 |
Nov. 30, 2004 – In the same way that digital cameras boosted demand for flash memory, newer, high-resolution cameras, cell phones and digital camcorders are expected to drive demand for next-generation portable storage. Executives at companies developing MEMS-based memory think that these high-density, portable applications will most likely be the first market for their devices.
But companies are still deciding the best ways to address that market — whether to manufacture their own products, work with partners, license technology to third parties or pursue a multi-pronged strategy.
 |
“When you can buy a state-of-the-art disk drive for $90, there has to be an incentive to pay more,” said Ed Schlesinger, director of the data storage systems center at Carnegie Mellon University in Pittsburgh. Such incentives could include greater portability, better durability or more capacity.
The companies — IBM, Hewlett-Packard and Nanochip, to name a few — have developed storage technologies that couple a grid of MEMS cantilevers with atomic force microscope (AFM) tips. The tips poke nanoscale dents in a polymer surface to store bits of data. They use various techniques to erase bits, such as relying on surface tension to pull material back up from the bottom of a previously made dent.
The differences in technology and the need to provide a compelling value proposition, according to company executives, mean these devices will not initially be hard drive replacements.
“It’s not a hard drive,” said Tom Theis, director of physical sciences at IBM, whose Millipede project is the best-known effort in this area. “Flash memory, portable storage — it has to compete for these markets.”
Conventional disk drives can be made both rugged and small, but they cannot fit into the smallest removable cards used in digital cameras. Flash memory is more expensive per megabyte, but its tiny size and lack of a rotating disk make it ideal for portable electronic devices.
Nanochip’s Gordon Knight agrees with Theis. “We don’t believe this is the disk drive market at all,” said the Fremont, Calif.-based startup’s chief executive. “It’s flash.”
The challenge is that the flash market is careening along a price/capacity curve that’s steeper than Moore’s law. A 1-gigabyte memory card whose manufacturer’s suggested retail price was $299 just a few months ago now retails for less than $70, the result of a pricing war that has slashed industry margins. At that rate, experts say that MEMS storage will have to come in at a much higher density to justify the price tag that would most likely be attached to a new product category.
“You have to be at the high end,” said Nanochip’s Knight. He and other experts say the technology could reach storage densities greater than a terabit of information per square inch by 2010 or 2011. As for initial products, both IBM and HP have said 2006 or 2007 are reasonable time frames. Knight said he agrees with the estimates.
In the meantime, the players are evaluating their options. Hewlett-Packard decided to seek a licensee for its MEMS memory technology, according to research and development director Tim Weber. Weber discussed the decision during a speech he delivered in early October at Chicago’s NanoCommerce 2004 trade show. Hewlett-Packard decided against launching an internal program after concluding that building such components is not part of its larger strategy.
IBM has not yet made a decision on whether to develop a product in-house or with a partner, although Theis said licensing out a technology while simultaneously developing a product is common at Big Blue. He said the company has been approached to use Millipede technology for scanning probe lithography, an application for which it could issue a license while still developing a storage product.
Theis added that the company will “know more” by year’s end, but that even if a decision is made it doesn’t necessarily mean there will be a public announcement.
Nanochip is already working with partners, according to Knight, who said such collaborations are a necessity for a small startup. The company will attempt to outsource the manufacture of its product and supply it to customers, rather than license its IP.