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.
Bought some AMD shares on the dip just now. Let's see how this turns out.
I wanted to see some independent benchmarks before considering AMD. I think they have a very competent team together which may be able to even outperform Intel's high end chips with the next iteration. No own fabs dragging them down anymore. By the way, Ryzen in my opinion shows just how capable Samsung's 14nm process actually is. I don't see a lead from Intel here anymore.
I am considering to buy some AMD shares today, after the pullback yesterday. Market cap is around 3.5 times sales, which is a lot, but considering the market potential of the processor, it still seems sensible to me (doubling of sales could be possible I think). Intel trades at a bit more than at 3 times sales (this is a flawed metric but in this case, a quick guesstimate).
I have read some comprehensive reviews of actual benchmark results of Ryzen and the results are really good. It is almost on par with Intel's high end processors. When has that been the case in the past years?
Regarding power, AMD actually is better than Intel, which many seem to oversee easily here. That makes it a very compelling product for servers and also has a big potential for laptops, granted that the right products will be released, which I think is possible since Lisa Su also showed to have a good understanding of the market when she turned around AMDs graphics card business.
For Intel, this may get ugly since they have to lower prices. This time, though, they cannot simply underprice AMD since AMD doesn't own the fabs anymore and I have doubts that Intel can compete in price with the foundries. Even AMDs margins could become better than Intel's, provided the right market volume.
This all could be the start of a next PC boom, which was the strongest when AMD and Intel fought toughest in the past. Maybe some of the PC vendors are a buy now.
AMD has to show that they can turn the momentum into better products, especially for servers and laptops. Designing good laptop processors will still be difficult for AMD, since we know from Haswell that much more than just process tech is needed to achieve this. We'll see how they do.
Ah, regarding the gaming benchmarks: Those are highly tuned to specific hardware, with lots of optimization, compiler tweaks etc. It is not unusual that they have issues with different hardware, no matter how performing it is. That's a matter of time and market penetration of AMDs processors. It's also the same reason that consoles with relative low power AMD hardware can outperform PCs with much higher specs. It's all about optimization.
While Borusa and Andy are patting each other's backs, here is some more information about the actual densities of upcoming processes of the foundries and Intel:
http://www.eetimes.com/document.asp?doc_id=1331185&page_number=2
Samsung already announced 10nm chips, its own Exynos and the Snapdragon. According to the article, this is with bad yields. It also states that Intel 10nm already has better yields than TSMC and Samsung, with TSMC somewhere in between.
Comparing the actual nodes, Samsung 10nm is less dense than TSMC 10nm, which again is less dense than Intel 10nm. From the chart it seems that they are all pretty close to each other this generation. At 7nm, the picture looks different (though I think there's even much more speculation involved), with Intel having its node close to what TSMC and Samsung claim as 5nm. According to the chart, Intel 10nm will be actually the same as TSMC 7nm, with Samsung 7nm being a bit denser than the two (this also tells you what TSMCs 7nm chest pumping is worth). Regarding the actual density of products, that's still a different story, I guess (also depends on tools, cell library etc.).
What is important in my opinion: All the 10nm nodes don't seem to be too far away from each other, with Intel slightly leading but Samsung introducing the first products at this nodes (at bad yields) as it seems. Can we call that an even? If not lost, Intel has almost no lead left. Bye bye the chance of "flooding" the mobile market and positioning itself as a high end foundry with premium prices. Screwed nicely ... Intel 14nm would have been a nice foundry killer, if, just if ...
I have had the privilege to visit a modern fab recently, including its clean room. It wasn't a leading edge 28 or 14nm fab but also not too old, able to process sub 90nm 300mm wafers. What I think to have learned from this visit is that it much more boils down to logistics and automatisation than I have thought before. It is a huge logistic challenge to process all those different products with different requirements, slightly different process steps, layers, quality requirements, testing etc. I think to understand a lot better what Intel faces as challenge when competing against a foundry like TSMC, which is able to mass produce a vast variety of products, from analog large structure ICs, mixed signal middle density to high density logic chips. This is much more of a logistic and automatization problem than it is a "be capable to produce small structures" issue. I have ever higher doubts that Intel has just close the skills and attitude to compete in this space, especially since they lost their process lead. Processing their own highly uniform products is a completely different animal.
Ah, and following the article, it seems to me that trying by brute force to be the leader in process tech no matter what the costs are is soon turning into a looser's game. Maybe the winners will be the ones staying at the nodes where the masses of customers find the best compromises instead of paying big for some prestige products nobody is going to pay the money for. Mask costs at 16nm are said to top 5 million now, compared to 2 million at 28nm and 1 million at 40nm. Sure, costs are expected to get lower with time but those are figures only a certain market share of the foundry business will economically benefit from - at some stage, the total costs of owning such prestige process tech will be higher than the return on that market share that is left to pay for it. We'll see who is acting the smartest from here.
Yes, SSDs aren't really suited for backup purposes since they need to be powerd every once in a while to retain data. If I remember correctly, things got worse with latest generation flash (MLC, 16 nanometer etc.).
Not sure if this would be an issue for datacenters, since they probably could power the disks routinely avoiding that effect. Also a refreshing cycle for all bit cells would be possible. It wouldn't make much sense though, since magnetic discs are the cheaper way to backup data anyway.
For private purposes, a magnetic drive is the backup drive of choice. Remember to buy one before SSDs take over all of the market!
Sold all of my Intel position on monday this week. I don't see Intel bringing the right attitude for the changes that are required to stay on top of the semiconductor business. When was the last time they actually delivered on time? This business is more competitive than it ever was, especially the foundry business, which provides the chip volume Intel so desperately needs. Stratix 10 still isn't shipped and the company doesn't even bother to explain what is going wrong there. With that kind of execution, I have high doubts that Intel will be having any meaningful market share in the foundry space.
Memory and 3DXP are nice but you can have that as a pure play with Micron (which had a very nice run up lately). In addition, I have some doubts about the market potential for a database server niche solution. Make it price competitive with NAND, bring it in SSDs and it could be a big success. As DIMMs? Doubt it.
The mobile business is essentially lost with Intel basically giving up without even further trying. That's billions down the drain. Standalone modems are a thing of the past, so not much for Intel to gain here either.
AMD is likely to have a better competitive position than they had in years. They may not take much volume from Intel but it will have an impact on Intel's margins. Even with a recovery of the PC market, it may not fully compensate Intel's PC business.
Last but not, much of the run up in Intel's share price over the last months priced in much of a recovery in the PC market and the Trump rally isn't here to stay. I see the US stock market as highly overpriced with such an administration which is about to harm the US economy badly. I'm pretty bearish for US stocks for the coming years, at current prices. Time for me to sell US stocks, including Intel.
This is what Nenni claims:
@Borusa, Andy Grave
What market segment can AMD penetrate with Ryzen? It is that of value desktops, right? Laptops, which is the best PC segment in terms of volume and margins, will likely not be a topic for AMD, at least if they don't improve a lot on power consumption and platform design. AMD has always been weak in the Laptop space.
Now, if only desktops are accessible for AMD, there won't be much left to them. First, it is a very small segment outside of business and enthusiasts, both which will be dominated by Intel since they will still provide the best performance, especially single core (enthusiasts) and best platform support (professionals). Those guys won't switch to AMD just to save a buck or two. What is left is value desktops, which is an extremely small segment these days.
Servers may be more interesting but that is also a question of platform design, trust, reliability etc. They may win some prestige project to put Intel under pressure but that doesn't mean it will make them lots of money.
If AMD will have to compete in price, the question is about competitive die size, which is unknown so far. If AMD made this performance improvement by sacrificing die area, they won't be able to price much lower than Intel's processors, making their products unattractive. Don't forget: AMD has to pay the foundries 40% or so margins while Intel keeps this margin to themselves.
I am not saying that Ryzen will be a big fail but it seems to put AMD in the ever same situation of close but not too close leaving them with breadcrumbs which destroys the value of the company. What exactly does make your investment in AMD so good and especially better than it always has been?
Just wanted to write a note on what I read in an article about IoT today. It was more like a summary, where the technology currently stands and what powers are pushing it in which direction, which I found pretty interesting.
Most people simply think: Plug in a modem into a controller, connect some sensors and that's IoT. There are some issues with doing this, such as power consumption, cost (die area of complex modems) but also things you wouldn't initially think of, such as robust operation deep inside buildings, where standard modems have issues with.
To solve all that, the industry is working on new protocols which allow for less complex and lower power consuming modems, which, at the same time, have a more robust channel coding, allowing for communication even in the worst condition. There are basically two streams, where one is based on fully new transmission systems which require a wholly new infrastructure and the other, using existing mobile resources.
The technology based on existing mobile infrastructure basically consists of reusing the existing GSM channel (first generation using it with a (way) more robust channel coding, certainly at pretty low data rate. According to Vodafone, about 85% of their mobile stations can be simply reprogrammed to support this new protocol, which is a big advantage of this solution.
What I think is important is that many mobile providers, at least in Europe, are really pushing this technology. Vodafone actually wants to start its operation in 2017 and already performs some trials in Spain.
What's unclear so far is how the providers will charge for it. Since they push so strongly, It seems pretty clear that the infrastructure will pretty soon be available from their side. It's the semiconductor industry which actually has to hurry this time
The reason why I am posting this simply is, that I couldn't really see which role the mobile providers will actually play in IoT, since I couldn't see breakthrough of it without their participation. As it seems, they are, at least some of them, actually pushing strongly into this direction which makes the time frame for introduction much shorter and removes some major obstacles for a major breakthrough within this decade already.
The market potential certainly is huge, just as it is for embedded systems in general. There are many applications with big volume outside of private use, such as public infrastructure (traffic, water supply, all sorts of sensors you can think of in urban areas). Logistics, car rental and industry, just to name a few. In terms of units, this definitely is going to greatly outnumber todays mobile phone business.
Still, it is in question how Intel can benefit from it, since this will be a commodity business, where performance isn't an issue in most applications and price is everything. It is also a business where many different solutions will be needed for different applications, just as the embedded business looks like today. Not much concentration and not much money to be made.
I still see a way for Intel to make money with this: Solve the flexibility issue with integrated FPGAs (good tools required here) and integrate its modem IP. Money will be made through semiconductor production, similar to the foundry business. Intel should be open right from the start to provide ARM or x86 based solutions. Intel won't make much difference with its own cores in this area anyway and ease of use, tool support and ecosystem is everything in embedded. The only reason to continue developing x86 in this area is to try and starve ARM, so it won't grow too strong with its big cores. Intel could sell its x86 based IoT solutions slightly cheaper than ARM based ones for instance. This certainly is only possible if Intel's x86 cores are competitive in every metric, especially area and power consumption.
As a side note: Did anyone read about the TAG Heuer smartwatch? It is actually one of the few smartwatches that have a classic look (display is permanently on) and Intel being inside doesn't seem to perform badly (though some lags and mediocre battery life have been reported). The watch is very expensive though.
Latest numbers from Samsung.
In the presentation, they break down the memory business and the remaining semiconductor business. From that, one can calculate the Samsung's revenue from its semiconductor business, which is (KRW Trillion) Q2 2016: 3,57 Q1 2016: 3,21, Q2 2015: 2,8. Looks like it is growing nicely. If you take the last quarters revenue from semiconductors, that is 14 Trillion KRW. Converted to USD, that's 12,5 billion USD. That's not the same as TSMC's round 28 billion USD of revenue (taken the last quarter times four), but for Samsung it is at least growing and together with the memory business, Samsung has around 40 billion USD of semiconductor revenue per year, which definitely pays nicely for the 10 billion+ Capex a year for new fabs.
We'll see how TSMC will be doing with such strong competition in the coming quarters. It's going to be interesting.
Oh, they are. You're looking at the wrong numbers. Free Cashflow, taken from Morningstar:
09/2015: 46,768
12/2015: 58,522
03/2016: 82,384
06/2016: 31,505
This is not even because of growing Capex but because of shrinking operating cashflow.
Yes, that doesn't mean that TSMC is already in trouble, far from it. They are losing market share and competition increases, so prices will get under pressure. Apple was able to compensate that a little, for strategic reasons, but that is surely at low margins for TSMC (maybe explains the higher revenue but shrinking cashflow). Apple is a big customer, but not big enough to compensate for all that competition that is coming.
If only Intel would be capable of taking some market share also ...
Yeah, I know, only somehow related to Intel, but Qualcomm seems to also let Samsung produce its upcoming Snapdragon 830 on its 10nm fabs. TSMC seems to be bleeding more and more.
We don't know the actual characteristics and densities of TSMC 10nm vs Samsung 10nm vs Intel 10nm and it is still only announcements, but it seems very likely to me that Intel actually is loosing this race. That's kind of an inflection point, don't you guys think? No process lead for Intel anymore!
For sure, Intel will tell us about how much denser their 14nm process is and that it is actually competitive with those 10nm processes. A matter of believe, I guess.
Google makes a lot of money with advertisment. Basically all their earnings. I suppose you all had this experience already: You search something via Google, e.g. some product and a little later, on some other website, you suddenly get offers for this very product sold by some company you haven't visited its website before. That's what Google sells: Personalized advertisment according to what you seek for, where you are (Google maps) etc.
I find this scary somehow, because Google therefore needs to:
1) Know who you are (some tracking ID I guess)
2) Where you browse the web (Google search, Google Chrome)
3) Where you are located and what places you did look for (Google maps)
I could just not use Google's tools and services but I must admit: They are really good at what they are doing. Google is a great software company when you compare it to Microsoft for instance.
Regarding the cloud business: I remember a report, think it was in the Economist, where they stated that actually Amazon is a bigger cloud service provider than Google. The biggest in the world, actually. They simply offer server capacity for rent at a very cheap price. I suppose Amazon is Intel's biggest customer when it comes to server processors. Google is also at the top, just as Microsoft. The problem is, that less servers are sold to companies and you have a lot of synergies on centralized server farms, e.g. by virtual machines etc. That means that capacity is going away from single machines on corporate level to global server farms with shared resources - less server processors required for this.
I think it is simply the general growth of data at private and corporate level that made Intel's server business grow nevertheless. Some price pressure may occur, once competitors, e.g. with ARM based solutions, can offer cheaper solutions at equal performance. There is still some way to go, but I would call it possible. Investing a lot into porting Linux software from x86 to ARM is not much of an issue for the big ones like Amazon or Google.
As announced by Google, Chrome OS is now capable of running Android Apps. A big advantage compared to Windows, since it allows a seemless interaction between PCs and smartphones. With this, a PC can run everything a smartphone does plus productivity applications. I think this makes a lot of sense.
In the US, IDC estimated that Chromebooks are already selling better than Macbooks. To Intel, this is chance and threat at the same time. Chrome OS could help people starting to buy new PCs again. On the other side, it is fully instruction set agnostic, so ARM has a chance. If Intel continues to lose ground, compared to the latest high end ARM SoCs, it will be very hard to justify those high prices.