Marvell: The Streak Is Alive
Yesterday, Marvell reported revenues of $135.9 million for its quarter ended Nov. 2, compared with $119.7 million the previous quarter, marking its 19th consecutive quarter of revenue gains.
That's not to say Marvell is profitable -- losses were $7.7 million, or 6 cents per share, compared with losses of $9.3 million the previous quarter, and $105.1 million for the same quarter a year ago. (See Marvell Reports Q3.)
Marvell shares traded around $23, down $1.33, late today.
Chances are, it's a simple case of Marvell attracting too much attention. The company beat estimates, but everybody expected them to beat estimates (the return of the infamous "whisper number"), and that caused a run-up on the stock.
"There was really no upside left because for the past two weeks, Marvell investors have been beating the table saying, 'Company's gonna beat! Company's gonna beat!' " says Analyst Ambrish Srivastava of Gerard Klauer Mattison. He points out that in the last few days, Marvell's stock has grown twice as fast as the Philadelphia semiconductor index (SOX).
Aside from its numbers, the company announced it has begun shipping 802.11b chips for revenue. Probable customers there include 3Com Corp. (Nasdaq: COMS) and D-Link Systems Inc., writes analyst Jeremy Bunting of Thomas Weisel Partners, in a research note issued last night.
Marvell also announced it's selling Gigabit Ethernet components to 3Com for use in network interface cards, an important complement to its ongoing sales of chips for Intel Corp.'s (Nasdaq: INTC) NICs. "That Marvell has taken business at 3Com suggests to us that the Marvell product is competitive and that, for now at least, Marvell will continue to be the leading player in the GbE PHY [Gigabit Ethernet physical-layer chip] space," Bunting writes.
The relationship with Intel is expected to fizzle once the chip giant gets its own parts perfected, but Marvell is preparing for that day with its Yukon chip, which competes with Intel (see Intel/Marvell: Who Was Stood Up?). Bunting thinks the companies' collision could be a year away, however. "It remains our belief that Intel is six months or more away from sampling its own product and a further six months away from production -- at least."
— Craig Matsumoto, Senior Editor, Light Reading