8 Comments

Summary:

Reports of Amazon considering the purchase of Texas Instruments’ mobile chip business for billions of dollars may not make sense at first glance. But there are several reasons why such a deal could make sense for Amazon in both devices and cloud computing activities.

omap5thumb

Amazon is considering a bid for Texas Instruments’ mobile chip business unit. The Next Web noticed that an Israeli news site, Calcalist, reported on Monday that the deal is in advanced negotiations and will be valued at billions of dollars; TI’s market cap is currently $32.35 billion. Texas Instruments’ OMAP line of chips powers various phones, ereaders and tablets, but the company has been steadily losing ground to other chip makers in this market, such as Qualcomm, Nvidia, and Samsung. TI will reportedly keep its other lines of business for sensors and automotive chips.

Indeed, TI’s OMAP line has had few major product design wins of late. It powers the new Kindle Fire HD tablets, the Barnes & Noble Nook Tablet, and a fair number of Motorola phones. Even Motorola, however, has recently switched to alternatives from Qualcomm: The company’s new Razr HD and Razr HD Maxx will use a Qualcomm Snapdragon chip; a change from the prior Razr models that relied on a TI OMAP.

Texas Instrument’s OMAP woes, however, could be a benefit to Amazon in several ways. First, although the up-front cost would be massive, the purchase could cut the unit costs to produce hardware such as Kindle e-readers and tablets. Currently, Amazon sells hardware roughly for cost, with hopes of recouping revenues and profits through content sales. By producing chips in house, profits could be seen on the hardware although it could time to realize such benefits; it’s not cheap to sink money into long-term capital such as chip fabrication production lines.

Kindle Fire HDHaving its own chip business could lead Amazon to customize the silicon so that its optimized for Amazon’s products. This approach is similar to that of Apple, which creates custom chips that are made to run iOS to its full potential. While Amazon doesn’t have a mobile operating system of it’s own, the company is using Google Android as a base with Amazon software for the user interface. With a chip business, Amazon could improve the performance of its tablets by optimizing the union of silicon and software.

This situation does cause one to think about Amazon getting into the smartphone business as well. As noted, TI’s OMAP chip already powers numerous Motorola handsets as well as my own Samsung Galaxy Nexus handset. I’d say there’s an outlier chance here as the smartphone market isn’t an easy one to crack. It’s not out the realm of possibilities, of course, and if Amazon is going to spend billions for a chip business, it will surely want to maximize its return on that investment.

That leads to one other possibility here if Amazon were to buy the OMAP business from TI: Servers. While Amazon is known as a retail and content company, it’s a cloud company when you peek inside.

The company relies heavily on its cloud services for the retail experience, Whispersync technology and its AWS, or Amazon Web Services which let users rent computing storage and processing power in the cloud. For some time, x86 chips have powered cloud-based servers, but cloud computing can be done on smartphone chips as well; an OMAP purchase could reduce Amazon’s cloud costs and even lead to custom hardware that’s more power efficient.

  1. Tsahi Levent-Levi Monday, October 15, 2012

    Kevin,

    This is a huge leap to make – using OMAP chipsets on servers. While I’d love to see that happen, I can’t see it come to fruition in the foreseeable future. Amazon must be able to bank such an acquisition from Kindle sales alone.

    Share
    1. I see your point for sure, Tsahi. If Amazon were to spend billions, however, I see it as a long-term investment; not a short timeline. Still, you could well be right. Thanks!

      Share
      1. Seems like quite a shift in strategy for them, but people have said this many times before. They are very good with their investment dollars so there may be an interesting angle as you suggest where they can justify it on their current business but have some optionality in other areas.

        Share
    2. using OMAP as a server is already being done for many applications. currently sandia labs as well a wide range of universities use OMAP3 and OMAP4 for parallel computing projects. it is a very short step to mass production servers….

      Share
  2. “chip fabrication production lines.” oh my what a stupid thing to say, are they buying UMC now too?
    Anyway,Amazon doesn’t have the volume now to justify making their own chip,maybe when they sell at least 20 mil units/q but that could be more if they make more than 1 chip that fits all.So if they do buy OMAP a phone is pretty much a certain thing since other phone makers are unlikely to use their chip and they will have to generate enough volume to justify the dev costs.
    Another big problem is that they won’t have a baseband and sooner or later they will have to integrate one so that would mean another buy (ST-E might be a better buy already).
    Moto isn’t using TI in it’s new models this year,last year they used Nvidia and TI,this year they went for Intel and Qualcomm.

    Share
  3. This would be crazy if Amazon did actually do it, as it would be difficult to justify the investment. Amazon would basically be buying IP and a huge team of engineers, which would cost hundreds of millions of dollars each year. The payback? Maybe $5 per device. If they could achieve Apple-like volumes, they could justify it but that’s unlikely. Also, Apple actually makes money on their phones and tablets.

    I know that Amazon say they make money when people buy stuff from Amazon on their kindles, but the important metric is not how much they make on those sales, but the differential between the revenue generated on these devices and what they would have made anyway. It’s not like all those sales never would have happened if people didn’t own kindles; most Kindle owners were probably already loyal Amazon customers. Can they justify $1 billion or so to buy the unit and then hundreds of millions more each pear to operate it (if their savings from owning the SoC IP is only $5/chip, they need to sell a lot of devices to just break even).

    The more hardware Amazon designs, the less they will sell from other manufacturers. They need to understand their real value, which is retailing. Designing hardware and software is not cheap, and is unlikely to be sustained from thin margins on content over a long period of time.

    Share
  4. Amazon start to be quiet popular in Europe and you can find Kindle E reader in public libraries. I believe they will be well in years ahead.

    Share
  5. Amazon Phone may be the ‘prime’ reason. They would have tighter control on their hardware costs. They would probably make money out of the hardware as well (unlike the Kindle series) and yet stay focused on selling content.

    Share

Comments have been disabled for this post