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.
The foundries 10nm process is only about 20% denser than Intel's 14nm which cost-wise can be made up in better yields and Intel's 14nm yield should be quite good by then just in time for Intel's 14nm Broxton/Sofia/Modems to be produced cheaply in volume. Intel's 10nm should then restore Intel's process advantage debuting around the same time as the foundries inferior 10nm variants. The problems Intel has faced at 14nm is understandable considering it is the world leader and 193nm lithography is still being used many years after debut. For a manufacturing guy though BK is more of a marketing BSer than Otellini lol. Like AE said he is unremarkable.
This whole process discussion while interesting is not really relevant in the short to medium term to INTC while the rough process differences remain between everyone. What is more relevant in the nearer term is the threat that Zen might be to Intel's x86 chips. AMD supposedly can get 32 SMT cores on a die in a 2-way SMP configuration which shows it's not a really hot design like the 'Dozy cores. It won't have Intel's single-thread performance but it will be closer like Stars cores were and it looks like they can get more cores than Intel on a die.
AMD now being fabless won't get into any market share wars wanting to maximize profit instead but it might be taking back some of the share it has lost since 2006 anyway just by having a better quality product than the post-Stars crap it has been churning out. The other danger is that a big cloud user like say Amazon goes ARM server in quantity just for performance/cost reasons.
So there are traditional threats to Intel outside of the most common popular fear these days of encroaching ARM mobile chips on advanced foundry processes which has always been overstated and exaggerated.
Don't be so sure, DMRJ will just ask for a lower conversion price as they did when dropping it from the original 26c to 8c. The time for debt exchange was when it was well over $1 as a new buyer may have accepted say 50c on that. This management has been addicted for too long on their regular DMRJ fix and while they have been leisurely conducting their current debt review the pps has halved which already locks in lower priced future dilution. Hard to see where the bottom will be at this rate, 10-20c ? The retail market wants to see debt/equity exchange action now not warm words before they return in mass.
Why should the SP be higher, Management ? The company is still losing money despite getting once in a decade large orders from ECAC and TSA and you still show no real effort in getting rid of the debt ? Current shareholders were expecting a profitable company by now given the size of the orders won and they have indirectly financed the debt up to now at a big loss to themselves through buying DMRJ's new freshly minted shares every quarter for many years now.
This is the typical shareholder-detached attitude/behaviour of management of small-cap companies that don't make a profit and who expect their shareholders to support them indefinitely through dilution. They don't feel your pain as they are still getting paid handsomely every month and obviously don't have the empathy/sympathy for you to do something urgently about it.
3D XPoint Steps Into the Light
As 3D XPoint memory chips move out of research and into the fab, an IM Flash executive gave more details about the novel memory technology, its road map and how hard it is to make ...
http://www.eetimes.com/document.asp?doc_id=1328682
He means this one ...
http://investorshub.advfn.com/boards/read_msg.aspx?message_id=119309090
Agree with your good detailed post and on a more general level your level of analysis and argument has lifted a level since your second bout of participation on this forum. Although I don't still completely agree with everything you say there is now logical self-consistency in your arguments and you are now bringing real value to this forum. FWIW to you I believe you have now reached the high level I originally thought you was capable of. Happy new year to you and the rest of this forum's regular high level contributors.
You keep on talking about fabless as though it had a basic cost.
... which would be the ~40% profit margin in die cost the independent Fabs charge them for producing their dies. As both Samsung and Intel have core business products which fund their Fabs outright this means individual series of products can be produced at raw wafer cost or even less to get market share. Unless the fabless ones have really great unique IP or great wealth, in the long run they are all toast with only Apple, Samsung, Intel and maybe Qualcomm left standing.
Of course it will work when measured over a longer timespan like 5-10 years. Where was Atom in tablets and phones a couple years ago ? Nowhere and now it is a significant player in tablets and an emerging player in phones when most said that x86 could not compete at that level. Once the modems are at 14nm too then Intel will have the integrated products to start capturing the general market from the ground up. Of course execution has been patchy and stop-start but over years the extra development and process potential that Atom has will allow it to catch up and go beyond the ARMy but with a better cost structure than its fabless competition.
but why attack from the low-tier? They're basically racing to the bottom against Spreadtrum/MTK etc...
A bottom from which the others cannot ultimately compete, at least the fabless others so everyone apart from Intel and Samsung. Intel can produce at raw die cost indefinitely until the opposition fades away. Atom also has more raw development potential than its ARM rivals which now look like PC/Server chips from a decade ago. This mobile strategy is like Intel's original Server strategy, a marathon one based on performance/price and will require considerable time to bear fruit, 14nm is when it starts taking shape for the long haul.
They can't sell dreams in 2016 . Typical lame AMD execution.
Looks a decent A57/A72 rival with the accent on performance/power. On first viewing it does not look bad to me but is clearly not excellent either. Of course it does not match Apple's latest Twister but that was always going to be a tough ask for anyone. This will be Broxton's direct rival in the merchant arena and it will be interesting to see how they shape up against each other in all areas.
60 years on, Mainframes just won't die eom
I don't think Abu Dhabi fully realized what a cutthroat ultra-competitive treadmill business semiconductor foundry work was before they bought AMD's FABs. They bought them just at a point when IBM's technology was about to go through a big fail, 32nm, and G-F has never fully recovered with it now basically having to license Samsung's 14nm technology to be competitive. It has basically been one big money pit which allowed AMD to survive but did very little for Abu Dhabi becoming one of the continuous semiconductor power houses like Intel, Samsung and TSMC as they just don't appear to have the staff to generate leading technology on a consistent profitable basis. Hector Ruiz mugged them good and proper offloading his billion dollar lemons to them before it was apparent how worthless they truly were.
ARM have been completely ridiculous for suggesting these ridiculous 20-25% marketshare figures for many years now. I have been saying for just as long that when they exceed 1% it will be a major achievement. They are inferior in performance and performance/power and if they have any advantages in performance/price no-one is interested in the absolute difference in cost.
Bryant said Intel actually went out and counted the number of ARM-based servers deployed today to try to estimate ARM's current market share. This figure, she noted, ultimately "less than one half of one tenth of a percent." i.e. < 0.05%
http://www.fool.com/investing/general/2015/11/23/heres-how-intel-corps-data-center-chief-views-pote.aspx
Or wait for GB4 in the hope it will be better.
There is no excuse for SPEC not releasing the retired SPEC_CPU_2000 for free now as everybody else's attempt at a cross-architecture benchmark are pretty flawed. It's a lot better than GB2/GB3/Antutu and all the others. The lack of a solid free alternative is why the flawed geekbench is so popular.
Indeed, only twice IMSC now . When do you think IMSC will see 1.40 again or even a dollar ?
Perhaps they should change the name of that supposed benchmark to NSPE - Non Standard Performance Evaluation
The IED Isis says it used to bring down the Russian airliner
http://www.independent.co.uk/news/world/africa/isis-releases-dabiq-magazine-claiming-to-reveal-how-it-brought-down-the-russian-plane-in-sinai-a6739311.html
So would an EDS/X-Ray normally pick this up if it hadn't been smuggled onboard ?
Market share between Intel and AMD has historically swung wildly on performance differences of only 10-20% which is the kind of performance increase Intel's process can deliver to A9X which at the moment is at or just below Core-M performance. Beef it up to Core-U performance and different thought processes will go through potential buyers minds. The gain in foundry profit would not be worth the potential a.s.p/brand damage done to Core.
You have to appreciate that Intel basically has a very profitable but always potentially fragile business model, it is continually dependent on combined design/process IA x86 superiority enticing buyers to pay a lot more for its chips than the competition. There is nothing preordained or permanent about this and could come crashing down quickly if ARM or even AMD unveil a killer competitive chip e.g. see the damage Athlon64/Opteron temporarily wrought. Why dull this current combined performance advantage by fabbing your direct competitors ? It's just not worth the extra incremental ~$10 die profit.
So you don't think iPad Pro will take business away from say a Surface Pro or another x86 solution ? You said this earlier about 64-bit Krait/A72 'The ARMy is aiming for "big core" performance in mobile form factors, further driving the disruption of the traditional desktop/notebook by mobile devices' so you are obviously happy to assist this process by fabbing their chips on superior Intel process and making them even faster.
I am sure Intel can hardly wait to build a Core competitor for Apple for just $10-15 foundry profit. This is where this silly fab for ARM idea is going to eventually lead, imagine an A9X at 2.5 GHz on Intel 14nm rather than at 2.25 GHz on foundry process then aimed against $100+ Core-M.
Looks like you might be getting your wish, Intel fabbing ARM chips at least for third parties. As long as Atom is continued with for those wanting IA mobile architecture in all o/s formats Intel may end up having the best of both worlds ...
TSMC's 16FF is like an equivalent Intel 18nm in size and from the looks of it performance too. Samsung's 14FF is an Intel 16nm equivalent.
Really, I can't understand how you can defend what is very clearly a lazy, half-baked effort from Intel with respect to Atom.
Remember it has to double up in micro-servers so performance/power is the overwhelming requirement not absolute performance and it's very competitive in that respect.
http://www.fool.com/investing/general/2015/11/02/intel-corps-upcoming-denverton-microserver-chip-lo.aspx
Broxton will have later ISA extensions as well as SHA ones which should beef up its GB3 benchmark which so many place inordinate store in. Once they also make it triple-issue in the future it will open up a big performance increase, it's there to be tapped and adding back SMT will also be useful. Like I said before I reckon an eight-core Broxton will be very competitive at the top-end and I reckon it will be efficient enough to get away with it i.e. not excessive throttling and they have had enough lead time to knock one up for next year. The 14nm Sofia will have the latest 7460 modem which should make it very competitive especially on price.
So if these ARM chips are going to be sold anyway, why shouldn't they be Intel-designed ARM Chips putting money in Intel's coffers rather than TSMC's?
This is where we differ because I believe a fair proportion won't be sold anyway as Atom will out-price them from the market if not now then eventually. Atom also has far more latent development potential than any ARM core, what dual-issue ARM core is currently comparable to dual-issue Atom ?
I think you are being unduly pessimistic on Intel's mobility future. Once Intel bring 14nm SoFIA LTE to market it will be a serious player in the mid-range skus as only Samsung will be able to match its price because Qualcomm and the rest will have to pay 40% Fab fees on top of die cost to match the prices Samsung and Intel will be able to offer customers. 64-bit Krait or A72 are not going to be worth 40% more in price to a customer. Also Intel can easily knock-up an eight-core Broxton (using its Denverton Server inter-core links) if it wants to go after the flashy benchmark sensitive top-end mobility range, the Chinese love their lucky superstitious eight-cores .
Mobility like Servers will be a marathon for Intel, it just needs to stay in the game year in and year out and grind down its competitors one by one like it has done in PCs and Servers. Supporting the ARM ecosystem by giving up Atom and building ARM chips on its superior processes is not the way to do that as the competitors will then price-target Core having got ridden of Atom. Think strategically in many years ahead not tactically based on this or next year's cpu models.
The ARMy is aiming for "big core" performance in mobile form factors, further driving the disruption of the traditional desktop/notebook by mobile devices.
It's more 'medium core' performance like A72 and Samsung's new cpu, ask AMD how that approach works when taking on Intel's traditional sku portfolio. The only ARM chip that would have a hope of being bought purely on performance is the Apple A9X but that is hiding in an expensive large tablet. The bigger effect will be on Atom though forcing it to stay in the device price range it is now unless Broxton is a big jump.
I think MSFT's new 2-in-1 ultrabook brings some interesting new combinations of qualities to the range although of course it's probably too expensive to sell in massive quantity but for a first effort by MSFT in this category it's not bad and in fact the reviewer says this in conclusion 'superb hardware being let down by immature software' but the latter can be more easily fixed. It's not a game changer but if you consider it as a pretty unique addition to Intel's device portfolio it's good. It's obvious what MSFT are doing here, trying to build a proprietary expensive device portfolio like Apple i.e. Surface/Surface Pro/SurfaceBook but it's not coming out as awful as one would have imagined before the devices actually arrived. None of these devices are bad and are worthy of consideration in anyone's buying thoughts.
Well it's a review of Skylake-U and also the next part of MSFT's hardware story, the Surface Book.
The type of explosives used by the attackers - who were wearing suicide vests - was TATP (triacetone triperoxide), according to the Paris prosecutor.
Sometimes referred to as “Mother of Satan”, TATP is a type of explosive that can be made with easily available chemicals that can be developed in a cost-effective, simple way. It’s also difficult to detect using police dogs and other conventional means.
Not surprisingly, it’s been a favoured choice of terrorists for some time. First developed by Palestinian bombers, it was used as a detonator in 2001 by the so-called ‘shoe bomber”, Richard Reid, a Briton who tried to blow up a transatlantic flight.
http://www.theguardian.com/world/live/2015/nov/14/paris-terror-attacks-live-news-updates-isis-france
Islamic State claim responsibility for the Paris attacks
http://www.theguardian.com/world/live/2015/nov/14/paris-terror-attacks-attackers-dead-mass-killing-live-updates
There was at least one American casualty according to witness reports
http://www.lemonde.fr/attaques-a-paris/article/2015/11/14/daniel-psenny-journaliste-au-monde-j-ai-senti-comme-un-petard-qui-explosait-dans-mon-bras_4809665_4809495.html
This chip is a bigger problem for Qualcomm who probably won't ever get back into Samsung devices now.
Most of the damage was done by suicide (vest) bombers, how does technology handle this particular threat in the long run ?
http://www.nytimes.com/aponline/2015/11/13/world/europe/ap-eu-france-paris-shootings-the-latest.html
Le Parisien has list of the deaths (~150) and injuries at the various attack sites around Paris, which it sources to the French authorities.
http://www.bbc.co.uk/news/live/world-europe-34815972
http://www.leparisien.fr/faits-divers/attentats-paris-fusillades-explosions-etat-d-urgence-13-11-2015-5273837.php
Bataclan: at least 100 dead, seven people in a critical condition, four others injured
Rue Charonne: 19 dead, 13 people in a critical condition, 10 others injured
Rue Bichat: 14 dead, 10 in a critical condition, 10 others injured
Avenue de la Republique: Four dead, 11 in a critical condition, 10 others injured
Stade de France: four dead, 11 in a critical condition, 39 others injured
Rue Beaumarchais: three people in a critical condition, four others injuries
Don't the extended warranties take it up to $25m and Morpho/Smiths used to have continual extensions of theirs over the years. Consumables have got to be a few millions more for 1170 units. I think ~$30m is probably what the order is worth just initially but the bigger the installed base gets the better the ongoing support revenue gets.