• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

[El Reg] Report: IBM to peddle its chip wing

Not unexpected after they sold of their server division to Lenovo.

IBM is simply pulling the plugs. Service and software is all they wanna do.
 
Not unexpected after they sold of their server division to Lenovo.

IBM is simply pulling the plugs. Service and software is all they wanna do.

But they only sold the x86 server division- their POWER server (and mainframes) division is still in IBM hands.
 
If intel doesn't open fabs - and no mubadala version 2 comes along.... won't it eventually only be TSMC & Intel?

Samsung has already proven it'd kill its manufacturing if profit is greater elsewhere.
 
If intel doesn't open fabs - and no mubadala version 2 comes along.... won't it eventually only be TSMC & Intel?

Samsung has already proven it'd kill its manufacturing if profit is greater elsewhere.

The endgame have always been between those 2, with Samsung only as a possibility. IBM was never any importance. And GloFo is a standing joke.
 
Well that would be unexpected. Would anyone else really want to produce POWER chips, whose only customer is IBM? And what would this mean for the Foundry Club, who rely on IBM R&D to stay vaguely competitive?

I posted this around november:

http://www.idc.com/getdoc.jsp?containerId=prUS24476413

Both Oracle and IBM have been experimenting crashing sales *and* crashing margins in their respectives hardware divisions. With IBM software sales slowing down, it was a matter of time to look at their hardware division to free up capital.

Oracle is a little trickier, because they acquired SUN by their own will when most of the market was against it, there's a lot of executive pride at stake, so we can expect them to wrangle a little more to try to make their hardware business work. The irony is that Oracle hardware business is far more pressed by x86 than IBM's.

And who would want to manufacture POWER chips for IBM? There's a subpar foundry whose shareholders have very deep pockets. I think they would be glad to manufacture chips for IBM, even at very low margins, as long as they could also snap all IBM R&D personnel. That would fill a capability gap they sorely need to.
 
Ah, is it just the fab that they are selling? From the way that The Register phrased it I thought that they were selling off their entire CPU design team too. (Though that may just be shoddy journalism from El Reg, as per.) Keeping the CPU team and going fabless would make more sense I guess. Though they will run into the problems AMD has- IBM needs very specific, expensive nodes for their CPUs (like their 22nm node which none of the Fab Club are using), and it seems unlikely that an independent fab would make those kinds of processes.
 
Ah, is it just the fab that they are selling? From the way that The Register phrased it I thought that they were selling off their entire CPU design team too. (Though that may just be shoddy journalism from El Reg, as per.) Keeping the CPU team and going fabless would make more sense I guess.
I will wait for IBM official statement, because it could be they will only sell their fab(s?) and keep both CPU and process R&D, or sell everything.
 
Ah, is it just the fab that they are selling?

According to the FT they are open to selling the entire hardware business. Fabs, chip design, R&D, the entire deal. The FT also mentions that they are open to other arrangements beyond an outright sale, such as a joint venture.

IMO this is not going to be a simple operation, because nobody but Intel and Samsung have vertically integrated semiconductor business, and I pretty much doubt that Intel would be interested in acquiring POWER (POWER is dying because of x86, it's a market that will be theirs in due time), and Samsung seems to be much more focused in developing their ARM business. They should be talking about multiples arrangement with multiples parties. Selling the foundry business and building a JV for the IC design business or selling the two business seems to be the most probable outcomes for me.

In that case, the subpar foundry seems the prime candidate for the foundry business, as that acquisition would help them with the R&D part. IC design is more complicated, who is going to pay to design and build chips to IBM, given that the POWER sales are plunging and there is no clear escape route for them?

Ed: Plus the subpar foundry already uses IBM foundry technology.
 
Last edited:
Don't confuse POWER systems with their mainframes.

The service revenue around mainframes is billions of dollars.

But the customers are fleeing from all IBM hardware segments.

The company’s sales declined 5 percent to $27.7 billion in the period and fell 5 percent to $99.8 billion for the full year. The real shocker was the implosion of IBM’s hardware business. Data-center equipment revenue fell 26 percent to $4.3 billion. Mainframe sales dropped 37 percent; Unix server sales, 31 percent; Intel based server sales, 16 percent; storage sales, 13 percent.
 
Real professionals don't deal in absolute terms. ( collapsing... Dead. Etc )

I wouldn't ever describe IBM in those types of terms...

Their power chips will continue to be great. No one can top their medical AI like servers.
 
Which they still can provide, POWER or no POWER. It's the same with HP and Dell.

Nope, services are bundled as part of a contract. Stop producing mainframes, lose the services business.

Anyway, POWER isn't going away. THE WSJ has the real article, IBM is looking to sell their fabs is all. Not get out of the chip business.
 
The vastly bigger story here is the FAB club (Common Platform alliance). This kinda leaves AMD, Samgsung and GloFo out in the cold if Samsung or GloFo doesn't pick this up.

Edit: Oh, and this could kill UMC.
 
Last edited:
In that case, the subpar foundry seems the prime candidate for the foundry business, as that acquisition would help them with the R&D part.

I don't see UMC having that kind of cash on hand. Unless you mean STMicro? No wait, it must be Samsung!

We're all well aware of your deep seated fear and loathing of all things AMD and GloFo, but they're not Beetlejuice. They're not going to turn up at your house if you say their name too often. 🙄
 
The vastly bigger story here is the FAB club (Common Platform alliance). This kinda leaves AMD, Samgsung and GloFo out in the cold if Samsung or GloFo doesn't pick this up.

Edit: Oh, and this could kill UMC.
Well Samsung is definitely competent enough and has the resources to continue moving forward.
 
They already laid off a lot of the engineers last year, before the 401k payment was due. They brought some back as contractors for their current contracts but none of the ones I know for how long. They even turned down some contracts as they don't have enough people to work on them anymore.

So its known there will be more lay offs if a sell does not happen this year.
 
The vastly bigger story here is the FAB club (Common Platform alliance). This kinda leaves AMD, Samgsung and GloFo out in the cold if Samsung or GloFo doesn't pick this up.

Edit: Oh, and this could kill UMC.

IIRC IDC(I'm not really sure but it's probably him) once said that the members of the FAB club doesn't really benefit that much from the Alliance since it was more focused on the needs of IBM rather than the group. If that's the case it shouldn't really hurt them that much
 
IIRC IDC(I'm not really sure but it's probably him) once said that the members of the FAB club doesn't really benefit that much from the Alliance since it was more focused on the needs of IBM rather than the group. If that's the case it shouldn't really hurt them that much
GloFo does benefit (as well as UMC), because they're not capable of keeping up with their competitors without IBM's aid. GloFo's got a little romance going on with Samsung to remedy this.

But yes, you're correct. IDC did say that, and it was in regards to gate first vs. gate last. IBM was pursuing the increase density provided by gate first, while the other foundries wanted the increased yields and performance of gate last.
 
IIRC IDC(I'm not really sure but it's probably him) once said that the members of the FAB club doesn't really benefit that much from the Alliance since it was more focused on the needs of IBM rather than the group. If that's the case it shouldn't really hurt them that much

No, he said the fab club doesn't get as much benefits as it should because the arrangements are centered around IBM needs, not around the needs of the members.

That doesn't change the fact that all of them except Samsung lack the R&D disciplines to develop a node and they will suffer if someone from outside grab IBM business and pull IBM assets from the alliance. That's why I think Globalfoundries is the prime candidate for the assets.
 
Last edited:
Back
Top