Back in late June, KitGuru reported that Microsoft (MSFT -1.16%) might be interested in acquiring struggling chip maker Advanced Micro Devices (AMD 0.19%). The author of the KitGuru piece suggested that Microsoft could be interested in AMD to save money on its game console chips and/or develop smartphone- and tablet-oriented chip solutions -- two arguments that I wrote at the time didn't hold water.
Now, according to a new report from Fudzilla -- which recently claimed that private-equity firm Silver Lake partners is about to buy 20% of AMD -- Microsoft is "seriously talking to AMD" about a buyout. In that same report, Fudzilla also said that its sources "suggest that Intel (INTC -2.15%) is interested in the acquisition of AMD."
Although anything is possible, I think that neither of these rumored acquisitions makes very much sense. Here's why.
Let's start with Intel first
There is a lot of overlap between AMD's and Intel's businesses. Both companies develop and sell processors into the PC market, and AMD has ambitions of once again becoming a viable provider of silicon into the server market, which Intel dominates.
If Intel were to buy AMD, it would essentially get a flailing discrete graphics chip business, a nearly nonexistent server processor business, a crumbling PC processor business (in no small part thanks to Intel), and a reasonably large -- but relatively low-margin -- semi-custom chip business, which is mostly a game console chip business.
One could reasonably argue that Intel could make good use of AMD's graphics processor technology, but I think Intel's organic graphics efforts have improved to the point where buying AMD wouldn't do the chip giant much good.
The only thing that seems as though it would be valuable to Intel would be AMD's CPU patent portfolio, which would include, among other things, the AMD64 instruction set extensions that Intel uses in its chips today.
However, even if Intel were interested in the AMD patent portfolio, it would probably make more sense for Intel to just buy the patents outright from AMD. Doing so would spare Intel from having to take on a business that's in poor shape and that overlaps heavily with its own offerings, and it would allow AMD to try once again to reinvent itself with the potential proceeds.
I think the odds that Intel outright buys AMD are practically nonexistent at this point.
What about Microsoft?
Although Microsoft has made some pretty questionable acquisitions in the past, I think the purchase of AMD would potentially rank among its most outlandish.
The argument that Microsoft would want AMD all to itself for Xbox chips doesn't make a whole lot of sense, but there are some other potential "synergies" that I've seen discussed that I think also don't hold water.
One that I've heard is that Microsoft might be interested in picking up AMD to build (semi) custom chips for its cloud infrastructure. If Microsoft wanted custom server parts from AMD, wouldn't it just make more sense for Microsoft to hire AMD to build such chips, as it did with the game consoles?
That said, it's hard to imagine that AMD would be Microsoft's first choice for such a job, especially since Intel -- which has time and again proved its capabilities in developing server platforms -- has also said that it will do custom chips for large enough customers. Microsoft, in this case, seems a "large enough" customer.
As to the rest of AMD -- which boils down to PC processors and discrete graphics chips -- I can't imagine that Microsoft would want AMD just for chips to power its Surface line of tablets. And also note that Intel's low-power solutions here have consistently been superior to anything AMD has been able to cook up.
I also don't think Microsoft would want to get into the standalone graphics-chip business, especially when it's trying to sell gamers on its Xbox platform.