The big chip news this morning is about a 7-year-old company coming out with a completely new technology that it believes has the potential to replace the type of memory used to store data in phones, MP3 players and solid-state hard drives. Unity Semiconductor also said it’s raised $22 million from Morgenthaler, Lightspeed Venture Partners and August Capital, bringing the total it’s raised to create its chips to $75 million. The Unity silicon will use the movement of ions, rather than electrons and transistors, to store information. The result is yet another competitor to NAND Flash memory, which retains information even when the device is powered off.
The Wall Street Journal quotes Unity CEO Darrell Rinerson saying that Unity plans to use its 60 patents “to keep others out, to not allow this technology to be commoditized.” I’m not sure exactly what Rinerson’s aiming at with that statement, but to a certain extent commoditization — or at least driving a standard — is the key to becoming a successful chip company. Making chips is all about scale, so the bigger your operations are, the better it is for the business, even for those who hold patents. This is essential in the memory market for consumer electronics goods, where companies try to cut their costs to the bone. For example, many are not shy about passing the costs of memory upgrades to consumers.
However, the technical specifications of the memory resulting from Unity’s CMOx process are exciting. It’s supposed to perform 5-10 times faster and can store four times as much data as current NAND Flash. There’s an ever-increasing demand for more data on devices, for storing more pictures, songs or even HD movies. To address this market, SanDisk has been making strides beefing up its Flash products using a specially designed controller, and other memory companies are doing the same. Today Samsung signed a partnership with Fusion-io to speed up Flash inside the data center. Unity plans to have a 64 GB chip out within two years, and because of its production process it believes it can build a fab to manufacturer the chips for about $1 billion — about a third or a fourth less than other chipmakers.
If it can deliver chips with comparable data capacity for a quarter or a third of the cost of competitors, it could create a technology that’s cheap enough to satisfy the consumer electronics market, while also delivering nice margins on each chip. But given how far out its first products are, the number of other memory technologies seeking to replace Flash and how cutthroat the memory business can be, that’s a big if.