You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!

  • Apple Going (IP) Vertical with no Imagination!

    What conclusion could we derive from the recent (April 3) PR from Imagination where we learn that the company “has been notified by Apple Inc. (“Apple”), its largest customer, that Apple is of a view that it will no longer use the Group’s intellectual property in its new products in 15 months to two years time, and as such will not be eligible for royalty payments under the current license and royalty agreement.”?

    Article: Why are AMS designers turned off by Behavioral Modeling?-a9png_575px-min.jpg


    Apple A9 Application Processor

    The first very basic lesson is “good business practices”: it’s always dangerous for a company to rely on a single customer for more than 50% of the revenues. The royalties paid by Apple to Imagination in fiscal year 2015 (ending in April 2016) was US$ 89.7, for a total revenue of US$ 177. If you consider the GPU IP only (without the MIPS CPU royalty revenues of US$ 34.7), Apple royalty payment represent 78% of Imagination royalty revenue linked with the company flagship product… (all figures in $ million).

    The second lesson is “differentiation”: Apple is continuing to deploy a strategy based on the internal development of the essential application processor functions. Since the acquisition of PA Semi in 2008, Apple develops an ARM compliant CPU (under an architecture license with ARM). By the way, the internal development of this CPU has allowed Apple to be the first to launch a smartphone based on a 64-bit CPU application processor. If internal IP development allows such kind of differentiation, that’s a win!
    Just to comfort this second lesson, just take a look at Qualcomm: the fabless chip maker has been, and is still the WW #1 application processor vendor. If you look at Snapdragon architecture, you will see:

    · ARM 64-b compliant CPU (architecture license)
    · ADRENO GPU (internal design)
    · Hexagon DSP (internal design)

    And, if you remember the last time that Qualcomm has integrated standard ARM Cortex 64-bit CPU, it was to react to Apple launch of 64-bit CPU while the internal 64-bit CPU was not yet available, and the result was a Snapdragon product showing such a high-power consumption that it was almost not usable by system manufacturers…

    The big lesson here is that if you want to design the highest differentiated product, you need to invest enough to develop the main IP (CPU, GPU or DSP) internally, the real lesson being that the leaders, system manufacturer (Apple) or chip maker (Qualcomm), did it and it was a strike!

    When I read the PR from Imagination, I am afraid that the next lesson will be “Be prepared to pay a fortune to lawyers, and this for years”. In fact, Imagination is preparing to fight on the legal field:

    “Imagination believes that it would be extremely challenging to design a brand new GPU architecture from basics without infringing its intellectual property rights, accordingly Imagination does not accept Apple’s assertions”. This may be a way to put pressure on Apple in order to push them to pay a license, even if ARM is no more using Imagination GPU… I am not a legal expert and not a GPU architecture expert too, so I can’t really comment.

    Last point (and maybe a business lesson): I can comment about imagination strategy during the last 5 or 6 years. When you look at their annual report, that I did as recently on Saturday (but it was a coincidence), you can see that, on top of the GPU and CPU IP businesses, Imagination is listing 8 to 10 other businesses or product line (like Ensigma IP). When you are the GPU IP #1 vendor like Imagination was, it’s really better to keep focus and stay #1, instead of diversifying into products of low synergy, if any.


    By Eric Esteve from IPnest