Advertisement

Blog

The Focus Dynamic

{complink 1989|Fairchild Semiconductor International} today became the latest in a long line of semiconductor companies to increase its focus on a specific technology or end-market.

“Fairchild is increasing investment in and development of a targeted portfolio of analog and power intellectual property (IP) to support handset manufacturers’ specific signal path requirements for audio, video, USB, signaling, sensing and timing, and in power management for peripheral cores, lighting and RF,” the company announced today.

Chipmakers for awhile have been facing increased costs to develop and bring new technology to market. Former broadline semiconductor makers such as {complink 3538|Motorola Inc.} and {complink 4268|Philips Semiconductors} were spread too thin, and spinoffs have found new life as smaller, more focused specialty companies. (See: Deals I Would Like to See, Part 2: Freescale IPO in 2011.) Others, such as {complink 3614|National Semiconductor Corp.}, are choosing their strongest and/or most profitable technologies and focusing on penetrating the market further and developing new applications.

“The cost to develop leading-edge application-specific chips continues to rise, ranging from $10 million to $20 million,” according to a recent report from {complink 7427|iSuppli Corp.} “Only the biggest and most successful companies will be able to afford such high development costs, effectively squeezing out the smaller companies or the second- and third-tier players.”

Chipmakers' sales channels are adjusting to this model. The distribution channel — which supports both suppliers and end customers — is adding more engineers to its sales and support staffs. These engineers are increasingly specialized, unlike the field applications engineer (FAE) model of old. While engineers are a fixed cost in the channel's sales-driven organization, distribution executives see this as another opportunity to embed themselves more deeply in their customers' design and supply chains. Customers can now choose from the best of the best in technology, but the selection process is more complex. By assisting customers in selecting semiconductor technology, distributors have the opportunity to sell compatible components on the rest of the board.

The broadline chipmaker used to be more of a one-stop shop where product portfolios worked together. Now, distributors are becoming the one-stop shops, and suppliers are moving in the opposite direction. Buyers: How is this model working for you? Does specialization mean you are working with more vendors for the products you need? And where do you turn for help: suppliers, distributors, or somewhere else?

5 comments on “The Focus Dynamic

  1. SP
    November 22, 2010

    The customer always looks upon the chip manufacturer for any technical help or for the intial stages of selecting the product. Big distributors like arrow and avnet are also very helpful. Distributors also do a great job sometimes in putting up with the right point of contact in chip manufacturer. I really enjoy working with avnet and arrow.

  2. Hardcore
    November 22, 2010

    Hi Barbara,

    I think that it will start to go the way of user customizable silicon, there seems to be a very long term trend in this area , going all the way back to  Monolithic memories (late 70's) with their PAL's/ HAL's , which were logic array silicon with 'blow-able' fuses , today we are at the   FPGA (field programmable array) (Xilinx,Actel) covering build logic from the bare minimum to gain functionality and  current trends seem to be towards the PSoC (Programmable System on Chip)(cypress) where you can tag together a number of 'pre-defined' building blocks.

    There will always be a need for Custom Ic's especially related towards   RF amplification/ extraction, but even this is moving into single chipsets that include  CPU/ memory and I/O

    All in an attempt to extend the useful life of a chipset, current issues relate to the interconnection and flexibility of the discrete logic elements , which in-turn  impacts the speed of the devices (anyone can design with an FPGA, but it takes a real pro to clean up the resultant 'net-list' of logic interconnections to gain the required speed from the device)

     

    I see sometime in the future distributors unable to keep up with the functionality of the chips that they sell, purley as a result of the chips-sets turning over so fast and the complexity of hte devices involved.

     

    It would be interesting to know how distribution channels intend to keep ahead of the wave edge.

  3. Barbara Jorgensen
    November 23, 2010

    Thanks Hardcore, this is helpful.

    My guess is that distributors will continue to improve on their programming capabilites, if nothing else. They've been doing ASIC programming for awhile, and depending on how much volume their customers' need, they will take over the programming function of toehr chips. How much this entails, I'm not sure, but I think the majors–Avnet and Arrow– have acquired companies that broaden their IP and programming capabilites.

    Another thing the channel could do is offer more tools on the programming side. I think most of these are already offered by the chip makers, but there might be something the channel can offer targeted to smaller-volume users.

  4. Barbara Jorgensen
    November 23, 2010

    Good news for the distributors.  Hiring engineers is costly, but as long as those investments provide some measurable return–even in goodwill or customer retention–we'll see more design help.

  5. Hardcore
    November 23, 2010

     

    Yes, the big problem with many devices including ASIC's are:

    1. it is mask programable (foundry)

    2. You have to order massive quantities to get the cost down.

    3. it is 'fixed' until the quantity is used (I worked with an 'engineer' that messed up an ASIC, it was not spotted until the final batch was produced, 1,000,000 pcs later we updated the part and it only took 3 years to use the stock…..)

     

     FPGA, PSOC, are re-programmable in some cases in the field, which means that design fixes and enhancements can be added to existing products even if the product has been deployed. Such devices and technology enable reduced reliance on  technology changes or shortages of stock.

    In the case of  FPGA, if you have a product range of 25 different products, then rather than having 25 discrete ASIC devices  you can stock a single  FPGA device then re-configure the device on the production line, which in turn means less lead times for distributors and manufacturers and significantly less capital tied up in potentially redundant stock.

     

     

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.