Somewhat slanted article. Seems like author poorly researched all the publicly available information on Ramtron's supply constraints.
Note that the major reasons for their total supply constraints were actually commented by the Ramtron CEO at the Third Annual Green Technology in Boston on March 8th, 2011.
Ramtron actually has a link to this webcast on their own website ( http://www.ramtron.com/investor-relations/calendar-events/ )
Here is the direct link to Conference webcast: http://www.wsw.com/webcast/kbro22/rmtr/
Note major reasons for supply constraints to their clients was commented at ~34:40min mark of conference webcast by Ramtron's CEO.
Net of reasons stated by CEO:
(a) Ferroelectric RAM (F-RAM) Process & Tooling very unique (not standard CMOS)
(b) Ramtron did not give adequate forecast of expected growth to their foundries. i.e. CEO states that Ramtron "gave heads up 9 months too late" to their Si Foundry in Dallas
(c) Back End Test capacity in Asia was not adequate to support demand, but they are now working towards improving.
Problems with getting FRAM on CMOS up and running? Who would have thought it. I mean it's so simple to take an existing 0.18um IBM CMOS process and us it as the underlay chip technology and the take a completely foreign 0.35um process and for the last 3 metal layers to come out with desired FRAM.
Just because the equipment that was supplied by Ramtron operated with a different operating system, different programs (using different programing languages than those on the IBM site) having no way to communicate with existing on site equipment or operators; there shouldn't be any long term manufacturing problems introduced from the start now should there.
Of course if one adds on some of those little recurring oversights that designers seem to keep having time to production keeps slipping. Can anyone remember the that simple little question that no one asked when low K dielectrics were the new production rage back in the late 1990's? Can anyone say Silk? How about Black Diamond? How about 1 billion dollars plus thrown away by the industry before someone finally asked what the coefficient of expansion and contraction was for these new wonder dielectrics compared to the other materials in the connecting layers of the chips?
One would think that after wasting over 100 million alone a IBM Burlington on trying to make chips using Silk work that that simple quarry would be added as a must ask question when designing a chip. Oops. Some one forgot.
That said Ramtron has a valid complaint with the extended time that IBM is taking to address all the little and big problems that have surfaced in attempts to ramp production of a product that works in according with the desired specifications.
Either IBM is not serious about this business or there are some process QC issues for the above said batches. I cannot in any way understand why IBM would want to scuttle a business which it can dominate if it became cost effective.
@goafruit: IBM is one of the largest fab companies- next time you play with your PS3, XBox or Nintendo Wii think about what company makes the chips :)
Very smart move by IBM to leave the consumer PC business... after all more game consoles are sold than PCs
-disclaimer I am a former IBMer