Inphi And Imagination Find Fast IC Sales Growth


I’ve never had much time for this Cloud computing bollox, but it made me sit up to hear the CEOs of Inphi and Imagination saying that this is where the semiconductor industry’s growth is coming from, and where start-up opportunities are to be found.

“Connecting chips to clouds is a hot area to be in,” says Hossein Yassaie, CEO of Imagination Technologies which develops SOC IP,  “there are huge opportunities for start-ups. SOC IP + domain expertise = the next round of successful semiconductor start-ups.”


Inphi, which sells very high-speed analogue components to server manufacturers, is growing a 45% a year. It’s: “In the right place going after the right space,” says Young Sohn, CEO of Inphi, who is as sharp as a whip, a director of ARM, and was at HarvardUniversity with US President Barack Obama.


The driver for Inphi’s business is that customers demand the fastest possible delivery of their information. Apparently, Amazon reckons that every 100ms of latency costs them 1% in sales; Google reckons that an extra 500ms in search page generation time drops traffic by 20%; and Tabb Group, believes that a broker could lose $4m per millisecond if their electronic trading platform is 5ms behind the competition.


The driver for Imagination’s business is that everyone wants locked-in customers. Even network operators have approached Imagination. “Operators wanting to get into SOC designs is quite amazing”, says Yassaie. Why do they do that?  “If you can deliver content directly to the consumer, you can create consumer ‘stickiness'”, says Yassaie.


Imagination’s Yassaie points out that most key SOC functions are now available as proven IP, and mature IP equals low development costs. “SOC IP allows small companies to do big things,” he says.


Inphi’s Young quotes IDC as saying that 25% of all incremental IT spend growth by 2012 will be spend on Cloud computing, and quotes Merrill Lynch as saying Cloud computing will be a $100bn market by 2011.


Imagination’s Yassaie points to the new platforms which the Cloud makes possible:

“The PC platform has been the only universal platform, but that is changing”, says Yassaie, “people are trying to establish that kind of platform in other channels – some with software, like Google’s Android; some with hardware, like Apple.”


So, you CEOs, instead of looking for the latest target for a cost-cutting exercise, get your place in the right space – connecting chips to clouds, and getting content from clouds to customers.


And, for God’s sake, make the air interfaces seamless, simple and solid. “Consumers want everything, effortlessly, and this is what drives our business”, says Yassaie.

TOMORROW MORNING:  Ten Best Reasons For Having Linux On Your Netbook



  1. Yes, DWL, I agree the idea is compelling but it’s the links which spoil it. I think the wireless networks too flaky to make Cloud Computing a reliable way to approach computing.

  2. Rather than being ‘bollox’ I think the cloud idea is rather compelling.
    You only have a certain power budget (~ a few W) at the handset; that has to be shared out around the screen, processor, modem, and memory.
    Given that it’s unlikely that those few ~W are going up by much anytime soon, enhanced user experience at the handset means that the other elements have to give (technology scaling accounted for…).
    It seems that one way out would be a very low power main processor; just capable of servicing the screen, and I/O.
    Pair this up with a(some) sensationally powerful low latency modem(s) and a reliable link to the ‘cloud’ and your device could become a modem with a screen attached.
    Granted there is lots to be worked out, but some smartphone software I have experienced has been so draggy that it sometimes seemed to be talking to the cloud after every command anyway.

Leave a Reply

Your email address will not be published. Required fields are marked *