CNET también está disponible en español.

Ir a español

Don't show this again

Tech Industry

2HRS2GO: Analysts take opposite tacks on Intel, AMD

Can two analysts be more diametrically opposed than Ashok Kumar and Tad LaFountain?

The pair are among the most-cited chip analysts on ZDII, largely because they produce very quotable reports. It also helps that they e-mail us their stuff.

This week features the latest round. Needham & Co.'s LaFountain continued his renegade campaign today by reiterating an "avoid" rating on Intel (Nasdaq: INTC). U.S. Bancorp Piper Jaffray's Kumar yesterday issued his latest in a long line of skepticism on Intel's chief rival Advanced Micro Devices (NYSE: AMD); although he kept an "aggressive buy" on the stock, Kumar has little faith in AMD's ability to meet its market share target.

Have an opinion on this?

LaFountain on Intel:

"The emphasis on control is not only misplaced, but counter-productive... Increase in capital spending seen as too little, too late ... management claims that the planning process failed to allow sufficient leeway for the short-term capacity effects of 'equipment reuse' ... This reason sounds better than 'the dog ate my silicon' -- but just barely."

Kumar on AMD:

"Road To El Dorado Is Strewn With Heartaches; AMD Has Tough Road To 30 Percent Share. ... That goal seems wildly optimistic. ... fraught with difficulty ... AMD's ability to reach 30 percent or even 20 percent of the performance segment is doubtful. AMD will also be hard-pressed to hold its gains at the low end."

Like I said, they're quotable.

The Needham analyst's Intel frown stems from the company's recent analyst meeting in New York City.

Intel sees 50 percent annual gains in the communications chip business that's supposed to drive future growth. Fifty percent is much faster than Intel's traditional PC processor business, but also much slower than other network chip companies such as Broadcom (Nasdaq: BRCM) and PMC-Sierra (Nasdaq: PMCS), LaFountain points out.

At the same time, revenue from Intel's core business has failed to keep pace with the uptick in demand. Intel's insistence on maintaining high margins has produced an unusual dilemma for a company with tons of money to spend.

"The failure to grow fab capacity (not just equipment capacity) at a rate sufficient to support 15-20% annual revenue growth while maintaining cash balances of $10 billion on the balance sheet is likely to go down as one of the industry’s stranger blunders," LaFountain asserts.

Executives from Intel point to a stronger second half. To which LaFountain says: so what?

"This scenario hardly constitutes news," he writes. "Intel always has more revenue in the second half of the year. The time that the second half fails to outperform the first half likely would signal the end of Intel as a money-making investment, at least for a protracted period. ... Mismanagement and the inexorable difficulties of growing a $30 billion business are combining to generate a long-term growth rate that is likely to disappoint investors who focus on the data, not the words."

He couldn't be more brutal with a baseball bat.

Meanwhile, Kumar questions some common assumptions on AMD.

Conventional wisdom credits the company's success with Intel's failure. Kumar lays the failure at the other end of the chip market. AMD's "biggest gains have come from the vacuum created as other low-end processor vendors left the market," Kumar believes.

"Many of these customers turned to AMD, whose K6-2 is the sole Socket 7 processor remaining on the market," Kumar says. "The K6-2's speed was fine for these low-end customers. This shift has kept K6-2 shipments from dropping over the past few quarters even as that processor has run out of gas.

Although AMD has seized 30 percent of the overall "value" chip business, it holds just 7 percent in the high-end segment, Kumar estimates. The new Athlon chips are powerful, but Intel's speedier Willamette chip is due for later this year; AMD's next generation product, SledgeHammer, won't reach production volumes until late next year at best, Kumar says.

"Without a performance advantage, AMD's ability to reach 30 percent or even 20 percent of the performance segment is doubtful," he declares.

Kumar also sees problems on the low-end, because Duron "will probably" cost more to make than Celeron, so AMD won't be able to compete on price. Intel also has an upcoming chip that combines CPU, graphics and other functions.

The Piper Jaffray analyst also repeats some long-time criticisms. AMD still hasn't made any significant gains in the business market, which consumes two-thirds of all PC shipments. And the company's motherboard technology choices -- which include old Intel designs and the new, Intel-incompatible Slot A, but not compability with Intel's current line -- force OEMs into hard choices.

"PC makers pondering a switch to Athlon must wonder whether to wait for AMD to settle on a single socket," Kumar says.

And of course, the competition won't lay down.

"Intel will use every trick in its book to dig in its heels," Kumar writes. "We don't see AMD getting much past 20 percent market share in the next year or two."

Imagine what he might have written if he didn't have an "aggressive buy" call on AMD stock.

There are holes in both analysts' arguments.

Kumar continues to raise some spectres AMD seems to have left behind, for the most part. The Socket/Slot issue just doesn't seem to be a big deal, judging by the use of AMD chips in PCs from IBM (NYSE: IBM), Compaq (NYSE: CPQ) and Gateway (NYSE: GTW). The business market hasn't embraced AMD, but that's at least as much the fault of the aforementioned computer vendors as AMD.

Regarding Willamette and the Timna all-in-one (sort of) chip as AMD threats, shouldn't we wait for Intel to resolve capacity problems with its current line before we drool over future releases? Right now -- emphasis on "now", as opposed to "forever" -- AMD seems more likely than Intel to meet its production deadlines.

On the other side, LaFountain gives short shrift to Intel's increasing reliance on investment gains. But what's wrong with becoming a successful investment company? As long as the strategy produces bottom line improvements -- and so far it has -- why complain?

It smacks of the IBM analysts who in 1994 criticized Big Blue's move to services: "But what about hardware growth? How do they jumpstart mainframe growth?"

Who cares? In the end, money is money. And let's keep things in perspective; in the case of Intel, we're talking about a single year of manufacturing problems after a decade of unparalleled success. Intel has the resources to bounce back.

There's another major problem with LaFountain's thesis: it expects a lot from the stock market.

"We understand that this viewpoint runs counter to the general perception of Intel as a core investment," LaFountain writes. "But it is derived from the facts of the situation. ... Mismanagement and the inexorable difficulties of growing a $30 billion business are combining to generate a long-term growth rate that is likely to disappoint investors who focus on the data, not the words."

Numbers rather than PR? Lots of luck on that one. 22GO

• AMD success gives Needham analyst right to crow
• Intel tops 1Q estimates, expects strong demand through 2000>