San Jose, Calif. -- The disk drive industry stands at a difficult double crossroads, a place where both its enabling technology and its market drivers are shifting.
Even as they navigate the transition to perpendicular recording, drive makers must choose a next-generation technology in 2007 if they are to maintain their 40 percent annual capacity growth and fend off the competition from flash memory. But industry leaders Hitachi Global Storage Technologies (formerly the IBM disk drive division) and Seagate Technology Inc. are divided on the way forward, with Seagate leaning toward thermal recording heads and Hitachi favoring patterned media.
And in the shadow of the debating giants, much of the rest of the industry has yet to make up its mind.
The next-generation technology, whichever one it turns out to be, could require historic levels of investment. That's especially worrisome because drive makers are poised for still more consolidation under business conditions that are both difficult and uncertain.
"We face more-complex engineering problems with the next generation than ever before, at a time when we are struggling with costs, so it's difficult to make the investment," A. Currie Munce, vice president of research at Hitachi Global Storage, said at the Diskcon convention in Santa Clara, Calif., last week.
PC makers, still the biggest buyers of drives, are hammering on costs harder than ever. Consumer electronics companies, forecast to become the primary users of drives by 2009, are almost equally cost-conscious and even harder to forecast.
"We are really at an inflection point in this industry," said James Chirico Jr., executive vice president of manufacturing operations at Seagate. "If I had to sum up the industry for the next 12 to 24 months in a word, it is 'uncertainty.' "
Meanwhile, the continued slide in flash memory prices has undercut the business for the relatively new, 1-inch disk drives, once thought to be a rising star for some consumer systems.
Apple Computer Inc.'s launch of the flash-based iPod Nano last year marked a turning point in the flash vs. hard-drive battle. The Nano demonstrated that 1- to 2-Gbyte flash MP3 players represented a bigger market than 8- to 12-Gbyte disk-based units. Meanwhile, cell phones have been slow to adopt hard drives, with the exception of a handful of image products from Samsung and Nokia.
Thus, Hitachi has delayed the introduction of any new 1-inch drives. And a Toshiba Corp. executive said demand for that company's novel 0.85-inch drive has been far lower than expected.
"I wouldn't say the 1-inch drive is dead," said Chirico of Seagate, which has a 12-Gbyte model, "but volumes aren't anywhere near what we thought they would be, and people are canceling programs."
Drive makers are in an overcapacity phase. Taking advantage of the situation, some PC makers have recently invited drive makers to "reverse auctions" where the lowest bidder wins the majority of the business, said Chirico. "They are long and aggressive meetings," he said. "The cost pressure has never been greater on us."
Conditions could worsen for drive makers as consumer electronics OEMs become their main customers. CE vendors are at least as frugal as their PC brethren, and much less able to forecast their diverse, seasonal and hit-driven businesses.
PC makers bought about 84 percent of all hard disks in 2002. But they will buy only about 61 percent this year and an estimated 46 percent in 2009, as new consumer applications continue to grow and drive demand. For example, the first terabyte drive, expected in early 2007, will probably be a 3.5-inch product aimed at high-definition digital video recorders.
"Companies like Seagate used to have about six major products, two for each [market category]--desktops, notebooks and servers. Now we have 42," said Mark Kryder, Seagate's CTO. "It really strains our resources to deliver that breadth."
Seagate hopes for further industry consolidation in the next two years, either among the handful of remaining drive makers or the few independent head and media suppliers. The company completed its acquisition of fourth-ranked rival Maxtor in May, increasing its market share to more than 40 percent and laying off thousands of workers in the process.
"We cherry-picked that operation. It was pretty brutal," said one Seagate manager.
Hitachi and Western Digital Corp. are jockeying for second place, with respective market shares of roughly 15 and 17 percent.
Seagate and Hitachi agree that thermal recording heads and patterned media are on the horizon and that they must decide within a year which to pursue first. But that's where the consensus ends between these players, each of which owns all the internal head and media operations needed to go it alone if need be.
Seagate thinks its Heat-Assisted Magnetic Recording (HAMR) technology, already under development with a government grant, is the next best step. It re- quires a laser heat source to raise the temperature of a tiny recording spot on the disk by several hundred degrees for perhaps 150 picoseconds. HAMR entails new recording heads and waveguides to align optical and magnetic focal areas.
The company has already demonstrated an optical device that can read from an 84-nanometer spot, a figure it needs to get down to 50 nm for commercial products. "This does not look terrifying, considering all we have learned with past-generation recording heads," said Kryder.
By contrast, bit-patterned media could require 12.5-nm lithography, which won't be available in the next decade, he said.
But Hitachi's Munce sees patterned media as a better choice for a next step. The company is investigating a nanoimprinting technique for mechanically pressing charge- able spots in a disk, an approach shown in university research to yield features as small as 5 nm, he said. The process still must prove able to meet the cost, throughput and yield requirements of mass-market disks.
Munce believes Seagate's HAMR process is more risky because HAMR's use of both heat and magnetism introduces essentially a new physics. That means it would require new materials and manufacturing processes for the heads, in addition to the new optical technology.
Either technology could push drives to an areal density of more than 1 Tbit/inch2. And both will be required in concert to get to 50 Tbits/inch2 and beyond. The only question is which to pursue first.
Seagate and Hitachi could travel separate paths, which would make for an exciting race of technologies. But the industry might be better served by an agreement on one direction that pushes other players into alignment and spreads out investments more broadly.
"We can't afford to commercialize both [techniques] in parallel," Munce said, speaking about Hitachi's resources. "Ultimately, I think Seagate will swing their ship around."
Time is growing short. The perpendicular recording just now emerging in today's drives will run out of gas somewhere between 500 Gbits/inch2 and 1 Tbit/inch2, experts estimate. At a density growth rate of 40 percent a year, that means a new technique will need to start shipping as early as 2010, and it takes about three years to transfer research into volume manufacturing.
"We are very close to a decision on what we have to do for the next generation," said Seagate's Kryder.
"If we want to stay ahead of flash, we have to have the next technology ready when perpendicular runs out of gas. We can't afford not to be ready," added Munce.
While the technology leaders debate, the remaining drive makers and independent head and media vendors are stuck in a wait-and-see mode. "No one is driving them," said Kryder.
No. 2 drive maker Western Digital has adopted a fast-follower strategy. Others note that the drive industry, unlike its big brother the semiconductor business, is too small to fund development of an industry road map. In any case, most drive makers are still focused on their transition from longitudinal to perpendicular recording, which is forcing changes in heads, media and other parts of the drive.
One component supplier said there are still changes in the works in perpendicular heads to better trap renegade particles and thus raise drive reliability. A media vendor predicted the transition to perpendicular recording will stretch out over the next couple of years.
Western Digital started shipping its first perpendicular drives just two months ago. So far, WD is using the technology only on its notebook drives.
See related image