Some data I kept on a notepad: Samsung presented PRAM papers in 2003. As of 2004 there have been papers describing 64b and later 256b parts but without being clear if those were functional prototypes instead of just papers. Functional prototypes at least by Sept '06: 200mm, 90nm, 512Mb, single die, 0.0467um2 cell size (best in industry), verticle diodes with 3D trannies. Working 512Mb prototypes also shown Feb this year. Production initially at 512Mb and scaling down later. Competitors: Newco, Tosh, Renesas, Elpida, Sony, Hitachi, IBM, Macronyx, Qimonda. Some licensed IP from Ovonyx. April '07 Intel prepares sampling 128Mb/90nm NOR part. PRAM might if highly produceable have the potential to replace other mem techs incl. DRAM, i.e. in addition to NOR. High speed, good density, non-volatile. AFAIK Spansion doesn't have access to PRAM and will have to acquire the technology and the experience somehow if necessary (e.g. via merger / acquisition). For now they apply rasor sharp focus on NROM only as basis for Mirrorbit (Dual/Quad), ORNAND(Dual/Quad), and Eclipse(Dual/Quad/hybrid).
My rather rough wag (I'm definitely not an expert here) is Samsung will start ship volume no later than EOY (Q4?). It'll take time to adopt especially outside Samsung. The ramp will take time too. Roughly the same applies for Newco. At this moment I guess serious impact for Spansion might arrive by '09. I'm ok investing heavily in Spansion this year but will review my position at least at EOY.
Pls do correct me where you find necessary; I haven't exactly had time to really check it.
Regards,
Rink |
