Maybe Micron's thinking is this: if they don't acquire scale by buying Elpida, they have 100% chance of being destroyed by Samsung over the long term. If they take a big risk on Elpida, they have a 90% chance or crashing and burning and a 10% chance of surviving with enough scale to beat Samsung at the memory game.
DRAM and flash will always exists (for a while), the question is if micron's cost structure will be competitive. Hynix walked away since this one ll increase their cost of production.
how can micron pull a magic?
the ironic part is elpida bought qimonda's munich team 2008 and now is their turn to fall apart.
while hynix's wuxi fab saved it after 2004.
is micron getting too smart this time?
DRAM and Flash are reaching the point where shrinking dimensions becomes prohibitively expensive. In the next few years, emerging memories, such as STT, Conductive Bridge, 3D NAND will go into production. Micron will have fabs in place to dedicate to these new memories, while their competitors will have to disrupt their current production to phase them in.
If micron dare to swallow elpida its fate sealed, at least need 2 years to merge, convert product lines, culture friction will be huge.
It will be teared apart very soon.
what MU doing is trying to harass Hony's effort, raise the price, it will leave in a while as toshiba did.
To complete the picture, we have Micron's 2.3 billion debt (I don't know if this includes the bonds just mentioned) to add to Elpida's 5.6 billion debt, so roughly 8 billion, but taking out the 2.5 billion bid leaves 5.5 billion, which was too hard for Elpida to handle, but maybe Micron will somehow miraculously get through it. If Micron (or Elpida) were a Korean company, no doubt the government would have done something; inappropriate arguably, but something. Hiroshima operating expenses reported by Elpida to be so high, which is why they had to set up Rexchip. I think Micron will end up trying to sell the Hiroshima fab. Who will dare buy then?
Micron has roughly 2 billion in cash and just sold 800 million in bonds so they definately have the cash to buy Elipida. They will most likely keep as much cash on hand as they can and finance the rest. If Hony won the bid, it could be another potential competitor, which could prolong the down cycle. Elpida spent 5 billion on their Hiroshama fab and another 2.5 billion for Rexchip. Buying Elpida will be a burden for Micron for a few quaters, but once things start rolling Micron should start constantly being profitable. With Elpida's extra capacity they finally have the scale to compete with Samsung and Hynix on volume. Both Samsung and Hynix have mentioned converting some dram capacity to over to app processor or something else.Try to name another semiconductor company that has been around this long competing against countries and banks. Once Micron's TSV technology start gaining traction, they will further break away from competitors. Dram prices are already starting to go up slowly. With this increase in revenue, Micron should reach top 5 largest semiconductor company by revenue. Hopefully with increased margin to go with it.
Micron doesn't need to have cash on hand to make this transaction. They can hand over some of their stock, or they can get a loan from a bank, or sell bonds. The key question is are Micron shareholders better off in the long run. Based on the mixed experience acquiring Qimonda's share in Inotera, it is a tough call.
Micron doesn't even have 2 billion in cash and had negative earnings the last few quarters. This 2.5 billion bid is fake report from Japanese media I hope. Otherwise, it's desperation on top of no financial control.
Blog Doing Math in FPGAs Tom Burke 18 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...