Rumor: Amazon in talks to buy TI’s OMAP division

Less than a month ago, Texas Instruments revealed a change of focus for its processor division, which makes the OMAP chips used in a range of smartphones and tablets. Instead of concentrating on those product categories, Senior Vice President of Embedded Processing Greg Delagi told investors the company would devote more effort to wooing industrial clients like automobile makers. Delagi expects slower growth in the embedded market, but he anticipates it will “generate a more stable long-term business” than competing in the wireless device market.

If the latest rumor is to be believed, there might be one player who is happy to chase that purportedly shorter-term wireless growth. According to Israeli site Calcalist, Amazon is in talks to buy TI’s OMAP division. The deal could be worth billions, the site says, and Gizmodo points out that Calcalist’s sources have some credibility; they were the first to reveal Apple’s takeover of Anobit last year.

There’s good reason for Amazon to be interested in picking up the OMAP division, of course. The company’s Kindle tablets, including the latest HD models, all use Texas Instruments processors. Amazon has been widely rumored to be interested in making a phone, as well, and the OMAP line has certainly proven itself in that field. Bringing processor development in-house should help Amazon compete with Apple and Samsung, which design their own chips. The move would also give Amazon an important edge over device makers who source SoCs from other firms.

Comments closed
    • Decelerate
    • 7 years ago

    If anyone needs to be scared, it’s Intel.

    Not a good thing when your [potential] clients are getting the stuff they need to replace you in-house, as opposed to simply going to a competitor, which is easier to counter (when you’re a 400lb gorilla).

      • blastdoor
      • 7 years ago

      Don’t forget Qualcomm and Nvidia. If we end up with Apple, Samsung, and Amazon all using SOCs designed in-house then it means that Qualcomm, NV, and Intel will all be fighting for business from what — RIM, Nokia, HTC, and Motorola? That’s pretty scary. At least Motorola can now be kept alive by Google, but those other three might not be around much longer. Maybe Google will buy NV so that Google can be vertically integrated, too….

      • Sahrin
      • 7 years ago

      >Not a good thing when your [potential] clients are getting the stuff they need to replace you in-house, as opposed to simply going to a competitor, which is easier to counter (when you’re a 400lb gorilla).

      The problem with this is that if what “they need” is x86, then they can’t replace it in-house.

      Desktop PC’s and laptops aren’t going away, which means their margins aren’t going away which means that Intel will always have a prohibitive fab advantage. If (and it’s a big if) ARM ever becomes serious for compute (ie ARM begins to perform the majority of the average user’s compute workload) then Intel will simply shift ARM to the current process node and every ARM manufacturer will go bankrupt overnight.

      Ask AMD what it’s like to try to compete with Intel’s fabs.

        • blastdoor
        • 7 years ago

        The weak link in your logic is that intel will always have a fab advantage.

        AMD was always at a disadvantage due to economies of scale and (related) being undercapitalized. Companies like Samsung and apple are operating at a very large scale, and I know apple is far better capitalized than intel (not saying Samsung isn’t — I just don’t know).

        The reason intel has a fab advantage over Samsung right now is that money isn’t the only issue — having the right people, the right IP, and the time to build the fabs are all issues. But in 5 years or so, the intel fab advantage might not be what it is today. Both Samsung and tsmc (assuming apple switches to them) could be noticeably more competitive.

          • Sahrin
          • 7 years ago

          >The weak link in your logic is that intel will always have a fab advantage.

          Huh?

          >The reason intel has a fab advantage over Samsung right now is that money isn’t the only issue

          [url<]http://en.wikipedia.org/wiki/Semiconductor_sales_leaders_by_year[/url<] Intel spends half of Samsung's annual revenue on R&D, and has for the last 30 years. (That's actually an understatement because Samsung hasn't been nearly as successful as they are now for 30 years). I don't think you understood my post. SoCs are low margin parts (as compared with volume x86 servers and notebooks). It doesn't matter how many hundreds of millions of A6's Samsung fabs for Apple, they will never be able to compete with Intel because they don't make enough money per part to be able to spend *all of it* on R&D. It's a question of economics, not "having the right team." AMD had the right team, and now they're all getting fired because they couldn't produce Sledgehammer fast enough and cheap enough to gain market share.

            • blastdoor
            • 7 years ago

            Interesting link, but not very useful for predicting the future.

            It absolutely does matter how many Ax chips are fabbed for Apple. CPU manufacturing is an industry with very large fixed costs. To make money, you have to spread those costs over a large number of chips. AMD’s problem was always that they were operating at too small a scale without enough money to invest in manufacturing. Apple (and whoever Apple chooses as its foundry, today Samsung, perhaps tomorrow TSMC) will be operating at a very large scale. And Apple has the cash to fund necessary R&D (again — far, far, far more cash than Intel)

            As far as margins are concerned, you’re focusing on the wrong comparison. You’re comparing the margins for Samsung et al to Intel’s margins. But what you should be doing is comparing the rate of return on investments made by Apple in R&D and fabs to the rate of return on Apple’s next best foregone alternative investment. Right now Apple has over $100 billion sitting in low yielding bonds and short term notes. That means that so long as Apple can get better than a 2% return on an investment in the supply chain, it’s a good investment. This is why Apple has spent billions investing in its supply chain — not just pre-paying for components, but actually buying and owning capital equipment that its suppliers then use. If it makes sense for Apple to do that for low-margin suppliers like Foxconn, it sure as heck makes sense to do it for SOC suppliers.

    • WillBach
    • 7 years ago

    I wonder if Amazon would extend it’s “everything is a platform” philosophy here? They sell access to their web store, to their compute clouds, to their database servers, and to their content delivery network. What if they sold access to their SoC design team, too?

    • Sahrin
    • 7 years ago

    Yes! MORE CONSOLIDATION! Just what consumers ordered!

    • deathBOB
    • 7 years ago

    What about ARM servers?

      • NeelyCam
      • 7 years ago

      Hmm… this is an interesting idea

      • brucethemoose
      • 7 years ago

      ^^

      • dmjifn
      • 7 years ago

      You know, if the instructions per watt were good enough, maybe Amazon would run S3 and their website off them.

    • albundy
    • 7 years ago

    lovin the comedic news on this site! a retailer that thinks it’s going to become a semiconductor, LMFAO!!!

    • sschaem
    • 7 years ago

    About Ti focusing on the automotive industry (and other use)… I have a feeling that the #1 target of ipad mini long term is in dash docking.

    So we might see a war of android/custom/ios car consoles in the coming years.

      • destroy.all.monsters
      • 7 years ago

      Possibly. QNX currently dominates that market now though (which is something I hope RIM can capitalize on).

    • jjj
    • 7 years ago

    Except Amazon doesn’t have the volume to justify the dev costs (for now),they wouldn’t have a baseband and that has to be integrated soon while ,long term, the costs will only rise so they’ll need a large slice of the market to stay competitive without spending too much to be worth it.Dev costs are not just what is obvious,there is more integration,more custom cores,advanced packaging.Ofc they wouldn’t have a GPU either and that could be a problem long term.
    On the other hand Bezos attacked Nvidia without much of a reason when launching the Fire HD and that would make sense if they were buying OMAP.
    So it’s possible,just not sure how wise.Maybe if they stop making their own devices and just sell a SoC optimized for their OS but at normal prices the hardware would be a hard sell for now and the strategy would work only when the OS has a solid market share.

      • sschaem
      • 7 years ago

      They would have sold what, 10 million kindle fire + kindle HD by the end of 2012?

      Lets assume that there is a $14 *profit* per SoC and related part, thats 140 million Amazon would save and can re-invest for free in their SoC business. (Ti gets paid 25$ per kindle fire)
      Amazon dont expect their kinlde market to die off, so that would be an increasing an re-occurring revenue.

      The issue is not so much $, but can they keep this business competitive .

      Apple leverage their chip design to have a competitive edge. If amazon can target a SoC to their market & product, they might differentiate themselves and for that alone amazon might be willing to spend hundreds of million on top of the million they would save,

      Ti see its business being attacked by China, the low cost Soc business will die off soon and will be between qualcomm, samsung & Intel for the high end, china low end.

      I think it would be good if Ti SoC can live on, and be used in tens of million of new devices every year.

      I really dont want to see a chinese Mips SoC in future Kindle…

        • WillBach
        • 7 years ago

        I would be shocked if Ti made $14 in profit per system-on-chip sale to Amazon. That said, if Amazon sees itself seeling 100 million Kindles over the next five years (or less!) it wouldn’t be a bad investment.

      • Hattig
      • 7 years ago

      OMAP SoCs currently use PowerVR SGX for the graphics (and the OMAP 4430 in my Playbook performs surprisingly well), I wouldn’t see that changing. TI must have already done most of the design for OMAP Series 5 and a lot for Series 6 as well – the future product line would be something that Amazon would have taken into account when considering buying the SoC design portion of TI.

      Maybe Amazon feel they need to design their own high end SoCs for their devices because they feel the market might not provide otherwise in a few years time. Apple has in-housed design. Samsung already in-houses design (although oddly also uses TI and Qualcomm SoCs rather than their own designs).

      Amazon wants to be one of the future “does it all”* media/technology companies, alongside Apple, Google and Samsung. Maybe LG, Sony, Microsoft. Certainly no longer Nokia!

      * Mobile OS, In-house SoC, App Store, Tablets, Readers, Computers, Phones, Games Devices, TV/Movie store.

      Yeah, I guess we’ll see Amazon phones one day too.

        • ludi
        • 7 years ago

        We have seen the future…and it is vertically integrated.

          • MadManOriginal
          • 7 years ago

          It’s like that theory about technology going in cycles and rediscovering or reusing ideas in a new form. I can’t remember what it’s called, but a common example is the cycle between parallel and serial interfaces.

        • MadManOriginal
        • 7 years ago

        You need to add cloud services to your list. It doesn’t get much attention from consumer-oriented media, but Amazon’s cloud services, and I don’t mean only consumer cloud services which could be considered part of ‘TV/Movie store’, are big and growing fast.

      • willmore
      • 7 years ago

      Admitedly, Amazon has never been afraid of investing in capital, but why they would tie up this much in TIs OMAP division doesn’t make sense. They’re a big customer and I’m sure TI would be willing to customize parts for them. What additional benefit would Amazon hope to get by owning OMAP?

      They may pay less per unit, but they would lose more money by tieing it up in capital. Unless they though they could use *all* of the output of the OMAP group and didn’t need any other chip customers to justify the capital investment. So, unless Amazon plans to use well more chips than they have so far, this would be a bad decision. Just buy the chips and let TI pay from all of the infrastructure and development costs.

        • MadManOriginal
        • 7 years ago

        Bezos has absolutely no problem tying up capital. The key question is whether TI will abandon the appropriate chip development altogether. It seems from TI’s own statements that mobile device and embedded won’t overlap enough to be part of the same development.

          • willmore
          • 7 years ago

          So, you don’t really see this as selling off OMAP, but splitting OMAP into the parts TI wants to keep and giving the rest to Amazon?

      • NeelyCam
      • 7 years ago

      Jeff Bezos is the next Steve Jobs.

    • Anarchist
    • 7 years ago

    … stupidest thing I ever heard, even during options expiration week.

      • link626
      • 7 years ago

      yeah. the market is a big con, manipulated by stupid rumors.

      all these tricks to try to drive stock prices

      • LocalCitizen
      • 7 years ago

      “options expiration week” … what kind of anarchist are you?

    • sschaem
    • 7 years ago

    Apple have its own chip design, Amazon will have its own chip design.. will Microsoft in a couple of years decide to follow ?

      • dpaus
      • 7 years ago

      No, by that time, they won’t even be relevant anymore.

      Oh, wait…

      • destroy.all.monsters
      • 7 years ago

      That’s a very good question. For now they seem focused on being available on different hardware and with different vendors but they may – depending on how they see themselves especially after there’s some consumer feedback on Win 8 – realize that that’s the only way to ensure the level of quality they want; particularly if they want to compete with Apple on equal terms. So we’d not just be seeing the shakeout of the handset chip business but of manufacturers as well. It sounds like the end result is Qualcomm/Google as the handset version of Wintel (with the possibility of intel or Nvidia in the AMD or VIA role).

      That would be nuts.

      • Sahrin
      • 7 years ago

      They already do, it’s called Intel.

Pin It on Pinterest

Share This