Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Ashraf, what was the limitation of XMM7260 again? Which carrier won't be supported? Where do you think Qualcomm might still have an edge?
Finally, BayTrail devices arrive in the market. It is a bit irritating that only China white box companies use it but I hope that is because the known brands need more time to develop a quality device. I am looking forward for Computex for Moorefield and BayTrail devices from A brands.
The Moorefield tablet rumour is very encouraging. It may be a perfect fit for decently priced allround devices and LTE as a cheap bundle option from Intel may be very compelling for customers. If it performs decently at standard use, especially web surfing, which it certainly will, it maybe the first real competition to Qualcomm's Snapdragon 800. Hopefully, we'll see phones with it very soon. There's a lot of volume to gain.
It looks very early to me that Moorefield is already available. Seems like Intel is acting much faster than it did before. That's a powerful sign. Let's see what wthdik2's and fastpathguru's posts will look like in a few months ...
Also the Chrome OS were great news. That's a clear indication that Silvermont was a must have for Intel and the losses in mobile are probably worth much less than Intel losing Chrome OS and not having a superior solution for Windows tablets than Windows RT. Intel's mobile investments already pay off.
Probably Intel is uniquely positioned for Chrome OS since it provides a mobile SoC solution with all the interfaces and drivers needed for running a full Linux. Also, x86 is where Linux is coming from, not ARM. That's basically a home run for Intel. Also the low prices won't harm Intel much, since they keep the foundry cut, the graphics and processor IP licenses and I guess Intel also keeps many of the payments needed for all these PC interfaces, like SATA, USB 3 Host, PCI Express and all that stuff. This will be very hard to crack for the ARMy and Intel shouldn't give them any chance to enter this market.
In addition, Chrome OS provides new growth prospects for the declining PC industry. A 2 in 1 BayTrail device for 300$ and an upgrade to Core I3 for 350$ and an upgrade to Core I5 for a total of 400$ would also keep Intel's margins high. They can basically keep the royalty Windows was taking in the past. Goodbye Microsoft, I really hope this is your end.
Intel really needs to work harder on its foundry ambitions. They need this business for their future fabs. I think we will have three high end foundries left in the end. TSMC, Samsung and Intel or Global Foundries. Global Foundries teaming up with Samsung makes it harder for Intel. On the other hand, if Intel delivers for Altera and they are first with 10nm, then they may grab significant marketshare in the fab business which allows them to keep their lead with their next fabs. Apple fabbing at Intel 10nm and significant market share in mobile together with stable revenue from PCs is the bull case. The ones left with high end fabs after a consolidation have the license to print money
Ok, let's have a short reality check!
The stuff you guys are writing about actually already exists, didn't you know? It's called world wide web. The web pages are generated by servers, doing all the hard work like fetching the database, preparing the content and so some calculations etc. The (thin) client, i.e. your computer, only performs the rendering part. Most popular browsers even use the GPU to support a fast rendering of the web page. So, your great vision is already here, today. And? Do you feel like this makes everything faster and removes the need for a faster computer? Do you feel that all this runs much faster today than computers did in the past? Do you think this is the end of computing as it was for the last 50 years with every increasing demand for compute performance?
Well, I don't. My computer has maybe ten times the performance and bandwidth for rendering internet pages than 10 years ago but it doesn't really feel much faster at all. Take website like Gamespot or Bloomberg (or even this one, sorry, awful website Investorhub). They are damn slow, just like the slow webpages in the old days. Yes, there are very fast ones, but they already existed ten years ago.
Why is that? It's always the same spiral you are seeing here with faster hardware generating more demanding software (often without increasing its quality actually). It's not about the software guys in general doing a bad job here. I have been doing a lot of programming and I know how this works. You can hardly finish your project in time so if it runs acceptable on an average machine, you are pretty happy already. Usually there's not much time left for optimizations (which can be a lot of work, it's not always low hanging fruits). If noone forces you by requirements, you're simply happy that it works. Hardware design is much more forgiving in that regard, by the way.
Anyway. Why should this change all of a sudden? Why should a Core I3 be all you ever needed as performance? That's ridiculous and against all the experience I have made with computers the last 30 years or so. Believe whatever you like but I am certainly not basing my investment on such a speculative theory (you know, this time it's different).
Instead of "the PC is dead" it looks more like a "the tablet is dead". Not that great of a form factor at all as it seems.
Man, still no Android on mobile. This stinks. That's about half a year late, this is such a big disappointment. Intel doesn't deliver, not at all, this doesn't convince me about Intel's future in this fast moving mobile market.
For Merrifield it seems that they need XMM2760 before having any chance in the smartphone market. Hopefully it will be cheap enough to find some buyers. If they have a BOM issue in their smartphone parts then I wonder what these ignorants actually thought they were designing. That is just ridiculous.
For BayTrail maybe an LTE modem is also getting more and more important in the tablet space as many new tablets include an LTE modem. 3G (i.e. Clovertrail+) is for the low end.
I really hope Intel will be ready with XMM7260 by end of june and deliver BayTrail and Merrifield, as well as Moorefield soon after that, in high volume. It's really about time for Intel to stick to the clear promises they have given.
Yeah, I know there are some crutches to circumvent these issues but these are definitely no OSs that provide the type of multitasking needed by a server OS, where the discussion originated from.
Mobile OSs are reduced in a specific way and that is done by purpose to reduce the memory footprint and usage of processing power. Actually Apple started that with iOS and that was smart move.
Ironically, the old Windows Mobile, used more than ten years ago, went the exact opposite approach. It used multitasking and never closed applications (unless when it was out of memory). That was due to the slow flash memory that was used at that time. On the other hand, Windows Mobile was a piece of shitty software from Microsoft and pretty slow. Still, Windows Mobile (or CE) wasn't a full desktop OS since it lacked many features and was stripped down.
Despite multitasking, the same is the case for Android and iOS. They have a reduced feature set. There is a reason those fit into around 1 GB of memory instead of more than 20 GB for a full blown desktop or even Linux distribution for servers. There is also a reason why they didn't use off the shelf Linux distributions for mobile, don't you think?
I have been reading an article in an electronics (plain old paper) magazine about Intel vs. Qualcomm. There have been some details stated about where Intel is standing in mobile versus Qualcomm and the others. I will give some extracts from the article, most information may have been already posted in this forum, I don't remember all. The information seems to come from an interview at Intel, probably from Mr. Eul but that's not stated clearly in the article. Expect the info to come from Intel so it may be biased somehow (although the magazine is independent from Intel).
- LTE Modem XMM 7260 will be in mass production end of june latest
- XMM 7260 supports carrier aggregation CAT 6 with up to 301,5 Mbit/s
- XMM7260 supports 30 LTE bands (how many are there overall?)
- Supported standards are W-CDMA/HSPA+/TD_SCDMA/TD_HSPA/EDGE
- XMM7260 is positioned against Qualcomm modem MDM9x35 produced in 20nm (release date unknown)
- According to their information, TSMC is still only providing samples of the new Qualcomm 20nm modem
- The RF frontend chip "Smarti 45" provides up to 23 carrier aggregation combinations and supports the modes LTE FDD/TDD-3GPP Rel.10, InterBand/Intra-Band LTE (whatever this all means), Non contiguous Intra Band LTE and DC-HSDPA/TDA-SCDMA and Edge
- Merrifield (Z34XX) and Moorefield (Z35XX) double the performance and energy efficiency in some benchmarks compared to Snapdragon 800
- Intel's 22nm process delivers 35% more performance/watt than Intel's 32nm process used for Saltwell
- Merrifield is clocked at 1,60/2,13 GHZ and Moorefield at 1,60/2,30 GHZ (Moorefield therefore close to BayTrail)
- Merrifield uses PowerVR G6400 and Moorefield PowerVC G6430, both clocked at 533 MHz
- HD videos are encoded at 60 frames per second, hardware encoder/decoder for Google's VP8, no mention of Ultra HD though (4K)
- integrated video processor for improving picture quality (noise, colors, contrast etc.). Camera with picture in picture and burst mode
- Hardware accelerator for sensors, similar to Apple's M7
- Merrifield Z3480 in WebXPRT 2013 216% performance of Snapdragon 800 and 116% of Apple's A7. 145% better performance in Mobile XPRT than Snapdragon 800 (benchmark not available for iOS)
- Silvermont pipeline uses optimized branch prediction and pipeline
- Intel reference design (4.3" screen) delivers 19.2h in BatteryXPRT compared to 16.9h of Sony Xperia Z1 compact (also 4.3" screen with slightly larger battery capacity than intel reference design)
- Merrifield graphics scores with 66 frames/s in GFX Bench 2.7 Egypt HD Offscreen vs. 69 frames for Snapdragon 800 and Apple A7 with 57 frames/s
- Intel expects 10-12% better GPU performance for Moorefield
- Intel expects 300 design wins this year (generally, I suppose this includes Android and Windows tablets - no more specifics given though)
- Goal to deliver 1 billion SoCs and 2 billion transceivers
- In 2013, Qualcomm had a 97% market share in LTE modems, Samsung 3%
The whole tablet market starts to stagnate. I think we'll have Gartner revise their numbers down soon. Nearly everyone that is interested owns a tablet now and people figure that these are actually only good for surfing the web and also there just for some type of web content and some nice to have apps. Most people, also in the emerging markets (Intel saw some growth there), will need a PC in addition to a tablet. That's not going to change anytime soon. The tablet boom was a one time effect. They will continue to exist as cheap gadgets next to a PC and a phone, replaced as slowly as the PC at home. Intel's low price "strategy" comes at the right time.
The PC business won't automatically benefit from that, it will just continue to grow slowly but steady I guess. At the momemt I think it is better to be invested in the PC business rather than to be invested in the tablet business.
The actual prize to win is the smartphone business. That will remain a solid margin and incredible volume business. Let's hope for Moorefield and Intel's next 4G modem to have some success there.
I actually think a Merrifield dual core with Intel LTE modem would make for a fine midrange phone (that would actually feel like a high end one in terms of speed). Not really sure whether consumers actually care about the number of cores built into their phone and I wouldn't really expect them to be that uninformed - often they read a few lines in the internet before buying something and if there's statet that a two core phone with Intel is running as fast as the fastest ARM quadcore (let alone octacore) ones, they will buy the product nevertheless. The OEMs have to think this way too, though.
Intel may just not be ready yet with Merrifield, be it driver issues or other stuff. Maybe they are working hard on 64 bit Android and this could be an advantage, but that also depends largely on Google when it will be available. Delaying the release of the platform for that reason is a very bad one in my opinion. They could release it with 32 bit Android first and guarantee for a 64 bit Android update later.
Anyway, it feels annoying that we don't get any information about particular design wins. I mean, how can you launch a platform and then not announce any design wins? Intel can't be that lousy in selling their stuff and they also sold Medfield, which was definitely not better at the time it was released than Merrifield is today.
Well, the call wasn't particularly exciting nor disappointing. Fact is that mobile currently is dragging the results down and the 60% share loss is simply due to Intel executing extremely badly. Something that has to be said. If BayTrail would have been available by end of last year including the LTE modem, they would have gained market share instead of losing it. Hopefully it's not too late anymore and the latest LTE modem together with BayTrail Entry will gain significant share in the market, even though that might cost a little (actually the modem could cover some of the losses).
It should still not be forgotten that Silvermont was a must and the PC group greatly benefits from it. Also competitive Windows tablets wouldn't be possible without it so it was a must to develop it and spend the money.
PC client group did ok but there's a lot of one time effects in it, especially the Windows XP replacements and also the next generation consoles are giving a lift for the high end desktop market. This may help us this year, but next year the actual consumer demand must improve. Hard to expect as long as Windows 8 stays the piece of sh.. it currently is (yeah, great, we're given a real start menu with Windows 8.2 - Microsoft really understands what this all is really about ...).
Regarding the future, Broxton clearly is more of a 2016 thing as it seems. At least it is going to integrate the modem part as well but the competition will look a lot different by that time. Intel will compete against TSMC 16nm I guess, where Finfets and higher density will eat quite some advantage of Intel's 14nm process. So their design must be a lot better by then, I'd say significantly better than an Apple A7 quadcore with higher clock - that's close to Haswell territory and the question will be about how much Goldmont actually will differ from the Core-I generation by that time (Skylake that should be). If Intel isn't clearly going to be leading in mobile by beginning of 2016, I'm out, hoping not to have lost too much money on that.
Also no word about Merrifield and design wins. That's really disappointing. Medfield and Clovertrail had at least some design wins. Also no word about why BayTrail Android tablets are that late since they were announced for late last year. Why are those analysts not asking such questions?