added to its investors' worries Thursday by mouthing two of the most unpopular words in a down market: share loss. Though it's too early to tell for sure, CFO Andy Bryant said the company's January price increases on memory may have cost it some flash business.
The unhappy twist: If he's right, the revenue trouble probably wouldn't be limited to one quarter.
That's because Intel's price increases on
flash probably hurt it in terms of new design wins rather than sales to existing customers. If so, missed design opportunities could weigh further on revenue growth down the road.
To understand why, consider the workings of the cell-phone market, where around two-thirds of Intel's flash ends up. Though flash is broadly considered a commodity, it's not completely interchangeable.
"Flash chips don't directly drop in and replace each other because they program differently and the software is a little different," explains Jim Cantore, principal memory analyst for iSuppli.
In other words, products from No. 1 player Intel and No. 2
Advanced Micro Devices
can't be easily substituted for each other. Once a handset maker designs in one product, it's typically reluctant to change -- even if that supplier boosts the price.
But it's a different situation with a new handset design, when a manufacturer has far more latitude in choosing its supplier. In that case, flash buyers might be more inclined to shy away from the expensive stuff. The bottom line: Intel's price increases may not have been enough to lose it existing customers, but they might have proven a sufficient turn-off to new ones.
Such a market-share loss would have longer-term consequences for revenue. "If they lost a couple of major designs as a result of their price hike, that would be very serious for Intel down the road," explains Cantore.
Says Lehman Brothers' Dan Niles: "When Intel originally said they would raise prices, what we said was that flash suppliers don't set the prices -- the market does, because it's commodity memory. It's hard to set a price when several other guys are more than happy to take your business."
To give Intel credit, it can claim impressive share gains over the past year. Between 2001 and 2002, Intel saw its share in NOR flash jump from 29.7% to 36.2%, according to preliminary estimates at iSuppli. Meanwhile, AMD's share slid, declining from 16.6% in 2001 to 12.8% last year.
Still, the minitrend doesn't favor Intel, as its share has more recently ebbed a little at the same time AMD has gained slightly.
Between the third and fourth quarters of 2002, Intel saw its share dip from 37.1% to 36.7%, while AMD saw it edge up from 12.7% to 13.1%, according to iSuppli.
AMD declined to comment for this story on whether it may have benefited from Intel's loss.
In January, Intel justified its price hikes to analysts by saying it had been shipping higher volumes of flash memory lately. But there's evidence that sales to the end market, cell phones, were underwhelming.
Shipment trends in the third and fourth quarters "tell me we probably had a surge of new products into the sales channel and demand tapering off at the end market," says Cantore. He believes consumers didn't adopt high-end cell phones as fast as expected -- a worrying trend for Intel, which draws the bulk of its flash revenues from that market.
Though the company's flash business was initially forecast to fall slightly in the first quarter, Thursday's update suggests it fell 15% to 20%, says Lehman's Niles.
Intel obviously didn't expect a sales shortfall when it pushed through the price hikes, but history shows price increases are a risky strategy in a period when the end markets are soft.
Take DRAM, for example. "Exactly a year ago,
jumped DRAM prices in January and was followed by every other major manufacturer. But it didn't hold," recalls Cantore. "Within two or two and a half months, market prices came down and even dropped a bit. It didn't stick because end-market demand for PCs was weaker, so the price increases were not sustainable."