Poor Old Numonyx, it never stood much of a chance. It was formed from two loss-making businesses and was still making losses when it was sold, today, to Micron.
Its birth was accompanied by pangs of financial pain. Under the original terms of the joint venture deal, Intel and ST expected to share a $900 million payout.
The banks backing Numonyx were expected to put up a $1.3 billion loan and a $250 million revolving credit to finance the joint venture, and $900 million of that $1.3 billion was to be shared as a cash hand-out to Intel and ST.
Then the credit crunch started to hit. The banks re-thought.
Instead of the $1.3 billion loan, the banks would now only put up a $650 million loan, while insisting that some $400 million of the loan had to go to Numonyx, and not to Intel and ST.
That left $250 million for Intel and ST. ST ended up with about $130 million of it which was further reduced by a $30 million re-structuring charge.
So poor old Numonyx got started on a loan of $400 million, and a revolving credit of $250 million.
Which was not much for a venture which, on the Intel side alone, had been losing about $300 million a quarter.
ST made it plain at the time that it was unprepared to put any money into the company in the future.
Numonyx’s own production never got onto anything better than 90nm. Its products never made it to densities higher than 1Gbit.
For higher density memories, or 3bit-per-cell technology, Numonyx relied on its partner, Hynix.
Almost as a fig-leaf to mask its problems, Numonyx made a song and dance about its phase-change memory technology as a possible replacement for floating gate flash, but it never made a part denser than 128Mbit, at a time when the leading edge for commercial floating gate memories is 32Gbit.
The technology was dubbed a Techno-Ponzi scheme – designed to keep investors sweet.
So poor old Numonyx was born sickly, lived sickly, and now departs to its Idaho resting place as a sickly soul.