Seeking Alpha
View as an RSS Feed

Mordechai Rorvig  

View Mordechai Rorvig's Comments BY TICKER:
Latest  |  Highest rated
  • Is Intel's Failure To Hike Its Dividend A Sign Of Things To Come? [View article]
    Alexander, I differ from you on a few points. Firstly, though it is true that INTC is late to the mobile market, you suggest that they will be unable to gain market share, which is simply unfounded and furthermore unsubstantiated anywhere in your analysis. They are in as good a position as anybody to take market share in 2014, with well known mobile product announcements coming in Q4 2013 and onward.

    Secondly, you seem to misunderstand the nature of Intel's capital allocation and reinvestment strategies. While it is true that Intel must reinvest huge sums of money to maintain its edge, these sums allow them to make the same products at better performance, cheaper cost (smaller area), and lower power. If PM could re-design their cigarette manufacturing facilities every 2-3 years to achieve the same outcomes, I am sure that they would also do so. My point is that your critique of this situation, while superficially viable, is not nearly as sensible when you consider what's really going on.
    Jul 30, 2013. 05:52 PM | 1 Like Like |Link to Comment
  • Intel's Dividend Raise: How Much This Year? [View article]
    Bill, nice article, seems pretty reasonable to me. I expect a raise same as last year's, $.015 per share, with the announcement most likely on Friday. While I think your logic and optimism about a rise to meet 4% yield is sensible, I'm guessing management will prefer to be more conservative in this time of transition. But if Krzanich wanted to make a statement, and give a higher raise like your 11% increase to $.025 per share, I think the community would really take notice & show appreciation in terms of volume.
    Jul 24, 2013. 11:17 PM | Likes Like |Link to Comment
  • With Haswell Faltering, Baytrail Is Make Or Break For Intel [View article]
    Ashraf -- Actually, I agree that Haswell is faltering. Global PC shipments are still in decline by 11% as reported for the previous quarter. I suppose it may be premature to say it's faltering before we get the numbers for the next quarters, but unless it turned something like 10-15% QoQ growth for Q3 and Q4, which seems highly unlikely, we will still be down for the year.
    Jul 16, 2013. 02:04 PM | Likes Like |Link to Comment
  • With Haswell Faltering, Baytrail Is Make Or Break For Intel [View article]
    Sneha, I think you make a number of good points here and give a fairly comprehensive summary of the situation. I strongly agree with the point about the problem with the PC market being the lack of PC software innovation -- I wrote an article about this a few months ago myself (http://seekingalpha.co...). If new, innovative PC software could start taking advantage of the available computing power, Haswell would be strongly reinvigorated. As it stands, too much of the computing power requirements have migrated to the server space.
    Jul 16, 2013. 02:03 PM | Likes Like |Link to Comment
  • TSMC, GlobalFoundries, And Samsung Can't Save ARM And AMD From Intel's FinFETs [View article]
    So, your thesis is Intel has better technology, and it took a lot of hard work to get there, and it will be difficult for others to catch up. So what?

    I mean, tell me something I don't know. Everyone knows Intel has a manufacturing lead, and yet it doesn't seem to be helping their stock price. What would be more helpful is if you could assess some of the more detailed questions concerning these issues, such as:

    -How long will it take TSMC to catch up, given that they are behind?
    -What about GlobalFoundries?
    -Why is TSMC showing so much more growth if its technology is this far behind?
    -When will Intel start seeing the stock benefits of its manufacturing lead?
    -And why compare an architecture and processor IP designer (ARM) with the foundries and the IDM, when they are doing completely different work?

    On the last point, I can help clarify this by saying how, to give one example, the architecture designers now must work closely with the foundries to make sure that their architectural designs can actually satisfy known design rules. Thus ARM can no longer function as a wholly independent IP entity. However, without getting into this discussion, your comparison is rather nonsensical.

    I encourage you to get into more specific technical questions which would improve the utility of your analysis.
    Jun 20, 2013. 02:12 PM | 1 Like Like |Link to Comment
  • Intel's 14 Nanometer: Setting The Record Straight [View article]
    First of all, even if the the 14 nm process was delayed a year, Intel would still have a major process advantage on its competitors, who will only be working with true 22/20 nm node scales in 2015, with the possible exception of IBM.

    I disagree with your analysis here -- I think that Intel will still be producing at 14 nm in 2014. I wouldn't read too much into these slides as it's quite likely that the presenters simply didn't want to make any promises that they couldn't keep. This seems to me to be the simplest explanation for the discrepancy. Better to err on the side of being conservative, than to put 14 nm on the slides before the 14 nm fabs have actually gone into production. Even more so if the presenter is someone from Intel Architecture or Marketing.

    And assuming Intel has the fab capacity, then it is much cheaper in cost/transistor to produce at 14 nm than 22 nm, so your logic of "So, from Intel's perspective, why should it release a new chip for a user base when that new chip will not improve meaningfully on any vectors that this user base cares about?" doesn't really play out. Even if the 14 nm process didn't offer significant improvements -- which is overly speculative at this point IMO -- it would be much cheaper for them to produce, so by all means, they would be applying the process shrink to every single product that they could. The reduction in cost/transistor is the main reason to develop the new process in the first place.

    Now, if you wanted to argue that Intel lacked fab capacity to handle its desktop chips in addition to the mobile chips at 14 nm, that might be a valid argument, but based on my current understanding Intel is running at substantial over-capacity right now, and I wouldn't expect that to change so much on the next generation.

    Many sources have indicated that the 14 nm process is in a mature state of development (see e.g. Mark Bohr's interview at semimd.com, or his presentation at IDF 2012), making use of the same Tri-gate design introduced at 22 nm. If it isn't released in 2014, I am confident this is not a serious technology issue, but more likely an issue with cost or general business management factors of which we're not aware of.
    Jun 19, 2013. 02:49 AM | Likes Like |Link to Comment
  • Intel Works To Develop Monolithic Integration Of Wifi And RF, Part II [View article]
    Right, well from what I understand, the issue with the RF360 solution is primarily just cost. Intel has repeatedly said that 3-D packaging, while useful, suffers from high expense. If we could find the cost comparison of the RF360 with the XMM 7160, that would be the best, but as you say, the Qualcomm chipset might not even be released yet.

    This might all be moot though in a year or two, since Intel is making huge strides right now and they might just be able to do that 20-40 band integration all on a single chip. We will know a lot more in 2014 when they begin releasing their Silvermont based smartphone chips and accompanying -- or possibly fully integrated -- modems.
    Jun 10, 2013. 04:09 PM | Likes Like |Link to Comment
  • Intel Works To Develop Monolithic Integration Of Wifi And RF, Part II [View article]
    joeguire73 -- thanks for this detailed and informative comment. As to your question, I think you are correct that Intel will have a notable advantage in programability of its digital radio. Intel recently announced the XMM 7160 at Computex (fact sheet http://intel.ly/19hboO9). It supports 15 LTE bands according to a press release (http://intel.ly/161EYmx), claimed to be the most of any in-market solution, as well as VoLTE, and of course a very small footprint as well. This product was reported to be in final stages of carrier interoperability testing, and readying for implementation in several Samsung products, I believe. What do you think of this?
    Jun 10, 2013. 11:29 AM | Likes Like |Link to Comment
  • Intel Works To Develop Monolithic Integration Of Wifi And RF, Part II [View article]
    DEC_Alpha -- Well, "on or off" is not the same as "1 or 0" I think. But perhaps this is a fair statement. I need to learn more about transistors.
    Jun 10, 2013. 11:21 AM | Likes Like |Link to Comment
  • Intel Works To Develop Monolithic Integration Of Wifi And RF, Part II [View article]
    John -- right, the first thing discussed in [4] after they introduce the transistor is how it can be used in simple amplifiers.
    Jun 10, 2013. 11:13 AM | Likes Like |Link to Comment
  • Intel Works To Develop Monolithic Integration Of Wifi And RF, Part II [View article]
    Glad you liked it, thanks.
    Jun 9, 2013. 10:08 PM | Likes Like |Link to Comment
  • Intel Works To Develop Monolithic Integration Of Wifi And RF, Part II [View article]
    Oh, you're right -- the hyperlink is incorrect. The reference I have been using is the one you said, at http://bit.ly/ZDKGKe. Thanks -- great to hear from another physicist and a past president of IEEE to boot ... incredible!
    Jun 9, 2013. 10:07 PM | 2 Likes Like |Link to Comment
  • Intel Works To Develop Monolithic Integration Of Wifi And RF, Part I [View article]
    Borpet -- right -- thanks for making this distinction clear. I understand you perfectly here. However, I am not sure what is the XCVR -- I take it you mean the transceiver, or RF portion of the chip? Also, I would appreciate your help or guidance in trying to understand what separates or differentiates the baseband from the transceiver.

    And also, you expect Qualcomm, Mediatek and others to eventually follow Intel in trying to digitize these portions as well, right? From what I understand, Qualcomm has perhaps the best radio expertise in the world, and I would be surprised if they didn't see this one coming.

    Finally, also, as your comment "... Intel reported a major technological breakthrough -- it developed an entire Wi-Wi in digital domain only ..." could you refer me to exactly which release or paper you mean?
    Jun 7, 2013. 02:50 PM | 1 Like Like |Link to Comment
  • Intel Works To Develop Monolithic Integration Of Wifi And RF, Part I [View article]
    TechInsider -- thanks, glad to hear that from someone with expertise. Don't hesitate to let me know if I make a mistake.
    Jun 6, 2013. 07:21 PM | Likes Like |Link to Comment
  • The Growing Performance Gap In Consumer Software And How To Play It [View article]
    Jagmohan, thank you for your detailed comment. I had not heard of these terminologies "design productivity gap" and "software productivity gap" before. Evidently what I was talking about in my article was mainly the latter gap.

    I am not sure if your discussion of MP3's versus video is a good analogy here, at least, for the problem I am trying to point out. Neither MP3's or video take full advantage of the computing power available now. What is needed is software which is akin to the cutting edge gaming software of the late nineties, which drives a serious need for computing power.

    Although there may always be some kind of software productivity gap, I do not think that means that software developers should be complacent or not look to use all the available power.
    Jun 6, 2013. 12:46 PM | Likes Like |Link to Comment
COMMENTS STATS
88 Comments
46 Likes