which has a position in Marvell. "
But this is the first time they've done a dilutive acquisition that's going to be significant in the short term."
According to Marvell's management, the X-Scale application processor and the baseband processor businesses it acquired generate about $100 million a quarter combined. Under Intel's roof, both businesses had operating losses.
The acquisition will hurt Marvell's earnings by 10% in the fourth quarter of fiscal 2007, and roughly 5% throughout 2008. But Marvell has set an aggressive timetable to turn things around by the fourth quarter of 2008.
The acquisition will give Marvell a long-term gross margin of 50%, which Stifel Nicolaus analyst Cody Acree said is lower than the company's previous long-term guidance of 53% gross margins.
Marvell CEO Sehat Sutardja, however, sees the deal giving Marvell a unique chance to move into the cell-phone market with a running start.
"Strategic opportunities like this are rare," Sutardja told analysts in a conference call Tuesday morning. The acquisition, he said, gives Marvell all the pieces it needs to offer handset makers a complete platform for building advanced, 3G cell phones, a market he said will drive Marvell's growth for years to come.
But analysts and investors questioned how well Marvell would do with a business that's been such a laggard for Intel.
While Intel has achieved a few marquee customers in the smart-phone and PDA market -- notably
Research in Motion
-- the company has had limited success selling the chips to the mainstream cell-phone handset market.
"Our caution stems simply from Marvell's ability to expand on INTC's limited success," wrote Stifel Nicolaus' Acree in a note to investors. Stifel Nicolaus makes a market in Intel and Marvell securities. The firm has provided noninvestment banking services to Intel within the past 12 months.
And Marvell will be competing against heavyweights like
, both of which have deep connections with handset makers -- not exactly an easy market to penetrate.
According to Marvell executives, the reason the processors never took off at Intel was because they were capacity-constrained due to Intel's model of manufacturing them with older equipment. Marvell's plan involves gradually shifting manufacturing to a third-party foundry, TSMC, so that the chips can be produced with 65-nanomer circuitry, which will cut costs and increase output.
"A significant amount of the move from the current operating loss position ... to an accretive position is going to be based on volume expansion," said CFO George Hervey. "The good news is the customers have already been obtained to do that ramp. The design wins are there. It's now a matter of ramping the production to those levels."
Marvell's founders have shown a prescient ability to detect markets where the company can flourish, even in the face of tough competition.
But even if the cell-phone move ultimately pays off for Marvell, some wonder why Marvell was not able to get a better deal from Intel, which was widely reported to be looking to unload the business.
"I'm just surprised
about why they didn't extract more pain from Intel," said Jefferies analyst Adam Benjamin. He points to the 1,400 employees Marvell will take on (a roughly 50% increase to its current headcount) as one part of the deal that could have been structured more beneficially to Marvell.
And given that Intel will continue manufacturing some portion of the chips for Marvell for the next two years, the companies could have shared some of the gross-margin impact, says Benjamin. (Jefferies makes a market in Marvell and Intel securities and has provided noninvestment banking services to Intel during the past 12 months.)
"Over the long term it's probably going to work out for them," Benjamin says. "It's just going to take a long time to play out."