Barrons article text
SATURDAY, SEPTEMBER 24, 2005 6:49 a.m. EDT
THE BOTTOM LINE
Once AMD spins off its flash business, the mounting profitability of its microprocessor operation will come into full view. In two years, the stock could be near $50.
On AMD's Menu: Eating Intel's Lunch
By RHONDA BRAMMER
THERE ARE A BUNCH OF REASONS WHY you might be tempted to sell Advanced Micro Devices.
Not the least of them is that the stock of this plucky upstart, No. 2 to mighty Intel in microprocessors, has run from 7 and change a little over three years ago when we did our first feature on the company to a shade under $25 and is still around 23. Taking a profit is usually not a bad thing to do.
Or perhaps you're worried that demand for personal computers, which has been robust for a spell now, just may -- in spite of what upbeat industry pundits predict -- be ready to roll over, and that obviously would not be good news for the chip makers generally.
Or, understandably, you might be concerned that AMD's flash-memory business, which accounts for about 40% of total sales, is capital-intensive and losing a ton of money.
But this time, it truly pays to resist temptation.
Pure and simple, the prospects for Advanced Micro Devices have never been brighter. While semiconductor stocks are always subject to bouts of vertigo and AMD would suffer with the rest if, say, PC sales started to flag, it would still likely fare better than the competition because it's gaining market share in high-end processors at a dazzling pace.
Moreover, the company is busily moving ahead with a plan to spin off the flash-memory biz, known now as Spansion, which would greatly ease that formidable drag.
Buoyed by the smashing success of its innovative new products that have left Intel (INTC) scrambling to catch up, AMD is in a strong position to rack up impressive growth in microprocessors as far as the investment eye can see. And once the spinoff of Spansion is done and flash no longer masks the sizzling profits in processors, odds are investors will focus on a richly profitable business growing 40%-50% annually. Result: Two years out, AMD shares could be pushing $50.
CEO Hector Ruiz has been quietly turning AMD upside down to go head-to-head with Intel.
Despite being vastly outspent, AMD has outfoxed and outmaneuvered its archrival Intel, not simply matching the latter's technical prowess, but outgunning it. Indeed, AMD has determinedly transformed itself from a dowdy also-ran in cloning chips for desktops into a technological powerhouse in designing state-of-the-art chips for servers, taking dead aim at one of the most lucrative markets in all of techdom, a market where Intel once ruled supreme.
Skepticism ruled supreme back in August 2002, when AMD shares were 7 and change, and we described how the company was set to release "truly extraordinary new products" with the potential to change "the balance of power in microprocessors irrevocably," venturing the stock could triple.
The new processor, dubbed Opteron, shaped up as a real threat to Intel's long-standing dominance in servers and workstations, a lush market where processors routinely command gross margins of 75%-90%. What Opteron had going for it -- that Intel's competing chip, the Itanium, did not -- was the ability to efficiently run both high-speed 64-bit software and lower-speed 32-bit software. That crucial advantage allowed customers to upgrade to 64-bit computing and still run all their old 32-bit programs as smoothly as ever.
The Itanium, it turns out, proved a colossal dud.
Intel eventually saw the writing on the wall and scrambled to introduce its own 64-bit chip with true 32-bit compatibility. But by then, AMD had secured a solid foothold in the server market.
The bold foray into high-end chips was a sharp departure for AMD . Traditionally, the company had hustled to compete with Intel in chips for desktops and laptops, while the lucrative market for x86-based server chips remained Intel's private domain.
But Hector Ruiz -- who took the reins of AMD from its founder, Jerry Sanders, in 2002 -- had been quietly turning AMD upside down: AMD would focus on chips for servers, laptops and desktops, in that order. The invasion of Intel's high-end turf was aimed at winning the minds and loyalty of corporate customers, an all-out effort to prove AMD had the right stuff to go head-to-head with Intel. As the normally soft-spoken Ruiz puts it, "We went for the belly of the beast."
Unlike the charismatic and flamboyant Sanders, who got his start in sales, Ruiz -- who is reputed to have an almost photographic memory -- boasts a Ph.D. in electronics from Rice University and comes across as savvy, quiet and careful. He spent 22 years working his way up the ranks at Motorola, running various operations, including Motorola's $7 billion semiconductor business. His right-hand financial man during many of those Motorola assignments was Bob Rivet, today AMD's chief financial officer.
Both Rivet and Ruiz insist that AMD has become "a world-class manufacturer" and that its drive to capture share from Intel is just now beginning to hit its stride. In Rivet's phrase, "This is a marathon, not a sprint."
Today, among major server makers, AMD boasts customers like IBM (IBM), Sun Microsystems (SUNW) and Hewlett-Packard (HPQ). Ironic, perhaps, that HP, which partnered with Intel to develop the ill-fated Itanium, is now the largest seller of Opteron-based systems. So far, more than 75 of the world's largest 100 companies or their subsidiaries have adopted AMD technology, including Honda, Chevron, Charles Schwab, Goodyear, Canon, Airbus, ConocoPhillips, 7-Eleven, MBNA and Deutsche Bank.
AMD's gains have been astonishing: In the second quarter, some 8% of all servers shipped had AMD chips inside, nearly double the 4.8% in the second quarter of 2004, reports Gartner Dataquest.
EVEN MORE TELLING, AMD'S SHARE of processors shipped -- server chips that will wind up in computers in the months ahead -- jumped to 11.2% in the second quarter of '05, from 7.4% in the first quarter, according to Mercury Research. That huge leap -- and server chips can carry price tags north of $1,000 -- translated into sequential revenue growth for Opteron chips in this year's second quarter of a whopping 90%.
Overall, in the first six months of 2005, on 35% higher sales, AMD's operating profits from microprocessors shot up a sparkling 60% over the same span last year. Second-quarter operating net rocketed ahead 89%.
But the sizzle of that performance was largely masked by the huge losses in flash. While operating income from processors in the first half of this year climbed to a record $202 million, from $125 million in first six month of '04, flash memory at Spansion lost $199 million -- virtually wiping out profits from processors.
Flash is the reason AMD won't report much in the way of earnings for the full year: Estimates run from 20 cents to 40 cents a share.
But all that may change dramatically with the proposed spinoff of Spansion, AMD's 60%-owned joint-venture with Fujitsu. The timing may prove fortuitous, since the prices of NOR flash -- the kind Spansion makes -- are starting to firm.
Flash memory -- cellphone makers use a slug of NOR flash -- is a classic boom-bust commodity business. Which is not to say Spansion is a bad company. In good times, it profits mightily. What's more, it boasts cutting-edge technology, and last year was No. 1 in NOR flash, with a 26% share.
But Spansion eats up lots of capital, and its main rival is Intel, which puts AMD in a strategically vulnerable spot. Since flash contributes a modest 7% of Intel's revenues, compared with 40% for AMD , should the urge strike, Intel can brutally slash prices. In that event, Intel may bleed a little, but AMD will hemorrhage massively. And when the money-losing Spansion needs more dough, AMD , as one of the joint-venture owners, has to pony up.
Why AMD Is Looking Chipper: Since Barron's last full-dress piece on Advanced Micro Devices two years ago, highlighting its bright prospects. AMD's shares have tripled while Intel's have gone nowhere. What now?
No details of the Spansion spinoff are available, except that AMD plans to take its ownership below 50% -- to perhaps to 40% if the market's hospitable -- and, concurrent with offering shares to the public, Spansion will do a debt offering.
Proceeds of the debt offering will go partly to repay $340 million AMD has loaned Spansion and AMD will no longer be required to provide capital. After a lock-up period, AMD will be able to sell Spansion stock as the opportunity arises. In addition, AMD's share of Spansion's profits or losses, instead of being consolidated, will be shown as a separate line item, allowing investors finally to get a full view of the fast-growing processor business.
With AMD no longer on the hook to fund Spansion, Intel may not feel the urge to compete as aggressively. In short, Spansion may be a more valuable business if AMD isn't a majority owner. Most important of all, the spinoff of Spansion will allow AMD to focus on its thriving microprocessor business and the huge potential in servers.
Right now, Intel can't match the performance of AMD's dual-core Opteron server chips. "AMD has the clear lead," exclaims Fred Hickey, who writes the High-Tech Strategist newsletter. "Nobody, but nobody, is saying that Intel's chip is better." And Hickey thinks it may be more than a year before Intel comes up with something truly competitive.
"AMD has terrific momentum," he says. "You can smell it in the tech world. They're gaining huge mindshare." It's also evident in the 90% jump in Opteron revenues from the first to second quarter of this year.
Ruiz reckons his company deserves a 20% share in servers by the end of 2006 (though internal AMD projections target a stake as high as 30%). Even 20% would be roughly double its present share. What's more, the spectacular performance in high-profile servers is creating keen interest in other AMD offerings.
Case in point: Turion64, a new chip designed to compete against Intel's Centrino in thin-and-light notebooks, where Intel has been dominant with a great product and neat marketing. But for the first time, Centrino will have real competition. In its first 90 days, Turion posted 60 design wins. Computer makers "want a choice," Rivet says, adding that he thinks in 12 months AMD's 10% stake in laptops can double.
Fine and dandy -- but the bottom line, of course, is, well, the bottom line. Analysts' earnings estimates for next year are all over the lot, ranging from 35 cents a share to $1.23 a share. Much depends on what happens to flash prices, and if, and when, the spinoff gets done and on what terms.
But strip out flash, do some back-of-the-envelope calculations, and it's no huge stretch -- even assuming growth rates way below those of the second quarter -- to see microprocessors alone, which last year (after interest expense and taxes) earned about 55 cents a share, netting about 85 cents this year and in 2006, even with beefed-up depreciation charges, around $1.15-$1.20.
Or look at it this way.
The stock market says Intel is worth $151 billion and AMD , less than $9 billion. Which makes no sense. After all, AMD has one-seventh of Intel's sales and the better growth prospects. If AMD had even one-seventh of Intel's stock-market value, it'd be a $53 stock.
Hickey, who has a paired trade -- long AMD , short Intel -- insists, in his usual understated fashion, AMD "could be the investment of a lifetime."
News 
Market Data 
Discover 
AI
