Dot Hill, Nvidia analyst conferences

"For every sad fly, there is a happy spider"

Two interesting things happened this week near me.

Dot Hill (HILL) barely met prior guidance for Q2, and investors were hoping a slew of new customers for Dot Hill's advanced storage systems would bring in results well above guidance. The stock took a major hit following the Thursday morning Dot Hill results release and analyst conference (my notes).

The main issue in the quarter was a particular customer, typically Dot Hill's second largest, bought fairly little in the quarter. This customer services large datacenter installations, and revenue variance each quarter is significant. Dot Hill believes the customers orders will resume.

In addition, July bookings have been strong, and 4 customers are likely to release major new Dot Hill based storage systems before the end of the year. Q2 was back-end loaded, so a fair amount of product went out at the end of the quarter that will book revenue in Q3.

So what did I do? I bought more HILL yesterday, basically buying back the shares I sold in March for $5.88 per share for just $3.51. Most of my shares I bought long ago for between $1 and $2 per share.

Of course Dot Hill management could be wrong, revenue and profits might not ramp in Q3 and Q4 and on into 2015. But I think it is likely they will, and Dot Hill will be far more valuable by mid 2015 than it is today. However, keep in mind that Dot Hill has disappointed in the past. Being a small-cap data storage provider is a difficult business.

The Nvidia Q2 conference seemed like it would provide no surprises until, near the end, an analyst asked about the bitcoin computing phenomena that had raised rival AMD's results in Q4 and Q1, but then hurt them when demand dropped off in Q2.

CEO Jen-Hsun Huang said why Nvidia GPU's were typically not used for bitcoin mining, or cryptocurrency in general. AMD's GPUs provide more performance per dollar. Mr. Huang argued extensively earlier in the conference (and in prior conferences) that because of its installed base and superior software, Nvidia GeForce GPU's are gamers' favorites. Maybe so, but it appears to be a well known fact in the computational community that if you want a lot of computation for your buck you want to go with AMD. The reality is that certain workloads do better with Nvidia's architecture, and certain workloads do better with AMD. I see this all the time, even within games, with some games getting better benchmarks with AMD, others doing better with Nvidia, assuming the same pricing range.

AMD, what to look for

Another post published at Seeking Alpha:

3 More Reasons to Stay Long AMD

This would not be the best article for someone who knows nothing about Advanced Micro Devices. But so much has been written about AMD. There have been rumors of an inventory issue (too much of it) with high-end GPU cards. That is supposed to be the result of the end of the Bitcoin mania. So now gamers can actually buy high-end AMD GPU cards at reasonable prices. If there is some excess inventory, it will work out of the system quickly. The only real problem would be if people just plain stop buying high-end graphics cards, which does not seem likely.

There is the PC unit decline (or not) issue, and the ancient market share, AMD vs. Intel issue.

And we are all waiting to see if there will really be a AMD chip in some new mass market device beyond PCs and gaming consoles.

Which will all come out when AMD reports Q2 results on July 17. As usual, I will read and listen to the presentation and put up my notes (at AMD analyst conference summaries). But if you are thinking of investing in AMD, you should listen yourself (the link will be at AMD calendar). Why give the professionals an edge?

Good luck. Keep in mind I am always looking for the long-run value; I don't try to predict stock prices in the short run.

I own AMD stock; I am a business writer, not an investment advisor.

AMD, Intel and Nvidia

My newest at Seeking Alpha:

Diverging Strategies Make AMD, Intel and Nvidia All Buys

Strangely, between the time I wrote this article and it appeared at Seeking Alpha, AMD indicated it would no longer be making new CPU-only processors for PCs. All of its processors models (probably starting in 2015) will have a GPU component.

This furthers my thesis that Intel will be able to increase prices for many of its CPUs. AMD will benefit from concentrating on its APUs (for PCs, which are category killers. It will also be able to put more resources into its server and pure GPU efforts.

I own AMD stock. I reserve the right to buy Intel (INTC) or Nvidia (NVDA) at any time it suits me. Or not.

NVIDIA Takes Up ARMs

Back in February I warned that "at $25.48 [per share], you might want to consider that it could take NVIDIA [NVDA] a couple of years of outstanding growth to justify this price."

Maybe it is just the stock market and macroeconomic turmoil, but NVDA closed today at $15.48.
What I think about the pricing of a stock is only loosely linked to what I think of a company. As to engineering and execution, I have long believed NVIDIA is one of the best technology companies in the world. As the long-term leader in computer graphics chips (GPUs), however, it is now at a crossroads where the old rules no longer seem to apply. It is a time of both tremendous opportunity and tremendous danger, which makes predicting future profit streams mainly guesswork.

NVIDIA reported fiscal Q3 results and held its analyst conference call on November 10, 2011. Note the quarter ended October 30, so its not a calendar quarter and chips going into consumer products for the holidays pretty much had to be out the door by the end of the quarter. Q4, ending January 30th, is historically a slower quarter for NVIDIA. Note also that rivals AMD and Intel had quarters ending September 30.

Revenues were $1.07 billion, up 5% sequentially from $1.02 billion and up 26% from $840 million in the year-earlier quarter. Net income was $178.3 million, up 18% sequentially from $151.6 million and up 110% from $84.9 million year-earlier. EPS (earnings per share) were $0.29, up 16% sequentially from $0.25 and up 93% from $0.15 year-earlier.

That is pretty strong growth over the past year. However, from a licensing deal and litigation settlement with Intel, NVIDIA got roughly $60 million in royalties it did not get a year earlier (the royalties run about another 5 years). That had some impact on revenue and a huge impact on net income and EPS. It is real money, but it did not come from chips sales.

At this point NVIDIA has two games going. One is in its classic discrete GPU chips, which are now also used as math and application acceleration co-processors. Its only real rival in that market is AMD.

The other is combining its graphics engine with ARM chip designs for mobile devices. That results in a chip called Tegra. The first version was interesting but did not generate much revenue. Tegra II sold a lot better, and has powered a number of smartphones and tablets. Tegra III, or Kal-El, became available in the quarter, so a few devices are available now and it will be widely available in 2012. NVIDIA management did not break out Tegra revenue for the quarter.

The NVIDIA vision is not just to dominate the smart phone and tablet markets. Future versions of Tegra are supposed to be powerful enough to go into notebooks, desktops, and even servers.

The problem with being in the ARM market is that anyone can license the ARM design. Graphics and cell phone connectivity chips and designs are available as well. NVIDIA seems tight with Google, but any ARM-based design can run the Android software.

Competitors each have some advantages. Apple, of course, is the perceived frontrunner. There is no guarantee that the iPhone is ultimately going to be defeated by Android-based phones or less likely competitors.

When it comes to ARM based chips for tablets and phones, each competitor brings some serious advantages to the court. Qualcomm has far more extensive experience in cell phone chips than NVIDIA does; so does Texas Instruments. Among a host of other contenders, Marvell (MRVL) should be noted, since they generate a lot of cash each quarter and have a lead in China, a much bigger market to fight over than the U.S. market. Then there are the Koreans, and Japanese, and numerous small innovators.

Another problem for NVIDIA is the lack of a x86 CPU chip design. AMD and Intel both are incorporating graphics into their base designs, eliminating the need for NVIDIA's GPU chips for most users. This already appears to be hurting them in the notebook segment. It will take some time and at least a couple of die shrinks before the pure GPU chip dies (except in specialty markets), but that time is coming at a fairly predictable clip.

So far profit margins have been good at the leading edge of smartphone and tablet processors, but at some point pricing could become more competitive.

2012 is likely to be a make or break year for NVIDIA. If they are able to dominate the Android phone market they might become the dominant semiconductor company, a position Intel has had for a couple of decades now. Myself, I am not willing to make that bet, but I'll keep a close watch. However, I do believe NVIDIA's price is now pretty reasonable, everything taken into account.

See also my Q3 NVIDIA analyst call summary.

Disclaimer: I have no position in NVIDIA and will not take a position for at least 2 weeks following the publication date of this article.

Keep Diversified!

AMD at the Earnings Crossroads

AMD (Advanced Micro Devices) has only two competitors in its niche: Intel (INTC) for CPUs (computer processing chips) that run x86 software and NVIDIA (NVDA) for GPUs (graphics processing chips). How much market share it takes in the PC chip market, and what margins it receives on the chips it does sell, determines its levels of revenue and profit or loss.

Historically, while AMD has been innovative, it has come in a far second against Intel and NVIDIA. In the last two years it has lost ground to Intel and gained ground from NVIDIA. The picture has been complicated further by the emergence of ARM architecture based processors as the preferred basis for smaller mobile devices like smartphones and tablet computers.

After years of development (usually corresponding to quarterly earnings losses) this year AMD is selling chips that combine a CPU and a GPU. Intel, also, has appended graphics to its new line of CPUs, but their chips are remarkably inferior, incapable of running the current Windows graphics standard, DirectX 11. As a result AMD has been selling all the Fusion chips it has been able to make.

Why then, the lack of excitement and lack of upward momentum in AMD stock? Today AMD closed at $6.92, well off its 52-week high of $9.58 and with an astonishingly low P/E ratio of 6.4, the kind you would expect from a declining industry stock.

For the moment the most visible cutting edge technology is in smartphones and Apple and Android based small tablets. That pretty much sums up tech investor thinking about AMD: that a tidal wave of 7 inch screens are going to replace PCs, including both notebook computers and desktops that can run 60 inch displays.

Let's say you have discovered the limits of small screen computing and think there is still life left in the larger form factors. How should AMD be priced then?

First—even if the economy lags, even if consumers are careful with their holiday electronics purchases, even if the economies of India and China don't grow quite as fast in 2011 as they did in 2010—in Q3 and more so in Q4 AMD will get a significant boost in profits from its new Bulldozer CPUs for the server market. They began shipping in quantity earlier this month, with most of the early allotment going directly into the supercomputer market, where they will replace, or fill empty slots in, the prior generations of AMD Opteron processors. Profit margins are better for server chips than for PC chips. AMD has lost a lot of market share to Intel in server chips these last five years. The new chips should help regain market share. They have a different architecture than the Intel chips, and hence are very cost effective at certain workloads. Bulldozer is not a conquer-the-world chip, but it will keep AMD in the most profitable part of the server CPU game.

On the down side, there are so many rumors about yields (% of good processors on a die) being poor for the Fusion chips, that I think it is fair we can treat the rumors as true. At the next AMD analyst conference there should be a question about that. At the Q2 conference the closest answer we got was that margins were good on the Fusion chips. If both are true, and AMD was right about 2nd half margin improvements, then what we have is upside potential. Yields usually improve over time; if margins are already good, they should be great when yields improve. The problem was doubtless forging the CPU and GPU on the same die; traditionally these chip types used different silicon technologies. Bulldozer yields are rumored to be good, but then these server chips don't have a GPU component.

For now I would take Q3 guidance as a fair range. The economy might push revenues down, but yield improvements could push margins up. Guidance was for Q3 revenue to increase 8 to 12% sequentially. Note that because of holiday demand, Q3 is typically the strongest quarter for AMD.

The numbers, when reported, give us hard data, but the technology trends rule long-term value. I think AMD (and for that matter Intel) are over-discounted. I think both will be taking market share in the tablet market in 2012 and 2013. I think the PC market will stay healthier than most pundits predict. Consumers and businesses who skipped a desktop or notebook upgrade to buy a tablet and smartphone will get back on the upgrade cycle.

The combination of full-powered GPUs and CPUs on a single chip may be more revolutionary than the smartphone. Essentially, we are introducing desktop (or even notebook) parallel supercomputing. We are just beginning to see software applications that utilize either a CPU plus separate GPU or the new Fusion chips. So watch for companies like Microsoft, Adobe, and Autodesk, as well as lesser-known companies and startups, to take advantage of this new paradigm.

Disclosure: I am long AMD.

AMD Key Question

AMD had a big jump in stock prices Friday, after reporting Q2 results after the market closed on Thursday. Q2 was not a stellar quarter, but AMD reported that its Fusion line of processors that combine a CPU and a GPU on single chip are ramping nicely. OEMs like them and consumers like them, so the demand is there. Intel, AMD's far larger rival, cannot match them. So it predicted a much better Q3.

If good time were more obviously ahead AMD's stock price would be far higher. Assuming AMD continues to release new Fusion projects, as well as its high-end CPU Bulldozer chips, on schedule, we still can't assume AMD will pick up appreciable market share from Intel.

Intel's law breaking, monopolistic practices days might be behind it, but it still has oodles of money it can throw at problems. It's problems are multiplying, to be sure, but it has a mountain of cash and a cascade of cash flow, unlike AMD.

An analyst at the Q&A part of the Thursday conference asked if Intel is going to lower its prices to compete with AMD Fusion. Let me quote my own summary of AMD's Q2 2011 conference, which is probably not an exact quote of the question and response:

Q. In the past when AMD has done well, we have seen price aggression from Intel. What are you seeing now?

A. It has always been a competitive market. The strength of our products will bring our plans to fruition.

In other words, not so much price aggression so far, but AMD is aware of the potential problem.

I believe Intel is somewhat constrained in its ability to lower prices. This is because investors are worried about Intel. Intel has about 80% of the market. To significantly lower prices on 80% of the market to keep AMD from gaining a percentage point here or there would cause Intel margins to drop, endangering the very cash flow that is such a competitive advantage. Also, the share AMD can take is limited by its production capabilities, which are very limited compared to Intel's.

Instead, Intel will rely on its marketing muscle. It will spend more on advertising, expecially the kind of "Intel inside" deals that convince retailers to promote computers with Intel CPU's instead of those with AMD Fusion chips. Intel will also keep trying to play catch up in the graphics arena. Intel is an entire generation behind AMD (and NVIDIA) in graphics technology, but they have been doing a good job in catching up, including by licensing technology from NVIDIA.

So again we have a situation where AMD has a window of opportunity, which should last until about the end of 2012. AMD will still be ahead in graphics and in integrating graphics with general cpu technology at the end of 2012, but going into the year 2013 Intel's combined GPU+CPU chips are likely to be good enough.

To really compound investor value AMD has to do great in 2012, generating enough money to keep up R&D and start exerting some advertising muscle of its own.

As far as the stock price goes, the Fusion chips are nice, but the real question is whether AMD will be able to pick up share in the server market with its upcoming Bulldozer based offerings. We won't begin to see if that is happening until we get results from Q4 2011 in January of 2012.

I own AMD stock, and I am cautiously optimistic, but I know how hard it is to compete with Intel's marketing machine, no matter how good AMD's chips may be.

See also amd.com

And my AMD summary page

Keep diversified!

AMD Fusion Upside

AMD (Advanced Micro Devices) recently revealed that it has been unable to meet the demand for its first Fusion processors. They call these APUs, for Advanced Processing Units, to indicate that each chip includes both a CPU and a GPU (graphics processing unit). Introduced in late 2010, the first APU variety was designed to bring low cost, high quality graphic capabilities to inexpensive notebook computers. Computer makers (HP, Lenovo, Sony, etc.) embraced these first Fusion chips as a big improvement on the Atom chip from Intel, famous for underwhelming netbooks (sub-standard but very portable notebooks).

Does that mean the sun is finally rising on AMD investors (that includes me)? Intel remains not only holding most market share, but dominating the most profitable segments of the market, including chips for server computers. On the other hand Intel has its own problems, pressed from below by ARM architecture based chips in smartphones and tablets, while failing to be able to deliver the graphic quality consumers now expect. To make a workable business or consumer computer manufacturers need to add an AMD Radeon or NVIDIA GPU to the system.

Intel itself is not exactly the darling of Wall Street anymore. Today Intel is trading at 10 times last year's non-GAAP earnings per share. No tech bubble there. AMD is trading at 8.5 times last year's EPS. Intel is wading in cash; AMD has a substantial amount of debt. If you wanted to buy a computer processor manufacturing company, Intel would be the more obvious choice.

To break out of its trading range AMD, at the very least, would have to start generating substantially more earnings than it has lately. What are the chances of that, when the sector itself appears to be in trouble, and Intel has a tradition of crushing AMD whenever Intel's market share starts to slip?

It really does come down to Fusion. Intel is scared of Fusion, so it is using considerable resources to catch up in graphics technology. Intel is now paying NVIDIA to use its intellectual property, which probably includes graphics designs. That should start to show up in Intel chips in 2012. My guesstimate is that by 2013 Intel will have closed the graphics gap. So AMD's future profits, and the value of its stock, highly depends on how much market share it can pick up in 2011 and 2012, and at what kind of profit margins.

It came as a surprise to me, and probably to everyone, that AMD became supply-constrained in Q1. I think the problem is that the first Fusion chips were on the 40 nm process, which is where most graphics chips are at this year. Both NVIDIA and AMD reported supply constraints at 40 nm in early 2010. AMD failed to anticipate demand for its new chips, and so did not book enough capacity in advance. AMD now contracts for fabrication of its chips, whereas Intel has its own fabs. On the April 21 AMD analyst call for Q1 2011, AMD executives talked about insuring future fabrication capacity, so that must be about the Fusion shortfall.

Right now AMD is about to launch its second round of Fusion processors, which will be more powerful, but also will require more power. So they will be for higher-end notebooks and for desktop CPUs. The will be on the more-advanced 33 nm process. That may have its own issues, but it won't run into the 40 nm roadblock. Later this year AMD will also introduce its Bulldozer CPUs, which are not part of Fusion (they don't have graphics processing on the die).

The upside for AMD investors depends on continued support from manufacturers (HP, Dell, Lenovo, etc.) and from retailers like Best Buy. The near-universal sentiment now is that consumers get more for their money with AMD based computer systems, particularly when it comes to graphics capabilities. Thus AMD should pick up market share this year in the sub-$1000 categories of notebook and desktop computers.

One commonly told story is that consumers are moving to tablets, so sales of PCs and notebook computers are about to come to a grinding halt. The actual evidence for that is scant. In the U.S., prior to the introduction of tablet computers (they've been around for nearly a decade, but did not become popular until Apple marketed a more attractive, if less powerful, version), everyone pretty much had at least one computer, either a desktop or a notebook, maybe a netbook. Most moderate to heavy users had a PC at work and a desktop and either a notebook or netbook for personal use. As the new thing, tablets are going to sell, and with budgets tight, they are going to cause purchases of notebooks and desktops to be delayed. But most tablet computer users have found they are not really a replacement for their more powerful cousins. So, the question becomes: what do they upgrade next? Are they going to get an even newer tablet, or are they going to refresh the old desktop or notebook computer?

You know what the gadget guys will have. A good, reliable, high-performance desktop, a good notebook computer for working away from home, a tablet to be cool and consume video, and a smartphone.

AMD is not going to compete in the smartphone space, but their Fusion chips are beginning to appear in Windows-based tablets. Their stand-alone graphics chips are now used by Apple for its desktop computers.

Also, this is a global market. The same low-cost high-graphics capability that appeals to U.S. consumers will appeal even more to first-time PC buyers in India, China, Brazil, and developing countries.

A recovery of market share in the server GPU market would also benefit AMD, but I'll leave that topic to a later story.

Any reasonable investor not already holding AMD should probably take a show-me the market share and EPS gains attitude. On the other hand, there is a good chance that 2011 will be a year when those of us who already hold AMD will see the upside of Fusion. Maybe as early as the report on Q2, if production was ramped for the new Fusion products.

See also:
AMD home page

NVIDIA Explains Smartphone and GPU Strategies

The NVIDIA (NVDA) analyst call on Thursday was characterized by a relatively short presentation and long question and answer session. Analysts pressed a variety of concerns while management explained their view of the technology space and their strategy for increasing revenue and profits. [See also call details at Nvidia Q1 fiscal 2012 analyst call news summary.]

The explaining was necessary because revenues were $962.0 million, up 8.5% sequentially from $886.4 million, but down 4% from $1.002 billion in the year-earlier quarter. GAAP net income was $135.2 million, down 21% sequentially from $171.7 million, and down 2% from $137.6 million year-earlier. GAAP EPS (earnings per share) were $0.22, down 24% sequentially from $0.29, and down 4% from $0.23 year-earlier. Those numbers include a $40 million payment from Intel for intellectual property (really, to keep NVIDIA a strong ally against AMD); take that out, and the y/y comparisons look worse, while Q1 last year was nothing to shout about either.

So what are the issues? NVIDIA is exiting the motherboard chip set business, which originally became necessary because of a dispute with Intel. AMD has cut heavily into NVIDIA's former dominance in the discrete GPU (graphics) chip and card business. NVIDIA's supercomputer graphics business actually shrank a bit, though management argued it would expand down the road.

That leaves the savior of the moment, Tegra, a chip that runs cell phones and tablets. The newest one is Tegra 2, which everyone admits is a big improvement over the original Tegra. Tegra chips generated $122 million in the quarter, which is a lot of money, until you compare it with losses from the humble chip set business, or the bulk of NVIDIA's revenue, or smartphone chip revenue of some competitors.

NVIDIA claims they are going to take market share in the discrete graphics chips for notebook computers segment. Their only real competitor, AMD, says the same thing. But the good thing about discrete GPU chip competition is there are just those two players. The competition between them is intense, but at least it is somewhat predictable.

In smartphones and tablet computers, however, almost everyone is licensing the core processor design from ARM and adding graphics, Wi-Fi, and cellular modems as best they can. The competition is multifold. Qualcomm was the pack leader two years ago, and still outsells NVIDIA heavily. TI is very competive and believes it now is closer to the heart of Google (maker of the Android smartphone operating system) than NVIDIA, 2010's sweetheart. Apple makes its own ARM based processor for the iPhone and iPad. If they all disappeared there would be Marvell, which dominates the hard drive chip sector and has slots in a number of phones, including the new OPhones in China. There are other players in Korea, Japan, and China. Then there are the non-ARM entries, mainly for tablets but eventually for smartphones, from AMD and Intel.

And anyone can license the intellectual property to make a smartphone chip. In other words, there could be more players in 2012, not less.

Which is too many players. Consolidation will take place, probaby around 2013 when smartphones have completed replacing the bulk of not-so-smart cell phones. While the competition so far has been on technology, with points for speed, usability, and low-power consumption, at some point price will become an important issue. Profit margins will be squeezed.

I certainly believe NVIDIA is as competitive as any of the other companies I named. But there are going to be losers. There will have to be losers. It is next to impossible to predict who they will be.

Given that, investors might want to think about going lower down the food chain, to the companies that make semiconductor manufacturing equipment like Applied Materials or printed circuit boards for smartphones like TTM Technologies.

Of the companies named above, I currently own stock in Applied Materials, TTM Technologies, Marvell, and AMD. In the past I owned stock in NVIDIA.

See also NVIDIA

Keep diversified!

AMD Llano APUs Shipping

In a analyst call today AMD mentioned that Llano APUs (Advanced Processing Units) are now "shipping for revenue," and should be appearing in computer systems later in this quarter (by June). This means that samples went out some time ago, sample systems were built, and the various computer makers are now moving to volume production.

Llano is interesting because it is outside the Bobcat and Bulldozer core paradigm most associated with AMD's Fusion program. Like the chips based on Bobcat and Bulldozer CPU cores, Llano also integrates a pretty high-end graphics processor (GPU) on the chip. However, Llano uses an updated Athlon (K-10) core. For most computer users that will make for a mean machine. Intel, AMD's bigger rival, is behind AMD in graphics technology. For most users graphics is now the bottleneck. Games and video need strong graphics processing, as does almost any content creation work. Llano-based computers should be very fast, very good with graphics, and very inexpensive.

For AMD investors the good news is that the profit margins on Llano chips are going to be better than those of older AMD chips. In effect, as older non-Fusion chips are phased out, everyone gets upgraded graphics computing ability. Gamers and other high-end users will still want to add a discrete high-end graphics card, but your everyday computer will have very good graphics capability without the expense of the separate card. In return the chip will be somewhat more expensive, and show a better profit margin. The impact on Q2 2011 will probably not be great. The computer industry is fairly seasonal. In Q3 there will be the usual seasonal ramp in production, getting ready for back-to-school and holiday shopping. OEMs will have Llano-based computer production in full tilt by then. Last-year's computers will be on fire sale.

Keep in mind that this puts what would have been, a decade earlier, a personal supercomputer on everyone's computer. These new PCs will have way, way more computing power than tablet computers based on ARM chips. Software makers are going to be able to do amazing things once they can count on APUs with parallel processing capabilities that can be used for many applications besides graphics. Many of these applications will make big computer screens even more desirable. Portability is great, but there is still a lot of future in big screens tied to truly capable computing machines.

Of course, rival Intel has a lot of marketing muscle. But the graphics capabilities of its new chips are seriously deficient, not even able to run the DX 11 graphics standard of Windows 7. That means for a good video or gaming experience anyone buying an Intel-based machine will also have to buy a graphics card based on AMD or NVIDIA graphics processors. Most consumers may not understand that, but the big OEMs do, and some of the sales people at Best Buy seem to understand that as well. Interestingly, Apple is currently building machines with Intel CPUs and AMD graphics chips, but I would not be surprised if Apple introduces Fusion based machines some time in 2012. Once the graphics leader, Apple can't afford to fall to far behind Windows in this race.

See also:

My main amd page

http://www.amd.com/

AMD Llano demostrated [October 19, 2010]

AMD: Fusion Without Dirk Meyer?

I admire Dirk Meyer, who recently resigned (under pressure) from computer chip maker AMD (Advanced Micro Devices). I also admire Hector Ruiz, his predecessor as CEO. But both men apparently had their flaws as AMD's leaders, and if you want to value AMD stock appropriately, you should understand the dynamics of the situation.

Dirk only served as head of AMD for about two and one-half years (he was promoted in July 2008). They were very eventful years. Notably, AMD became a fabless company by spinning off its foundries to GlobalFoundries, and the first models of the Fusion line of combined CPU/GPUs on a chip (APU, for Advanced Processing Unit) were released.

The problem is that the Fusion concept was in place when AMD bought graphics chip maker ATI back in 2006. Note it is now 2011. It took AMD five years to put a graphics processor on the same chip with a CPU.

That was simply too long. Rival Intel, a much larger company, brought out an admittedly inferior set of chips combining CPUs with GPUs this month. But that is probably good enough, since Intel can heavily out-advertise AMD.

Imagine, now, that the first Fusion chips had come out in the summer of 2010. It would be a whole new ball game. Intel would try to stall their OEM partners, but it would be a hard sell because they would have to get OEMs and consumers to accept an inferior product late. A six month lead in the computer industry can be an enormous advantage, as AMD showed by bringing out DX-11 capable discrete GPUs about six months ahead of rival NVIDIA, which had been the dominant discrete GPU company, until then.

While no one would say getting out a new line of chips is an easy task, the blame for the delay really has to be taken by Dirk Meyer. This is the second time in the last decade this has happened. Mid-decade AMD had a temporary advantage over Intel in the server market with its Opteron chip. Dirk was then chief operating officer, and when Opteron went from single to dual core, there were serious delays. Those delays enabled Intel to introduce new products and beat back the Opteron challenge. AMD has never been able to regain the market lead in server chips, although its newest 12-core processors are much better than Intel's at certain tasks.

But what investors really want to know is, what is next? Yesterday's call (see my AMD Q4 2010 analyst conference call summary for details) demonstrated that the board of AMD is looking for faster execution on plans to bring out new, specialized processors that can better compete with Intel in certain markets. The typical press view, and Wall Street sell-side analyst view, is that it is about AMD pursuing the tablet computer market. That would be only one small facet of it.

The new Fusion chips run the new DX 11 graphics standard. The Intel chips can only run DX-10 (it is fair to say Sandy Bridge is instantly obsolete). But Intel just made a deal with NVIDIA that doubtless allows it to import their DX-11 designs. How long will it take Intel to move that to silicon? Probably not that long.

Which means AMD has to execute faster. As the smaller, underfunded company, it has to stay ahead technologically and offer a good value proposition to OEMs and end consumers. AMD needs to deliver on the promise of Fusion and perhaps APU's that incorporate other special functions. Maybe even cell phone technologies.

It is a big task, and it needs a leader who can speed up the ball game without causing any fumbles. As we saw in the last decades, fumbling when up against Intel amounts to losing the game.

Dirk Meyer would be a superhero if he had gotten Fusion into consumer laptops in time to make them the mass-market choice for the 2010 holiday season. Instead the products are going to start ramping in Q1, which is a slow sales quarter, and have to sell against the confusion Intel is trying to create in comparing the two technologies. It is going to be a hard sell.

See also:
AMD main page
AMD Fusion
my other AMD articles and conference summaries

Nokia N8: CPU, GPU Roles Shift

Most days I check Anandtech because the staff there does an excellent job of reviewing new computers, smartphones, and peripherals. By excellent I mean they actually give details about the software and chips used in the devices, and also run benchmarks. Today the only chip stocks I own are AMD and Marvell (MRVL), but my customers own (far larger) positions in other technology companies, so I follow the industry as best I can. Today's review of the Nokia N8 smartphone should be of particular interest to technology investors since it illustrates a fundamental shift in computing technology.

The N8 is "the first Nokia phone to have a discrete GPU." GPU is Graphics Processor Unit, as opposed to the more general purpose CPU, Computer Processing Unit. Smartphones are now expected to have excellent displays, which is hard to achieve without a GPU assist. But for years now AMD and NVIDIA (the rival makers of high-end GPUs for personal computers) have been talking about how computer work loads are starting to shift from being CPU-bound to GPU-bound. This is not just because of the need to have large, detailed, rapidly changing displays for gamers. It is because many ordinary computing tasks can be done faster if they are broken into parallel processes that accelerate results.

The Anandtech N8 article is the first time I have heard someone say the transition has already been made in an actual device (aside from professional video content creation machines). "You see, pretty muhc everything in the N8 runs around the BCM2727 media processor. I would hazard a calculated guess that appart from lightweight low-level OS functions and interfacing with the baseband and other radios, there isn't much else for the CPU to do on the N8."

I think that in effect the Broadcom (BRCM) GPU is acting as a DSP (digital signal processor) in addition to doing graphics processing, leaving the CPU little to do. Note that the 2727 (like AMD's new combo CPU/GPU chips) can output 720p HD video through an HDMI port.

NVIDIA has tried to place itself at the center of the GPU revolution, but it's anybody's game. There are two fronts right now: graphics for notebooks/PCs and graphics for cell phones/tablets. The graphics for smartphones can't display games on big screens at a high frame rate (which smooths the action) yet. But in 5 years, maybe sooner, they should be able to do that. The big differentiator is that anything that runs on batteries has to do its graphics work with minimal watts, but a machine plugged into the electric grid can do a lot more a lot faster, using a lot more watts (say, 200 watts versus 1 watt). NVIDIA's strategy is to keep rolling with its high end discrete graphics units while rolling out mobile combined CPU/GPU chips based on the ARM architecture. These would integrate its current offerings that work with ARM, but with the GPU on a separate chip.

AMD's strategy is called Fusion, which combines CPU and GPU on a chip based on the 8086 architecture.

Intel's strategy is to advertise, and hope that the reviewers dependent on its advertising revenue do a good job not mentioning that its current generation of combined CPU/GPU chips, code named Sandy Bridge, are instantly obsolete because they are a generation behind both NVIDIA and AMD in graphics capacity (ask for DX11 capable computers, and you have eliminated Intel, which had trouble implementing the now ancient DX10 standard) and in low-power consumption. Also, Intel is using its vast resources to get back into the ARM-based architecture. They sold their ARM mobile chip unit to Marvell Technologies a few years back. Marvell has since become one of the bigger players in ARM devices, most notably with its chips inside some Blackberry devices and the XBox 360 Kinect.

Then of course, in addition to Broadcom, we have Qualcomm, Apple, TI, Samsung, etc., scrambling to combine graphics and CPUs into single chips or small chip sets to power smartphones and tablet computers.

Software programming is changing too. Want a job? Show you can recode older non-parallel software for parallel processing on GPUs. Lots of shortages in that department.

For investors, you might want to buy a piece of every company in the race. I am certainly not certain who will emerge a winner. On the other hand, there are no pure plays here. A victory in smartphones would add very little to Intel's fortunes, or Samsung's, but it is essential to Qualcomm. Broadcom itself has a diverse set of chip products that cover Ethernet, set-top boxes, and Wi-Fi, among others.

See full Anantech Nokia N8 Review

NVIDA Hopes for Tegra processors

Graphics chip maker NVIDIA (NVDA) has had its ups and downs these past few years, and not just because of the recession. Increased competition from AMD in the discrete graphics card market has hurt. So has the accelerating loss of its "chip set" (chips for integrating a cpu to other components) business. At the very high end NVIDIA is getting traction for its Tesla computational chips, but those revenues are not likely to make up for other losses. It is possible NVIDIA will do better in its competition with AMD in 2011, but it is telling financial analysts that Tegra chips for mobile devices are its next big thing.

At its November 11, 2010 analyst conference call and report on its third quarter (Q3, ending October 31, 2010), Tegra was touted as a replacement for lost chip set revenue and more. Note revenue were down 7% y/y to $843.9 million in a period when most semiconductor companies ramped revenues by double digits.

Note also that Tegra has been around for a number of years. What we have been promised in 2010 is a new, improved Tegra, with an improved software stack make the new devices using it something device makers feel can compete with the iPad. This has resulted in a least a half-year delay in releasing product; the products are largely a 2011 story. Management believes that touch-based systems are going to wipe out older systems, and that Tegra will make NVIDIA a serious player in the field.

Maybe, but I am not the only analyst who is a bit skeptical, and with good reason. Tegra 2 might be improved enough to be a revolutionary epicenter in 2011, but it might get lost in the forest of competing platforms. Essentially Tegra combines an ARM CPU with a GeForce GPU, in much the same way that AMD's Fusion chips combine an 8086-based CPU with a Radeon GPU. ARM is a low-power architecture that is being widely used to address the mobile device market.

The problem for NVIDIA is that lots of companies are selling ARM-based processors for mobile devices. Apple designed its own. NVIDIA's advantage would be its graphics technology, but it is not yet clear how much of an advantage they really have. Many companies are using graphics chips, or integrating graphics architecture onto a chip, that was designed, like ARM, to be low-power from the ground up.

The history of Tegra is not one of blazing success. Microsoft Zune and KIN were flops. Maybe the 2011 devices will be better than Apple, but will they sell? They won't be competing with just iPads and iPhones, but with a wide variety of devices. Note too that some of NVIDIA's competitors have huge advantages in the mobile market derived by expertise in areas like Wi-Fi and cell phone modem chips.

For device makers there are a number of competing strategies to choose from. Do you want to start with the best graphics, or perhaps the best cell phone 4G technology? Or if devices become largely indistinguishable, maybe the low price supplier is the key to success.

I have owned NVIDIA stock in the past, but right now I think there are better bargains to be had. If Tegra ramps as rapidly as management would like, then sure, today's stock price looks nice. But I believe this is a wait and see situation. The new Tegra devices may come on the market as early as Q1 2011, but I want to see how they sell through. One popular device could make Tegra viable again, but only Apple seems to be able to guarantee the popularity of its own devices. Once you step outside of Apple, what I mostly see is ruinous competition.
See also:

My NVIDIA analyst call summary for Q3 2010
http://www.nvidia.com/

AMD versus Intel: Show Me the DX 11

New computers in 2011 based on competing processors from Intel and AMD will initiate a new era in personal computing: the combination of a CPU and GPU (graphics processing unit) on a single chip. Intel's offering is refered to as Sandy Bridge; AMD's is Brazos (part of the Fusion program). In both cases a line of processors will be introduced that will be distinguished by part numbers. Systems are being built now and should be first available to consumers in January, with many more systems and variations on the processors available as 2011 progresses.

Over the last decade in particular graphics processing has increased in importance for the vast majority of computer users. People watch and even edit video on a regular basis. 3D virtual realities, including games, are a common part the computer experience. Even business applications are increasingly visual and three dimensiona. Starting with Windows Vista, computers needed improved graphic computation just to allow the operating system to present all of its graphic features.

If there is one thing consumers (including business buyers) need to know about computer graphics, it is that DX 11, introduced with Windows 7 in 2009, is the graphics standard for today's applications. DX 11 is short for DirectX 11, which Microsoft designed to handle multimedia, including video and 3D graphics. The prior generation, DirectX 10, was introduced in 2006. While a powerful advance in that era, it is now seriously out of date. While many applications still use DX 10, most new software introduced in 2011 will run best with DX 11. When DX 11 is not available, they will default to the lower graphics standards of DX 10 or earlier.
When a computer runs a DX 11 game or application by substituting DX 10, it loses graphic details that enhance the visual experience.

Sandy Bridge cannot run DX 11. If you buy a computer with an Intel Sandy Bridge processor, you will have two choices. Sub-optimal graphics, or buying an add-in graphics card from AMD or NVIDIA. AMD's Brazos chips (and all their Fusion chips to be introduced in 2011), on the other hand, do run DX 11. If you think about the computer replacement cycle, this is a remarkable difference. Intel computers bought in 2011 can be expected to remain in use for 2 to 4 years. By the end of that cycle they will be running a decade-old graphics standard.

Intel is going to spend a hefty amount of money trying to convince people that Sandy Bridge based computers have graphics on par with AMD Brazos based computers. I've seen some of how that will work already. One online technology reviewer in England used a factually correct article that first appeared at Anandtech to argue that Sandy Bridge is about as good, possibly better, than AMD offerings. He used the following graphic to support his argument [see also the full Sandy Bridge Preview at Anandtech]:



What you see hear is that a desktop Sandy Bridge Core i5 (the top bar) shows a good improvement on Intel's earlier integrated graphics attempts. It is also significantly better, for this particular game, than a AMD Radeon 5450 card. That is great, but it is not a valid comparison for people buying new computers in 2010. The Radeon 5450 card, as you can see from the chart, could do remarkable things for an older Intel CPU with integrated graphics. But it is a card you can now get, retail, for $33.99 [See HD 5450 at TigerDirect]. It supports DX 11. It is the very bottom of the AMD discrete graphics line.

Fair enough, though, Sandy Bridge has the graphics equivalance of the cheapest, slowest discrete video cards made with AMD graphics chips, except it can't do DX 11. Brazos has, as its graphics engine, the equivalent of either a Radeon HD 6250 or 6310, depending on the exact chip used, but those don't correspond to any discrete chips released by AMD. However, graphics capabilities of the Brazos APU are only slightly less than for the 5450.

The Brazos chips, draws just 9 watts of power in the C versions with 6250s, or 18 watts in the E versions with 6310s. The Intel Core i5 2400 draws 95 watts (according to Anandtech). In fact [See Sandy Bridge Preview] they list no Sandy Bridge CPU that draws less than 65 watts.

In other words, in order to make Intel integrated graphics look better than AMD APU graphics, you need to take an expensive, power-sucking chip designed for desktop computers and compare it to a relatively inexpensive, power-sipping chip designed for netbooks and notebooks. And the Intel based desktop computer won't do DX 11.

The way it works out, AMD is releasing its Fusion (combined CPU/GPU) notebook & netbook chips in January. Intel is releasing its desktop chips with integrated graphics January. So direct mobile to mobile and desktop to desktop comparisons are not available yet.

AMD will be releasing more powerful Bulldozer chips designed for desktops later in 2011. They will support DX 11. Intel uses Atom for its netbooks; expect no DX 11 support there.

DX 11 adoption is well underway. Games tend to adopt a new graphics standard most quickly, but only when new games are introduced. See a list of games with DX 11 support. Note that Windows 7 itself supports DX 11. Expect new versions of most major application programs coming out in 2011 and 2012 to support DX 11.

What consumers need to know is that Intel based computers are essentially defective as they come off the assembly lines unless (1) you just do simple tasks like e-mail that are not graphics intensive or (2) they include a discrete graphics card from AMD or NVIDIA.

Who will tell them that? Not Intel. And Intel's advertising budget is such that you can expect a lot of obfuscation in media outlets, including technology magazines and web sites dependent on Intel for much of their advertising revenue. Intel also pays many large retail chains to promote its products over AMD products (by paying for ads).

If anyone is going to get the truth out, it is the millions of ordinary tech people who help everyone else with their buying decisions year-in and year-out. If they can get the typical buyer to ask the typical seller, "Does it do DX 11?" then AMD is going to pick up a lot of market share in 2011. On the other hand, if they stick to the "Intel is the premium brand," line, AMD will continue to have a hard time getting its message heard.

AMD Ships First APUs on Analyst Day

Computer chip maker AMD announced it began shipping its new APUs (Advanced Processing Units) to computer makers this morning from its Singapore plant. At its analyst day presentation AMD executives described and demonstrated the APUs, which combine an advanced CPU (general purpose computer processing unit) with an advanced GPU (graphics processing unit) on a single chip.

Emphasizing what a groundbreaking point has been reached, CEO Dirk Meyer held up a typical sized CPU and then a mid-range GPU card, which was about the size of a small paperback book. Then he held up the APU that will have the equivalent CPU and GPU computing power. It was smaller than the CPU chip, about the size of a postage stamp. I know that the GPU card contains not only a GPU chip but memory, a fan, and connections for video output, so the comparison was a bit of an exageration. But it is a sort of computing grail achievement that goes beyond mere size comparisons.

Of course, knowing AMD has been working for years to achieve this feat, much larger rival Intel has announced that it will also have an integrated cpu/gpu product release for 2011. AMD executives mocked it, as well they might. We know it is an inferior product. It supports a graphics standard called DX10, which is now four years old. AMD supports DX11. It is true that most older software and games can't take advantage of DX11.

But many games already can, and most graphics software updates are moving to DX11. So Intel will be making an offering that can't cope with new games or software. When you buy a new computer, it is often because you want to take advantage of new software. Intel will be leaving consumers in the lurch.

Nevertheless, Intel is the Goliath, and AMD's previous attempts to take on the giant have had mixed results. Intel's profits are usually higher than AMD's revenues, and Intel spends way more on R&D than AMD. A few years ago Intel was so far behind in graphics, it is remarkable that they are maybe only 2 years behind now.

Intel will heavily outspend AMD in marketing, and will omit to tell consumers that its chips can't run DX11. So for AMD to take a lot of market share in 2011, it has to get its story out. In my experience retailers are more interested in Intel advertising subsidies than in making sure consumers make an informed choice between computers based on AMD and Intel. I would hope that tech "geniuses" would tell show off their stuff by telling the public to choose AMD if they want good graphics and video capabilities. But it seems that a lot of technology mavens are employed by Intel and Apple.

If the word does get out that Intel cpu/gpu combination chips are not good enough, AMD's ability to take market share could become capacity constrained. Intel has a huge production capability to match its market share; AMD's capacity can only be expanded so much in the short run.

Still, even a 10% increase in revenues for AMD in 2011, with maybe a 1% increase in market share, would be a boon for AMD.

Watch this space closely. The actual computers will start being available to the public in January, traditionally a slow period for computer sales. Public acceptance of the new AMD products, or resistance to Intel advertising, should be knowable by March or so, and act as a predictor for the remainder of the year.

For investors a key element will be margins. AMD believes that with the new processors (and server chips introduced in 2010) it can improve its non-GAAP gross margin from about 40% for 2010 to about 44% to 48% in 2011. If that turns out to be true (if Intel does not start a price war), then earnings will rise nicely and AMD will be in an even better position to compete with Intel in 2012.

Dirk Meyer showed an HP thin light notebook running gaming level graphics using the new APU chips. He claimed it could run 8 to 10 hours on one battery charge, and would cost less than $600. I want one, and it would work a lot better for me than a smaller form factor tablet computer. This ability to reduce power consumption is being introduced across the range of new AMD products in 2011: for netbooks, notebooks, desktops, and servers. That is good news.

NVIDIA (NVDA) Risks, Opportunities

Last Thursday, August 12, 2010, NVIDIA (NVDA) reported its results for the quarter ending August 1, 2010, its second fiscal quarter of 2011. The results could be taken as a glimpse into July for the semiconductor industry since most other companies in the space reported second quarters ending June 30th. There is a concern that the recent boom in chip sales could slow appreciably because of slowing (but not really reduced) end demand by businesses and consumers.

Here I want to discuss the big trends engulfing Nvidia and the future of computer processing, including graphics processing. If you are interested in details of Nvidia's 2nd quarter, see my NVIDIA Q2 Analyst Conference Call summary.

Since graphics on personal computers began to be important in the late 1980s there has been a divergence between the central processing unit (CPU) and the graphics processing unit (GPU). Fans of history may recall that before that there was a similar divergence between the CPU and the floating point processor (FPU); you could pay extra for a computer with an Intel 8087, if you wanted to do math, or use math to accelerate graphics calculations.

This divergence between general computing needs and floating point math and graphics needs is fundamental. In an era that may be coming to a close Intel and AMD (and once upon a time, Motorola) produced most of the CPUs for the industry while ATI (now part of AMD) and Nvidia made the graphics processing unit. Many computers had no GPU, allowing the graphics work to be done by the CPU, which is fine for slow work like word processing. Intel, Nvidia and AMD also all made motherboard chipsets that included some GPU capabilities. As usual, those became more capable with time. Your $1000 GPU card of 2000 is not as powerful as the lowest end Intel graphics of today. On the other hand graphics demand has gone up with both the introduction of high definition video, fast rendering needs of games, and graphic content production systems. You don't want Intel graphics for any of those.

Conversely, the powerful GPUs from Nvidia and AMD can now be used to accelerate many processes that CPUs do more slowly. Mostly scientists, engineers, and graphics designers are taking advantage of these capabilities, but they will trickle down towards the mainstream. Moms may use them to convert an HD video to lower definition, or vice-versa.

For the most part FPUs were killed when floating point processing started being integrated into CPUs like Intel's Pentium. As transistor sizes shrank, and Windows became the common operating system, that became a natural way to give users more for their dollar.

2010 is the year that GPUs and CPUs are being merged onto a single chip. Unless you are an industry insider, you won't see the products until 2011.

Because high end GPUs actually use more silicon space than the current CPUs, there will continue to be life for independent GPUs, typically on video cards, for at least a few years. Video and graphics content makers, scientists and engineers will want workstations that include both a high end GPU and a high end CPU card. But the rest of us will probably migrate to the new combined cards (AMD calls them APUs, advanced processing units) between 2011 and 2015.

While Intel might be behind AMD in graphics, I would not want to bet on its losing a lot of market share in the new combined CPU/GPU arena. Intel's profits dwarf AMD's revenues, and their R&D budget is correspondingly higher. They will find ways of keeping their market share, most likely.

For Nvidia there is a bigger problem: they don't make GPUs at all. But before addressing that, let's look at the new tide coming in: mobile processing, mostly based on ARM-based processors.

Mobile devices are rapidly becoming more powerful, as you can see from the iPad, iPhone, Droid, etc. The key is low voltages with corresponding low power consumption, plus the usual shrinking of transistor sizes making it possible to integrate more processing capabilities into the processing units.

It is conceivable that just as PCs wiped out minicomputers and even, for the most part, mainframes, the new mobile CPUs could start invading the notebook, desktop, and even server computer space. It is already being tried, the reasoning being that a swarm of small, low-energy units can make ideal virtual machines for serving web pages.

Nvidia has the Tegra platform, which you'll be seeing a lot of later this year. It integrates an ARM CPU, an Nvidia GPU, and motherboard chipset functions on a single chip. It is a beautiful piece of work.

It also appears to be winning out over AMD in the high end GPU for non-graphic computations market.

I am not ready to declare the long line of evolution from 8086 chips to be over. They may just absorb what is best, and morph to meet changing user needs. But the competitive dynamics are certainly going to change in the next few years. If Nvidia's Tegra is a competitor, so are chips from Apple, Marvell, Broadcom, Qualcomm, TI, and many others. In fact too many companies have pinned their hopes on ARM processors. Even if ARM becomes the architecture of the future, it is likely that a couple of these companies will gain a marketing or manufacturing advantage, and the field will narrow over time.

Nvidia's best hopes for future profits are probably in specialized graphics processors for high-intensity computing. These may no longer be necessary for the average person's desktop or notebook, but high-end GPUs are subject to much less competition and therefore have higher profit margins on each unit sold.

AMD Earnings At Bat

Another earnings season got underway today when Alcoa reported. Of the companies I follow closely the first to report on its second quarter (Q2) will be the computer chip maker AMD.

Q1 2010 was the best first quarter in AMD's history, with revenue of $1.57 billion and GAAP net income fo $257 million. Guidance for Q2 (given at the AMD Q1 2010 analyst conference) was that it would be down seasonally from Q1. That is a typical seasonal pattern. In Q3 and Q4 each year the computer makers who are AMD's customers start ramping up for back-to-school and Christmas holiday sales.

I suspect that, within that context, Q2 has been a pretty good quarter, largely based on watching advertisements for consumer and enterprise computer equipment. In graphics chips NVIDIA did not come out with its new, competitive chips until well into the quarter, and even then supplies seem minimal, with AMD dominating key price points. I saw a number of adds for AMD graphics in machines that use Intel CPUs, which is a positive surprise.

In the CPU department I think the results may be mixed. Intel seems to continue to dominate in notebooks, while AMD has made some gains in the desktop department. Desktops are going a bit out of style as consumer items, and margins are low, but they are still critical for people who are doing non-mobile work where having computational power and large screens is still a big plus.

Finally, there is the server CPU market. I am seeing mixed signals here. AMD and Intel are more differentiated in servers than they were a few years ago. AMD has gone for more cores per CPU and more processing power per dollar, but Intel leads in pure performance. I doubt AMD made much headway against Intel in server market share this quarter, but it should still be a rising tide situation.

We'll get AMD's numbers on Thursday, July 15th. I'll be taking notes on the analyst conference and post them at AMD Q2 2010 analyst conference. You can also listen to the conference yourself live on the Internet, starting at AMD Q2 live conference. That will be at 2:00 PM Pacific Time (5:00 PM Eastern Time). If you miss it you can also listen to an archived recording later.

Can AMD Ignite Fusion?

AMD is the perpetual also-ran in the CPU and GPU (graphics processing unit) sweepstakes. According to AMD, their next big thing is Fusion, in which a CPU and a GPU are combined on a single chip, or APU (Accelerated Processing Unit). Is it possible that in 2011 AMD will make serious profits from this innovation? Or is this more like the power from nuclear fusion program, which after decades of work and billions of dollars of investment has yet to make commercially available electricity?

In my April 17, 2010 story, Paint AMD Black?, I discussed AMD's latest financial results. The second quarter ends with June, and that readout will tell us how well demand is holding up. But the real value in AMD stock, or lack thereof, is in whether it has made the right choices in its roadmap, given its difficult position versus rivals Intel and NVIDIA. While there are other aspects to that roadmap, Fusion is probably the most important in the 2011 time frame.

There are two major forks in the computer processor road in the 2009-2011 time frame. One is about graphics processing, including using processors designed for graphics to compute other types of problems. The other fork is between traditional PC processors and mobile processors, where low power requirements are as important as computing and graphics capabilities.

Intel is the largest of the three rivals, so I will describe its known roadmap first, then contrast that with the roadmaps of AMD and NVIDIA. Intel dominated the market for PC computer chips almost from its inception, but it has traditionally supplied only minimal graphics capabilities. Keep in mind that a regular CPU can do graphics calculations. A GPU speeds up graphics calculations. GPUs can do this because graphics calculations use a relatively small number of well-known algorithms to accomplish their tasks, and these algorithms can run in parallel, all at the same time. In addition to letting the CPU do graphics, as that became unsatisfactory even on low-end computers, graphics were added to motherboard chip sets. This is called integrated graphics, and Intel (as well as NVIDIA and AMD) developed low end-graphics for motherboard chip sets.

Intel also tried to develop what are called discrete graphics chips. These are not on a motherboard, but on a separate card, called a video or graphics card. Intel recently abandoned its Larabee (see Anand's Thoughts on Intel Canceling Larrabee) project for a high-end discrete graphic chip, but will be putting its expertise to work by adding the failed Larrabee circuitry to its own CPUs, some time in the future. So right now Intel has no clear strategy for dealing with GPU computing on desktop or server computers. If you want that, you need to add an NVIDIA or AMD ATI GPU card to an Intel-based machine, and we are already seeing a lot of that.

In the mobile space Intel sold its floundering mobile application processor unit to Marvell, which in retrospect was a mistake, as Marvell is quickly becoming a big player in mobile devices. Instead Intel concentrated on Atom, which was originally used in the netbook space. Intel's strategy is to make Atom more powerful, yet less energy needy, to move more into the smartphone and tablet computer space.

AMD has decided to focus on the PC (desktop and notebook) and server market with graphics capabilities being a major weapon to win market share from Intel. The last time AMD won a lot of market share from Intel was also due to a roadmap decision, when Intel tried to bifurcate customers into 32 bit and 64 bit users. AMD instead integrated 32 bit and 64 bit capabilities on its Athlon and Opteron chips, which was so popular that Intel had to change it roadmap.

Will it work this time? One problem is that AMD is typically a half-step behind AMD in the process technology it uses to manufacture its chips, so it either needs a larger chip to accommodate the same number of logic gates, or must make due with fewer gates on a similar sized chip. GPU chips for high-end discrete graphics are big, and so are the CPUs for PCs. Combining both on a single chip could lead to a lot of defects, low productivity, and negative margins. The first Fusion chips will probably have the equivalent of mid-range GPUs mixed with mid-range CPUs. Yet if defect rates are low enough, these could be formidable chips. There would be no need to run to an external motherboard graphics chip or discrete graphics chip in order to do the processing. For many applications that lack of communications lag will make up for the lower processing power of each unit.

AMD has simply decided not to compete in the mobile space except to the extent that its notebook processors are able to pick up some of the netbook or tablet market. Given the importance of the mobile market, that may not seem smart. But the mobile market is fiercely competitive; even Intel has floundered there, against the likes of Qualcomm, Broadcom, Marvell, and many others. AMD has limited resources compared to Intel. It can't afford to enter markets where it has no competitive edge.

NVIDIA has an interesting strategy too, and it involves a different take on each of the forks. It is going heavily into mobile with its Tegra processor. It is ahead of AMD in high-end graphics, and graphical computing, with its Fermi and CUDA technologies. Clearly NVIDIA would like to allow their GPUs to also do the work of CPUs and thereby crash the PC motherboard party. They could do that by adding ARM based processors to Fermi. But it would be a whole new marketing world for them. They might simply wait and watch the ARM/mobile space eat up the old PC space the way PCs once cut into the minicomputer space.

I own AMD stock and think it is currently undervalued, but there is no denying that Intel is much larger that AMD and has lots of profitability, while AMD has just squeaked by for years. Until AMD shows several successive quarters of profits and market share gains, and it is clear that it chose the best roadmap, I don't expect much short-term upward movement of the stock. On the other hand, waiting until all the good news is in means missing out on almost all of the stock appreciation, if it happens.

See also:
AMD Q1 2010 Analyst Conference Summary
AMD 4/15/2010 earnings release
AMD at Wikipedia
my main AMD page

Paint AMD Black: Profits at Last!

AMD, contrary to expectations, posted a profit in the first quarter of 2010. This is remarkable. The maker of computer processing unit (CPU) and graphics processing unit (GPU) chips was not expected to show a profit until well into 2010, if ever.

In Q4 2009 AMD's rival Intel paid it $1.25 billion to settle (probably true) allegations that Intel had engaged in illegal, monopolistic practices that lowered AMD's sales earlier in the decade. Whether or not AMD was profitable in Q4 is problematic. GAAP (Generally Accepted Accounting Principle) net income (profit) was $1.18 billion. But take away the $1.25 million from Intel, and you have a $70 million loss. On the other hand AMD claimed in had non-GAAP net income of $80 million. But to get that number they had to exclude GlobalFoundries, which is the new company holding the fabrication facilities AMD had formerly owned 100% of. Still, Q4 revenue of $1.65 billion was up an astonishing 42% from Q4 2008.

Q1 2010 had positive net income any way you look at it. Revenue was up 33% from $1.18 billion in the year-earlier quarter. But there were a number of one-time charges and benefits related to spinning off GlobalFoundries that need to be examined by investors. GAAP net income was $257 million. Most of the time companies give non-GAAP numbers because they are higher that GAAP numbers, but in this case AMD owned up that on a comparable non-GAAP basis net income was only $63 million (then again, prior guidance was for negative that figure). See my AMD Q1 2010 Analyst Conference Summary for a more detailed explanation, or the AMD 4/15/2010 earnings release for a number of reconciliation tables between GAAP and non-GAAP numbers.

Because of the cyclical buying patterns of both consumers and enterprise buyers, the first half of the year is typically slower for AMD and Intel than the second half. Even with the global semiconductor industry climbing rapidly out of the recession (or, as I prefer, the Panic of 2008), AMD was expecting to show a loss this quarter. What made the difference?

AMD has introduced some very competitive products lately. In servers, it offers chips with more cores than Intel (although Intel cores allow more threads per core, which helps in some multi-threaded software applications). In notebook computer chips it has been coming from behind, which makes it relatively easy to gain market share. In graphics chips AMD picked up a lot of market share from competitor NVIDIA these last two quarters, and with very good profit margins (there's a manufacturing capacity shortage, so no price war at present).

As we move into 2010 the most probable trajectory is fairly sweet for AMD, but of course competition with NVIDIA and Intel will remain fierce. One way to categorize servers is by number of processors; most low end servers are 1P, while 2P servers are very popular. There has been a historic price premium for 4P and up servers because of the difficulty making that many chips communicate effectively with each other. Now, at least for 4P, AMD has solved that problem. The new 6000 series 2P chips are ready for 4P. At first they will be used mostly for 2P boxes, but the incentive to move to 4P boxes, especially in virtualized server farms, will be significant. A 4P box would have 4 AMD processors with as many as 12 cores per processor, or 48 cores. Each core can easily run as a virtual machine, so for instance for Web server farms, each box would be the equivalent of 48 Web servers. The economies of scale will be quite attractive.

On the other hand, a lot of tech guys just wait for Intel, no matter how slow Intel is. So how much actual market share AMD will pick up while those guys are waiting for Intel is not predictable.

There are other new products being introduced this year, but the next big thing is Fusion, in which chips can act as both CPUs and GPUs. Look to hear more about that this coming fall.

AMD Q4 2009

AMD reported on its fourth quarter of 2009 this Thursday. It was a quarter of impressive improvements, with sales up 18% sequentially and 42% from the year-earlier quarter. That is excluding the almost $1.25 billion payment from Intel, which was entered in the operating expense accounts, and so did not appear in revenues. AMD also excluded the payment from non-GAAP numbers.

The hot money bailed out Friday, sending the stock price down over 12% to below $8. That is the way of hot money. Looking at the charts, my guess is a lot of hot money bought in above $8 during December and January, which makes this another example of hot money being stupid money.

Investors, the people who actually reap profits from business activity, need to look at the moving parts within AMD to see whether staying long is justified. The main reason to suspect it is not is that AMD is competing against Intel and NVIDIA, two very tough competitors. In addition, ARM-based processors are coming on strong against the x86 brethren, so in the next few years companies like Marvell (MRVL) may be nibbling even at mighty Intel's heels.

First, look at measures of profit. I'll use three: GAAP net income, non-GAAP net income, and EBITDA. Again, I exclude the $1.242 billion Intel settlement. GAAP net income was negative $64 million. Non-GAAP net income was positive $80 million. EBITDA ((Earnings Before Interest, Taxes, Depreciation and Amortization) was $282 million, impressive except that AMD has a lot of debt and so interest, and has invested a lot of capital and so depreciation does represent cash spending in the past. Still, I think it is fair to say that by some investor standards the quarter was profitable. The non-GAAP and EBITDA numbers exclude the former fabrication business. This makes sense to use since that business has been spun off (to GlobalFoundries). Starting in Q1 2010, AMD will be fabless.

As to the future, we want to know whether revenues will continue to ramp, and whether profit margins on revenues will continue to improve. Which means we have to look more closely at product trends.

AMD is now in the graphics processor business, and after years of struggle is showing some success. NVIDIA is the company to beat, and while that is still a distant goal as far as chip revenues, AMD almost certainly gained market share in Q4. AMD has the only GPUs (graphics processing units) that work with DirectX 11, the newest graphics standard. This attracts forward looking purchasers, and resulted in graphics revenues of $427 million, up sequentially 39.5% from $306 million, and up 58% from $270 million year-earlier. The worry would be that NVIDIA will have a better DirectX 11 capable processor some time in 2010. It is not a for sure thing, given how NVIDIA has been stumbling the last couple of years. While I don't expect AMD to have another year of 58% GPU revenue growth, I suspect whatever NVIDIA does, AMD will see healthy growth this year. We are going into a computer refresh cycle, and NVIDIA's slowness to market means it lost opportunities to be built in to systems.

In the CPU business revenue growth was robust, if not as eye-popping as with GPUs. Revenue was $1.214 billion, up 13.5% sequentially from $1.069 billion, and up 39% from $873 million year-earlier. Growth was described as broadly based, but Opteron 6-core server processors were described as having been a source of strength.

If profit margins expand along with revenues in 2010, today's AMD market capitalization of $5.3 billion is going to seem ridiculously pessimistic. On the other hand, since AMD still has almost $5 billion in long term debt and liabilities, any further stumbles of the kind we saw this last decade could have very serious consequences.

Q1 is typically seasonally down for CPU and GPU manufacturers. Given that AMD was not able to ship enough of its new GPUs to meet demand, there is at least a possibility that Q1 revenues could be closer to flat for AMD. That would be a very good sign for the full year 2010.

On the whole I would say I think AMD stock looks undervalued at $8 per share, but then I am a long-term AMD investor, so maybe I am being overly optimistic.

For more details, see my AMD Q4 2009 analyst conference summary.

See also amd.com

NVIDIA Q4 Glows; Disses ATI & AMD

NVIDIA (NVDA) posted soaring revenue and earnings gains for its 4th quarter of fiscal 2008 ending January 27, 2008. At the analyst conference on February 13, 2008, management told how this was done and predicted a bright future. For details see my NVIDIA analyst conference summary.

Revenues of $1.20 billion were up 7% sequentially from $1.12 billion and up 37% from $879 million year-earlier. Net income of $257.0 milion was up 9% sequentially from $235.7 million and up 57% from $163.5 million year-earlier. Q1 is expected to be down only slightly, instead of the usual seasonal dip.

Management declared this "the era of visual computing." I agree. NVIDIA has led the charge on ever-better graphics processing and presentation capabilities for personal computers. The chips (GPUs, graphics processing units) they sell today are marvellous, and the chips they promise for the future will be even more capable.

Even though everything reduces to 2 dimensions on your LCD screen, for games and other 3-D applications the math processing power needed to make realistic (detailed and fast-changing) scenes is only now beginning to come online. NVIDIAs GeForce 8800 GT GPU has a larger die size and more transistors than most CPU chips. Because of that there was some extra expense in Q4 trying to get good yields of this new top-of-the-line chip out of the fab.

NVIDIA management believes that for most people it now makes more sense to buy a medium or even low-end CPU and combine it with a high-end GPU in order to get maximum use from their computers. This would be especially true if you mainly use your computer for gaming or viewing videos. On the other hand the ability of programmers to find new uses for any extra CPU processing power is a well known trend that is not likely to break up. Right now programmers are behind the curve created by multiple core processors, but I expect new products in 2009 will bring that more into balance. Of course if you have a CPU that mainly just idles while waiting for you to type, even the CPU's of ten years ago are powerful enough. Still, the trend towards doing more computing on the GPU is a good one that will provide programmers with increased flexibility and power.

As to the competition, which is the ATI devision of AMD, NVIDIA management was not worried about them. On the one hand, to illustrate the new trend towards mid-range PCs with high-quality graphics, management picked the Gateway Effects 7020 desktop PC built with an AMD Phenom quad core CPU and an NVIDIA GeForce 8800 GPU.

On the other hand they thought ATI graphics chips are really competitive only in the low-end market, particularly in the low-end notebook computer market. Management went out of its way to dismiss the new X2 cards from ATI as not fast enough and occupying too much space in a computer. These cards have two HD 3870 graphics CPUs on a single card. NVIDIA favors the one-chip per card design, but hedged a bit by saying that they could produce a two-GPU per card competitor if the X2 design catches on. One analyst was surprised enough by this assertion to ask a follow up question and get the same dismissive answer. CNET reviewers, who I respect but sometimes disagree with, said the X2 outscore NVIDIA's GPUs on many games.

I am not a gamer, so I'll leave it to the pros to decide.

NVIDIA is a great technology company and its stock is cheap right now due solely to the liquidity squeeze and global economic uncertainty. I don't own the stock but have been very tempted to buy it lately.

Keep diversified.

See also my NVIDIA (NVDA) main page