Mannerisms

If, as Saifun and Spansion believe, traditional floating gate technology does not scale much below 40nm, then the world may turn to trapped-charge technology, and then Spansion will have the world by the balls.

Of course the world thinks there are alternatives to floating gate beyond trapped-charge, alternatives like MRAM and Ovonic Unified Memory (OUM) based on chalcogenides.

But, MRAM and OUM have only delivered low density parts, and the only technology which has produced parts which are competitive with floating gate in terms of density and speed, are trapped-charge parts based on Saifun’s NROM technology which Spansion calls its Mirrorbit product line.

It has taken Spansion ten years, working with Saifun, to get NROM technology to this point.

Since buying Saifun, Spansion is now projecting a surge in its licensing business. Of course there will be licence demand from people wanting it for embedded memory in SOCs, but how about the discrete memory makers?

Boaz Eitan, CEO of Saifun, reckons: “The floating gate people will have to look for a new solution, and there is one in the trapping field.”

So will the world’s flash memory manufacturers have to beat a path to Spansion’s door? Or can they design around the Saifun/Spansion patents? Or can they risk ignoring those patents?

“We haven’t see litigation of any sort, at all”, says Eitan, “Saifun’s technology is very well protected. I don’t think people will make investments of hundreds of millions of dollars and betting it on litigation.”

On the other hand, who is going to take a licence from Spansion to make trapped-charge memories, and then try to compete with a company which employs all the engineers in the world who know best how to make trapped-charge work?

You’d have to be pretty desperate.

TOMORROW: THE TEN BEST CHIP APPLICATIONS