It's Time For Samsung To Buy AMD, And AMD Should Let It (Op-Ed)

Earlier this week, we learned that three of AMD's top executives have quit the company, after a new CEO took over last fall. The new CEO likely wanted new people to help her run the company. However, even with a new team, chances are slim for AMD to make a strong comeback in the chip market over the next few years.

Why AMD Should Sell

While AMD is struggling to survive in the PC chip market, the whole PC chip industry is facing an increased threat from the mobile chip makers such as Qualcomm. AMD can't even address this mobile threat right now due to a lack of sufficient income. AMD scored a major win with the Playstation 4 and Xbox One contracts, but even with those, the company is still rapidly losing revenue and profits.

Getting acquired might become inevitable for AMD in the next few years, but it would be better for both the company acquiring it and AMD itself if an acquisition came while its products were still relatively competitive.

AMD's foundry partner (and former division) Global Foundries is also falling behind lately, having to license Samsung's 14 nm FinFET process in order to keep up with other foundries. By choosing Global Foundries, AMD is usually a couple of generations behind in process node technology, which only deepens the gap in performance and power consumption between its chips and Intel's chips.

The x86 baggage is still real for mobile, but at least Intel can hide it by being ahead of ARM chip makers with at least one generation-newer process node. When AMD is typically behind ARM chip makers in adoption of new process nodes, that gives it no chance in the mobile market.

AMD also has a problem of image that can only begin to recover once the company's products are more competitive. To do that, AMD will need to be on cutting-edge process nodes, hire great chip engineers, and have enough financial backing to do the required R&D. A stronger investment in marketing and branding would also be highly beneficial.

Why Samsung Should Buy AMD

One company that could help AMD solve all of its problems is Samsung. Samsung's priorities align quite well with AMD's, making it a great match.

For starters, Samsung makes billions of dollars in profit every quarter. Samsung could give AMD as much money as it needs to compete not just in the PC chip market against Intel, but also in mobile. AMD is now working on its own custom ARM CPU core, and there have been rumors that Samsung was working on one as well.

It's likely AMD has more experience designing CPUs, though, even if they are on an ISA AMD hasn't used too much until recently. AMD could potentially help Samsung design a better CPU than it could on its own.

Samsung also tends to just use ARM's Mali graphics, but it could instead start using AMD's GPU technology, which is among the best in the industry. AMD could help Samsung build its own CPU and GPU in order to differentiate itself from all the other mobile manufacturers who have now begun using either Qualcomm at the high end or Mediatek at the low end.

Samsung could simply follow Apple's strategy of designing its own CPUs in order to fully control the hardware that goes into its devices. However, in this case, Samsung would go a step further, as it would then be designing both the CPU and the GPU. This is in contrast to Apple, which only designs the CPU and licenses the GPU from Imagination Technologies.

In the mobile market there also seems to be a need for more competition right now, as OEMs don't have too many choices left. Samsung could take advantage of this situation to establish itself as a main competitor to Qualcomm. Samsung sells a huge number of mobile devices. By using its own chips in its own handsets and also by selling them to other OEMs, Samsung could actually dominate the mobile chip market in terms of market share.

The company would also have the advantage of using its highly-competitive foundry to manufacture the chips, where it could further cut down the costs. This could allow the company to save money in its own devices, or sell the chips at a lower price than Qualcomm to other manufacturers.

Samsung has also been toying with the idea of entering the server chip market with its ARM chips, mainly because it's a highly profitable market. However, even for a company as large as Samsung, that could prove quite difficult. Samsung is going head to head against Intel, the dominant player in that market, not just with a new chip, but with an architecture that's also new to the server market.

Through AMD, Samsung would have access to years of experience in the server chip market, as well as both an x86 ISA and an ARM one. Using both architectures would make Samsung more adaptable to the various needs of its enterprise customers than Intel. Both AMD and Samsung are also founding members of the "Heterogeneous Systems Architecture" (HSA) Foundation, making it even more clear that the goals of the two companies align. The HSA should give the two companies increased synergy between their two main architectures of choice.

Transferring AMD's x86 license to Samsung could become somewhat legally challenging, but that's nothing a little more money sent to Intel can't solve.

To recap, by allowing itself to be acquired by Samsung, AMD would gain significant financial backing to achieve its most ambitious goals, strong marketing campaigns, chips built on cutting-edge process nodes, and access to a large portion of the mobile market (including Samsung, but also other manufacturers' devices) in which to put its chips.

By acquiring AMD, Samsung would get a chip company that has much more experience in designing its own chips, access to the x86 architecture as well as AMD's advanced GPU architecture, and entrance into the profitable server chip market thanks to both AMD's relatively solid brand in that market and its experience.

The acquisition would greatly benefit both companies, but it remains to be seen if they see it the same way. Samsung may believe it can do just as well without AMD (unlikely, especially in the server market), while AMD might want to keep its independence for as long as possible (which may weaken the company too much in the future).

Follow us @tomshardware, on Facebook and on Google+.

Lucian Armasu
Lucian Armasu is a Contributing Writer for Tom's Hardware US. He covers software news and the issues surrounding privacy and security.
  • TechyInAZ
    I'd prefer this over AMD going bankrupt.

    I love Samsung products, the quality is very good on all Samsung products I've owned. so they both would make a good partnership.
  • therealduckofdeath
    Why would Samsung buy a company that's primarily building x86 processors at a loss when Samsung seems to be getting themselves out of the PC industry slowly? They've already announced they'll stop selling PC's in Europe.

    That would literally be like a PC manufacturers like HP buying a failing mobile company right after they'd just killed of their own mobile production. And looking at how that went for HP, I'd say fairly confidently, Samsung has no interest in AMD.
  • mapesdhs
    Merging company tech in such a way can be horribly complicated. Even when it
    sounds like a grand idea in theory, in reality it can all fall apart over differences
    between management, etc. SGI ran into this sort of problem when they tried
    to merge with Cray.

    I was moaning last year that it if only some outfit like IBM could buy AMD, maybe
    they'd finally have the resources to get back in the CPU game, but one has to
    ask, is the desktop CPU market something AMD needs to get into once again?
    Does it really matter given the current tech trends for tablets, smart TVs, mobile,
    phones, etc.? Last quarter saw a further 2.4% decline in PC sales (personally
    I think this may be partly because the current generation of PC tech is overpriced).

    I'd love to see AMD recover, we need the competition, but a buyout may not help
    even if it does sound ideal on paper.


  • phatboe
    Intel's x86 license that they grant to AMD will not be granted to Samsung if they buy AMD, so since that is where most of AMD's assest lie (x86 CPUs) I highly doubt samsung would buy AMD. Though if AMDs new custon designed ARM processor turns out to be highly competitive to other ARM processors Samsung might be more interested in buying AMD.
  • SkyBill40
    I, too, would favor an acquisition as opposed to outright failure and bankruptcy. While there would undoubtedly be growing pains when considering all the tech they'd invariably be forced to merge together, they'd end up helping each other along the way and likely coming out much better at the end for it.

    If AMD is to survive and once again be truly competitive in the PC market, they need a huge influx of cash as well as some new thinking. They could get both in one fell swoop from Samsung. This move could also give Samsung even more power versus Apple.
  • SonSon1
    Such amount of BS, firstly, AMD needs to make a comeback in the CPU market, otherwise the writer is denying, or ignorant, that GPUs are also chips, and that AMD is getting to be very competitive in that segment.

    On the other hand, GloFo does not license 14nm FinFet from Samsung, both developed the process in a partnership. Now, perhaps the writer is not up to speed, TSMC just reported to experience another 6 month delay in their 16nm FinFet, so from where does the write pulls the argument that AMD is always a couple of generations behind. Nvidia and AMD BOTH are on 28nm right now.

    In all, one of the most pathetic articles I've read on tomsh
  • SirTrollsALot
    AMD should stay in the US. Id rather see AMD go bankrupt than get raped by some foreign entity.
  • FCsean
    <Mod Edit>
    <Moderator Warning: Watch your language on this site>
  • FCsean
    This is one of the most dumb articles I've ever seen.
  • CaedenV
    Not sure that it needs to be Sammy, but I agree that someone needs to rescue ATI's GPU technology from the sinking ship that is AMD.
    But Samsung would be a bad pick for a company like AMD. Samsung is a product based company; meaning that they find what is hot, put in bare-minimum effort in acquiring enough IP to put the product together, ship it, and move on to the next product. They still manage to make a few excellent products (SGS phones, Galaxy Note, and 840/850 SSD lineup to name a few) in a few market segments, but their hits are far fewer than their flops.

    Still, someone needs to buy them. Maybe in a few years they will have fallen to a point where VIA will be able to pick them up and finally do something interesting with a patent expansion?