Founded in 1999, ZettaCore is betting its 20 employees and $23 million in venture capital financing on delivering standalone memory chips as well as process and device intellectual property for embedding memories within systems-on-chip. The company claims that its molecular-memory element can be embedded in a standard CMOS process flow.

The ability to shrink traditional memories to smaller process geometries is becoming more limited by materials issues than by lithography. Hence, new breeds of memory technologies are being explored, such as electromechanical nanotube memories (Nantero), molecular memories (ZettaCore) and MRAM (IBM, Infineon, Motorola). Each of those technologies has its opportunities and challenges.

ZettaCore's memory technology, like existing memory technologies, is based on charge storage principles. However, the company's devices store charge in organic molecules that self-assemble on a silicon substrate. The molecules look like lollipops, with the stick part adhering to the substrate through chemical forces and the sucker part storing charge. A sea of these lollipops is spin-coated onto the substrate and then lithographically patterned. Once the sense and drive electronics are added to the memory cells, the bit density is on par with DRAM, assuming one storage bit per cell.

So why is this better than DRAM? First of all, charge storage in ZettaCore's memory elements occurs in quantized states. This means that it is possible to store multiple bits per cell. Second, although ZettaCore's initial devices are volatile, the refresh cycle is on the order of seconds, leading to substantially lower standby power as compared with DRAM and SRAM.

ZettaCore has one of the most impressive pedigrees of investors, advisers and managers that I've seen in a startup-and I've seen a lot. Still, the company has its work cut out for it. As mentioned at the start of this column, other companies with new technologies are also racing for the same prize. Also, those with traditional memory technologies may yet have a few tricks up their sleeves.

If ZettaCore can prove the manufacturability and reliability of its technology, then it has a good chance of capturing a market with enormous potential.

Jeremey Donovan (jeremey.donovan@gartner.com) is chief analyst at Gartner Dataquest.