Mannerisms

But things could change if one scenario for the future of non-volatile memory plays out:  i.e. if traditional floating gate flash runs out of steam, and Spansion's trapped charge flash becomes the mainstream non-volatile memory technology.

That's because Spansion seems to have a bit of an IP lock on trapped charge, now it has bought trapped charge pioneer Saifun, and because Spansion has a multi-year start on everyone else in the volume manufacturing of trapped charge memories.

If oxides get so thin after 32nm that floating gate ceases to work effectively - and the flash industry is very close to using 32nm processes - then Spansion will be in pole position.

If the industry turns to trapped charge, Spansion has indicated its willingness to license trapped charge technology, which would be a nice revenue stream, while its experience in manufacturing trapped charge memories should give it a market edge over its licensees.

It looks a nice situation to take advantage of. Those $2 shares look a bit of a steal.

And yet, of course, the scenario may not play out like that. Three other scenarios could unfold:

1. Floating gate could find a way to stay effective after 32nm.

2. One of the alternative technologies, phase change memories, could overcome its development difficulties and compete effectively with trapped charge.

3. Another alternative technology, MRAM, could (by some miracle) become a high density, mass-manufacturable technology.

If I don't buy Spansion's shares, trapped charge will become the industry standard way of making non-volatile memory and I'll be telling myself: 'Sucker'.

If I do buy Spansion's shares, floating gate, phase change and MRAM will no doubt demonstrate startling technological improvements and come storming into the market with higher densities, lower power and cheaper costs, and I'll be telling myself: 'Sucker'.

It's a bitch.