<<Rambus' most serious DDR/SDRAM patent is for the register bit(s) that distinguishes CAS delay in SDRAM and DDR. [I.e. CL2 vs CL3 SDRAM.] But the patent (if it is valid for SDRAM) not only covers that register bit, it also covers the technique of how you change the state of that bit. Every time an SDRAM is powered up that register has to be set, and the thing that sets it is the memory controller. Thus Rambus' patents cover the memory controller.
If the memory controller is in a chipset, costing $20, then Rambus presumably takes 4% of that $20. If the memory controller is in a high end integrated CPU that costs $500, (as I expect the industry to converge on within 3 years or so), then Rambus presumably takes 4% of the $500.>>
Thanks Carl.
Stupid questions now.
What is CAS?
Can the memory controller be stand alone, or is it always in a chipset?
Would there possibly be a memory controller and an additional memory controller in the chipset as well, or is there only one?
Why are you presuming Rambus is asking for 4% of $500 when a chipset sells for $20. I would presume Rambus is and should be asking for 4% of the latter, although the former certainly would be a boon to Rambus shareholders. Is it possible that this is just all part of the negotiations now that is holding up a settlement?
Is there any way to know what Toshiba and the others that have signed up are paying, or are they only in the memory business and not the chipset/controller business?
Is a chipset really $20? Chipset prices for DDR, SDRAM, and RAMBUS?
Instead of bickering, I am just trying to figure out what RMBS might be worth if things do go their way, but 4% of a CPU does not seem reasonable to me.
Based on the above questions and comments, care to try again on a value?
The next phase will be to guess as best as we can possible scenarios and liklihoods of Rambus IP standing up on all these things. Based on assumptions people are willing to make about that IP standing up, we should be able to come up with a risk reward scenario. |
