it's hard to say that ati/graphics division hasn't been pulling it's own weight lately. However, at best this news is neutral regarding BD, at worst it is bad. Especially now, the last thing AMD wants to do is distract from the ongoing SB recall saga.
Edit: Here's a thought: What if the board is in discussions of a buyout...with Nvidia. That could explain an exodus as many of the higher ups would probably rather flip burgers than work for JHH.
If they managed to live with this executives till now, they could have waited one or two months, till Bulldozer comes out.
With a brand new product, a new CPU that replaces Flinstone era CPUs, it would be just a matter of management turnover.
I have the feeling they grasped this SandyBridge issue just on purpose: "it is going to reduce the impact in times when Intel gets hit with one billion bill and the partners have nothing to sell till April."
It would be pure speculation to st´retch the above in less pleasant waters "it is going to dilute the sensation when we present a CPU that can barely hold up with a non overclocked SandyBridge(and we know it happily works at 4,8GHz), 3 months before Intel gets out with IvyBridge."
Buyout?
The X86 architecture is a dinosaur.
If we would be at the height of X86 it would make sense to buy AMD. However X86 at the stage of a star going supernova; the only thing still fueling X86 is the
miniaturization.
Sandy Bridge relies on 32nm for "new generation" performance.
Ivy Bridge should offer 20% more performance thanks to 22nm.
Rockwell is the last stage-16nm.
The architecture is, with smaller tweaks, the same.
After 16nm things get difficult. The cost of machinery and the complexity of producing such miniature is going to present both a technical and an enormous financial challenge.
Just look at TSMC problems at 40nm (Nvidia made a complex chip; great innovative effort almost ruined by manufacturing; only the second generation that we have now (500 series) managed to be at least properly implemented).
So, Ivy Bridge comes out this year.
Then Rockwell, and after Rockwell?
More cores? You would have to make an unthinkable revolution in the whole concept of programming to use more than 4 cores. And 4 cores are long mainstream.
Nvidia has already caught the opportunity that the 28nm and smaller manufacturing offers-integrating ARM cores inside GPUs. 16nm will make possible to throw X86 out of a gaming PC. And we all know that the majority of us does not run nothing more intensive than games. TSMC is already running sample 28nm GPU (2,4x performance).
Besides, regulatory organs would never give the permit to Nvidia to buy out AMD; on the other hand Nvidia needs every available dollar to push the GPUs they are developing right now (Maxwell; around 20nm).
For market regulatory reasons it's not possible for Intel to buy out AMD.
To Intel the brave new world created with the Larabee project is making available every delight that the incubus presented in the mentioned science fiction.
There is only one possibility of a buyout. A company that wants to shape the future of consumer computers, has the financial means and it's brave and visionary enough.
There is only one such company.
Oracle
Apple is to concerned packing chinaware into glossy plastic.
A buyout would be greatly easier and less expensive if a 4core Bulldozer fails to reach Sandy Bridge.
Dreams about a 6core Bulldozer selling for 4 core SandyBridge are illusive. The cost of manufacturing is always going to be higher, besides, Intel has more efficient productive capacities and considerably more room to lower prices.
Intel management cannot just give the market share to AMD; the investors would throw them through the window if AMD increases market share.
Intel's future?
Intel is going to pay any price to buy Nvidia. Any.
"Hsung Hsung", CEO of Nvidia is probably already counting one billion due to pour on his personal bank account.
Facebook
Twitter