Showing posts with label microprocessors. Show all posts
Showing posts with label microprocessors. Show all posts

Tuesday, May 19, 2009

AMD: The Dream Renewed

Lately there has been an uptick in the price of AMD stock following news that the European Union will fine rival Intel some $1.44 billion for illegal anti-competitive practices earlier this decade. This lends credence to the AMD v. Intel lawsuit in the U.S., which could result in billions of dollars of damages being awarded to AMD. AMD's market capitalization even after the recent run up stands at less than $3 billion, largely because it has been years now since it has shown a profit.

We probably are not at a the start of a new era, for reasons I will give shortly, but consider the possibility that AMD is going to emerge as a financially strong competitor to Intel in 2010. Intel would have to be prevented from using its substantial cash reserves to force AMD into another price war that would keep AMD from making significant profits. AMD would then be free to continue to introduce superior technology as it did earlier in this decade, only going forward it would be competing in a free market where computer manufacturers could incorporate AMD products into their designs without being threatened by Intel.

What happened earlier in the decade with AMD's technological innovations is instructive. AMD, which had mainly made Intel CPU clones until 2000, made three major design changes that rocked the computer world.

First, Intel had decided that the switch from 32 bit to 64 bit computing would be made by segmenting those two markets. If you wanted to do 64 bit computing the Intel way, you would have to switch to Itanium processors, and your old software would not work on those new processors. At the low end you would stay on the 32 bit road map. Intel ignored feedback from its customers that they wanted chips to run 64 bits on the old 32 bit instruction set, making them capable of easily running older software, and allowing for software makers to easily design 64 bit software. AMD listened to customers and introduced chips that could run both 64 bit and 32 bit software. A lot of people liked that.

Second, AMD introduced HyperTransport, which fixed a major bottleneck in Intel chips that prevented data from being moved from the CPU to memory, other CPUs, or other devices. Only this year, in 2009, did Intel introduce a similar feature in its processors.

Third, AMD's design took into account the problem of energy consumption and overheating of chips. Intel's plan was to make chips run ever faster, with no major changes in architecture, but fast meant hot. AMD changed the architecture to get just as much performance from its chips, but using less energy and running cooler.

The Opteron chip for servers incorporated all of these features, and for a couple of years steadily gained market share. Of course, AMD had almost no server market share before Opeteron. The same features were introduced in AMD Athlon chips for personal computers.

In 2004 I thought AMD was going to blow Intel out of the water. I thought the natural conservatism of the IT community would be overcome by the clear demonstrations of AMD's better technology. But in addition to AMD's need to build credibility and overcome the close relationships Intel had with manufacturers, it ran up against Intel's war strategy.

Internally, we now know, Intel admitted the AMD processors were better designed. Externally, Intel used three classic tactics: it used its massive ad budget to create the perception that its products were still superior; it lowered its prices so that AMD could not generate the profits necessary to maintain its process development efforts; and it used well-documented illegal and unethical tactics to keep manufacturing partners from using AMD chips in their designs.

Then, gradually, Intel followed the AMD roadmap. They started making greener chips that had the 32/64 architecture. Finally, they licensed AMD's hypertransport technology and built their own take on that into their chips. Given all that, with Intel's admittedly superior manufacturing capabilities (they can make smaller transistors than AMD, and so get more into similarly-sized chips), today it cannot be said that AMD has any kind of clear technology lead.

In retrospect, Intel used its market dominance to punish AMD for innovating. Prior to the described era of competition, Intel was quite happy to set extraordinarilly high prices for processors. AMD's clones were cheaper (and a year or two behind in design and process technology), but still profitable. If AMD had been content with 10% of the market, Intel would have kept it prices high, in turn guaranteeing AMD profits. It is almost universally acknowledged that Intel wanted to keep AMD alive and profitable, if not really competitive, simply to have plausable deniability to charges of running a monopoly.

Consumers were helped by AMD's technology innovations, but AMD stockholders lost out.

Hopefully the AMD v. Intel lawsuit will sort all that out. Billions of ill-got Intel profits should be rewarded to AMD.

Given the two corporate cultures, I suspect AMD has a roadmap that is in many ways superior to Intel's. To execute on the roadmap, however, they need money, and they need to know that Intel is not going to engage in ongoing anti-competitive practices.

AMD has at least one trick up its sleeve that should be played soon. It has a graphics division that is vastly superior to Intel's. Intel is pouring money into its graphics design division, hoping to catch up before it loses the lawsuit and AMD has the money it needs to pursue its vision of CPUs that do what clients want, rather than what Intel dishes out.

Disclaimer: I own some AMD stock.

See also my main AMD investor page at OpenIcon.com

Friday, January 23, 2009

AMD on the Ropes

When you go up against Goliath, you had better not make mistakes. The biggest David and Goliath story in the semiconductor chip industry in this decade has been AMD and Intel. Right now AMD is badly battered by the giant Intel, partly because AMD stumbled in introducing quad core microprocessors.

AMD reported fourth quarter 2008 results at its analyst conference yesterday, and they were a disaster. Even when you strip out special or non-cash items, AMD lost a lot of money on its basic operations producing microprocessors and graphics processors.

For details on what happened in the quarter and what management thinks of the first quarter of 2009, see my AMD analyst conference summary for q4 2008.

AMD made its biggest mistake, or double mistake, when it and Intel were transitioning from dual core to quad core chips. Much of today's problems stem from that era. Intel, which had been rapidly losing market share in the high-margin server chip market, decided to kluge together two dual core chips, each a separate piece of silicon. AMD insisted on the more elegant solution of actually putting all four cores for the Opteron on the same silicon chip. Then it ran into design and production problems, delaying its quad core introduction. Intel, with its dominant market share and huge cash flow, had no competition in quad-core chips for about a year. That was enough to regain most of the market share it had lost and to using pricing to keep AMD at bay even after quad core Opterons were finally introduced.

AMD also acquired graphic chip maker ATI, which was a smart move in itself, but they paid way too high of a premium for the acquisition and were saddled with debt that was unsupportable given ATI's marginal profitability.

AMD has always been behind Intel in what is called process technology. AMD's superior chip designs are typically made using larger silicon features than Intel's, which means Intel can put more transistors on any given size of chip.

The fourth quarter of 2008 should have been a magic moment for AMD. The new version of its quad core Opteron processor is at the same process technology, 45 nm, as Intel's current competitors. It is a very competitive chip. AMD also has very competitive desktop Phenom processors and notebook Turion processors. Some people even admit that it is ahead of NVIDIA in graphic chip technology for the first time since ATI was acquired.

But demand for computers fell across the board in the last quarter of 2008 due to the banking crisis. Instead of making a profit, gaining market share, and preparing for the next round of competition, AMD was badly hurt. Intel has a huge inventory of unsold processors, so they will probably be dumping them on the market. And Intel has huge cash reserves, so they can better develop new technology while making little profit during the downturn.

So is AMD finished? By no means. The spin off of their manufacturing operations to The Foundry Company has been well publicized and should be completed in February. At that point the core of AMD will be a design and sales company. AMD's much smaller team of engineers has consistently out-designed Intel for a decade now; maybe they can keep up the heat.

AMD will emerge from its spin off with a substantial amount of cash and somewhat decreased debt. That leaves it vulnerable, of course, to Intel's "aggressive" sales tactics (illegal, according to some nations) and ability to flood the market with cheap processors.

So visibility is practically non-existent here. If the economy turns around before Intel is able to introduce 32 nm based chips, AMD might be able to gain crucial high-margin market share in servers. But once Intel starts selling 32 nm chips, AMD will be at a disadvantage again, for a while, no matter how good their chip designs are.

I still own some AMD stock. It has been the worst performing stock in my portfolio. It is a reminder that even geniuses can lose money for you.

So keep diversified.

More data:

www.amd.com

Sunday, October 19, 2008

AMD Breaks New Microprocessor Ground

AMD is feisty. Despite the brutal monopoly tactics used by rival Intel (illegal kickbacks to customers; using the cash accumulated by earlier monopoly pricing to try to drive AMD out of the market; etc.), AMD is not only still in the game, it presents a growing strategic threat to Intel and a growing opportunity for eveyone else in the computer industry.

In the third quarter of 2008 AMD made progress on four major fronts. It showed a degree of creativity and flexibility that we have not seen from Intel since the 1980s. While Intel, which "owns" about 80% of the global microprocessor market, continues to have enormous advantages, it is AMD that is leading. Even if you just bought a PC with an Intel processor, its main advances over processors from 5 years ago - combined 32 and 64 bit capability, multiple processors on a single piece of silicon, and lower energy use - were all invented at AMD and then reluctantly introduced by Intel.

You can read some of this history in articles listed at my main AMD page for investors. I'm going to focus this blog entry on recent stategic developments.

One place AMD has consistently lagged Intel is process technology, which is the technology of creating semiconductor chips with ever smaller surface features. Each new generation allows more gates and hence capacity to be fit on the same size of silicon (or you can put your old design on a smaller piece of silicon, which costs less). It is very, very costly to develop each new generation, both in terms of R&D expense and then in creating the equipment that the new fab (semiconductor factories are called fabs) needs to make the improved chips. Intel made so much money from its microprocessor technology that it was able to keep well ahead of everyone else in the industry, often having as much as a two year lead time. For makers of chips requiring less transistors, this was not a big problem, unless they competed directly with Intel for a particular product.

AMD met this challenge by doing more with less; their processor designs were significantly better than Intels from about 2000 until 2006. But if AMD could be at the same level of process technology at the same time as Intel, it would be in a better position to press its design advantage. During the last decade AMD, working with IBM, has created very sophisticated fabs in Germany, and it has narrowed Intel's lead time. But it is still 6 months to a year or more behind Intel, and Intel has the money to drive the research and investment to keep that lead or even extend it. AMD is now shipping chips using 45 nm technology, but Intel started shipping theirs at the beginning of 2008.

In the past decade many major semiconductor players have emerged that don't have their own fabs. These fabless companies, like NVIDIA and Marvell, have done very well. But the independent foundries that produce chips for them have not been able to keep up with AMD in process technology, much less Intel. You might think that independent equipment makers like Applied Materials and KLA Tencor, would focus on getting ahead of Intel, but most chip makers don't have to put but a fraction of number of gates on a chip as AMD or Intel do, so the economic incentives are not there.

AMD's new stategy is to start allowing other companies to use its foundries when they need the advantage of cutting-edge chip manufacturing. As this happens research and development costs can be spread out over a larger number chip makers. With IBM, AMD and others working together it is possible that the AMD coalition could catch up with the Intel axis in process technology over the next few years.

To accomplish this strategy AMD brought in new investors, basically the nation of Abu Dhabi. AMD is contributing its fabs and some of its intellectual property to the Foundry Company, but Abu Dhabi will be the majority owners, in return for a multi-billion dollar cash infusion. In addition to continuing to expand the AMD plants in Dresden, Germany, AMD will break ground on a new plant in New York State. At a time when most new foundries are being built in places like Malasia and China, you might wonder why New York State. Well, what some people believe is the most sophisticated chip factory in the world, one of IBM's fabs, is in New York. These factories are mainly run by robots. There is no need for a lot of low-paid laborers doing hand work.

Except for the cash infusion, we won't see the results of this strategic turn until 2010 and beyond. But there are the three other fronts I spoke of: graphics, servers, and profits.

AMD aquired NVIDIA rival ATI for about $5 billion back in 2006, just in time to see ATI lose much of its share of the graphics chip market (at about the same time re-designed Intel microprocessor chips caught up with AMD's and Intel lowered prices to make all microprocessor sales unprofitable). I want to repeat what I've said many times, I think the NVIDIA team may be the smartest guys in the tech world. Still, they are AMD rivals in graphics. Recently ATI (now an AMD division) stole a march on NVIDIA with a new line of chips and graphics cards that became very popular very quickly with game enthusiasts this summer. All NVIDIA could do was lower the prices on their own chips while they rushed back to the drawing board (okay, these guys never leave the drawing board, they sleep on the drawing board, but you know what I mean).

NVIDIA still owns the graphics card market, but in Q3 AMDs graphics division revenue jumped to $385 million from $285 million in Q2, a 55% increase in 3 months. More important, AMD/ATI has two advantages over NVIDIA that might come into play in the next 2 years and allow them to eat further into NVIDIA market dominance. AMD leads NVIDIA in process technology, precisely because NVIDIA does not have its own fabs. And AMD can integrate its processor (CPU) development with its graphics development, which NVIDIA can't. This is also a three way race because Intel has long made low-end graphics chips and is trying to upgrade that division in response to the AMD threat. All this will be very good for consumers, who are already seeing inexpensive video cards that are targetted at gamers but that can easily handle high definition video and that can be used as super computers those of us who like to crunch numbers.

In server chips, so important to the Internet world, AMD is shipping its first 45 nm Opterons. Depending on what they are used for, they can be seen as better or worse than Intel's current server chips. AMD chips use less power and are better at math and sending information to and from memory. Intel chips are better at heating your building (actually, they are quite good, having finally incorporate much of the Opteron circa 2002 design improvements). The same advantage you see in Opteron server chips will soon become available in Phenom chips for desktop computers. AMD already has great notebook computer chips, the Puma series.

And for the final strategic advance, we have: profits. While still working towards GAAP and even non-GAAP net income, in Q3 2008 AMD showed operating profits. Much of the industry was concerned that if AMD continued to lose money it would be forced into bankrupcy, and Intel would go back to its high-prices and low performance everyone came to expect in the 1999s. By lowering costs, including by making the Dresden fabs highly efficient, and raising sales, AMD would have been back in the game even without the Abu Dhabi deal.

I own AMD stock and two computers containing AMD processors. I also own stock in Marvell. In the past I owned NVIDIA stock, but do not at present.

See also www.amd.com

Tuesday, September 30, 2008

AMD Shanghai Opterons Could Restore Profitability

Microprocessor and graphics processor chip maker AMD let slip that its new Shanghai version of the Opteron processor is actually ahead of schedule for production. Samples are available now for testing and it may begin appearing on servers as early as Q4 of this year.

AMD was doing well in its efforts to become a true rival to Intel until the Barcelona version of Opteron ran into problems. It came out over a year late. It had some merits: the four core version had all four cores on a single chip. Intel's four-core Xeon server processor actually was two two-core chips packaged together, which resulted in a variety of performance issues.

The Shanghai Opterons will also have four cores, but they will be on a 45nm process enabling them to have larger internal memories (caches) and other acclerators that were not present in the Barcelona versions, which were done with a 65nm process.

In looking at Q3 2008 results, which will be released October 16, we will be seeing the results of revenues from the current generation of Opteron processors. This generation has its merits, and may have sold well into energy conscious server customers. But if there is a surpise bump it will be from the recently reinvigorated ATI graphics unit, which came out with highly acclaimed products that may have sold very well in Q3. Notebook CPU and graphics chips, which were once a weakness for AMD, also should be showing a strengthening trend.

So if Q3 was better than expected, and Shanghai starts shipping in Q4, and the economy does not melt down, AMD investors will have something to cheer about for the first time since 2005. Given the almost vanishingly low current market capitalization for AMD, that might generate quite a pop.

Then again, I am a long-term investor in AMD, so this analysis could just be wishful thinking.

For more data see my AMD page and www.amd.com

For the technical news article on Shanghai see AMD says new 'Shanghai' chip is ready to go by Brooke Crothers at CNET.

Its always best to ...

Keep diversified!

Sunday, January 20, 2008

AMD Beats Expectations

Wall Street's gleeful expectations of the demise of AMD, and return of monopoly profits at Intel, were shattered when AMD delivered its Q4 results on Thursday, January 17, 2008. While AMD is far from being out of the woods, its situation is much better than most analysts thought.

The headline numbers offered an excellent demonstration to paying attention to the difference between GAAP and non-GAAP numbers. Revenues under both systems were the same, $1,77 billion. That was up 8.5% from the previous quarter but flat from Q4 2006, so it mainly represents the seasonal nature of the personal computer business.

GAAP net income was negative $1.77 billion. That would normally mean a company spent twice as much money producing and selling its goods as it was able to sell the goods for.

But non-GAAP net income was negative $9 million. Losing $9 million in a quarter is not good, but it is a far cry from a $1.77 billion loss.

The difference has to do with AMD's acquisition of ATI in 2006 for over $5 billion. At the time that seemed more than ATI was really worth, but AMD thought it was worth it to obtain ATI's graphics chip and motherboard chipset (the chips that make the microprocessor of a computer work with other chips on the board that holds them all) capabilities. Unfortunately ATI's competitor NVIDIA came out with new product that kicked ATI's behind in 2007.

So for Q4 2007, under GAAP rules, AMD admitted they paid to much for ATI, writing off $1.6 for lost "goodwill." You can still look at it two ways. AMD admitted they lost $1.6 billion. Or they had already paid for ATI back in 2006, so this is what is called a non-cash charge.

It's not like the ATI acquisition was all folly. AMD actually has, at least for the present, a strategic advantage over rival Intel because of it. Intel makes processors and chipsets, but only low-end graphics chips. NVIDIA makes high-end (and low-end) graphics chips and chipsets. Only AMD can deliver all three to computer makers: advanced microprocessors, chipsets, and high-end graphics chips.

The other part of AMD's report that was a surprise to most of us was the "Barcelona" quad-core Opteron processor situation. Bacelona chips had been delayed in Q1, and an error was found in the chips, so AMD announced it would not be making general shipments until Q1 2008. Actually the error was easily fixed in the Linux operating system, so large data centers that had been waiting for the chips decided to buy them despite the bug. Nearly 400,000 were shipped. Of course the revenues from most of those won't show up until Q1 2008.

Because of the quad-core shipments ASPs (average sales prices) were up for processors, and number of units shipped were up as well. Graphics revenues were up, but not so much.

Intel, of course, has fairly decent chips out now, in many ways playing catch-up to AMD designs, but having an advantage of introducing 45nm (higher transistor density) technology before AMD. But AMD is sampling 45nm chips, so the race is hardly over. On the Intel side they are working to catch up with AMD and NVidia in graphics capabilities.

With demand for personal computers (mostly for business use) exploding in India, Russia, China, Saudi Arabia and other rapidly developing nations, there is room for both AMD and Intel to grow. Behind in technology, and having been forced to cease, or appear to cease, illegal marketing practices due to an AMD law suit, in 2006 Intel started a brutal pricing war to keep its market share. In 2007 Intel introduced better chips and relented on pricing issues, which allowed it to start raking in money again. AMD has benefitted from the easing of pricing pressures, but must regain a clear technology lead if it is to be free from that danger in the future.

AMD's stock price received a nice pop on Friday after the Q4 results announcement, but the stock price is way down from its 2005 price. AMD is a risky investment, but management believes AMD will turn profitable in Q3 2008. If that is true, the stock is significantly undervalued at today's price.

I own AMD stock.

See my notes on Thursday's AMD analyst conference
See my main AMD page, with links to more articles
http://www.amd.com/
My Intel main page

Wednesday, January 16, 2008

Intel Disappoints, But Stock is Undervalued

Sell-side analysts (those who work for Wall Street brokerage houses that make commissions by getting clients to buy and sell stocks) did not like Intel's Q4 results or forecast for 2008. Pricing by auction in the stock market made the price of Intel stock plunge today (by 12.25%) - no one wants to buy, if at the end of the day you can buy cheaper. In addition Intel stock was already cheap, like most stocks, because of the current liquidity squeeze.

In the Intel v. AMD tech war, I favor AMD, both because it is the traditional underdog and because it was innovating processor design in the opening years of this decade, which benefits everyone who uses personal computers. I think AMD stock is way cheap right now, but that does not blind me to the value of Intel stock.

If you want a lot of detail on Intel's (INTC) Q4 results and management's explanations, take a look at my analyst conference summary.

Here I'll focus on a couple of points. The first in Intel's guidance on Q1 2008 margins. Margins are the difference between revenues and expenses, so at a given level of revenues, higher margins are good. Intel expects margins to dip in Q1. The Wall Street geniuses acted like that was a big surprise and a reason to dump Intel stock.

But there are two main things accounting for the Q1 margin dip. First, demand for CPUs for microcomputers is seasonal, always peaking in Q4. Q1 revenues are expected to be lower; with many costs fixed, that in itself will lower margins.

Second, Intel sold its cell-phone processor division to Marvell (MRVL) last year. It made 2007 a trying year for Marvell because they paid cash for it. It was great for Intel because they got cash for a money-losing unit (Marvell expects it to become profitable in the very near future). Under the agreement Intel would continue manufacturing the processors until Marvell could arrange offshore manufacturing for them. That transition began in Q4 2007 and will be completed in Q1 2008 except for some legacy processors that Marvell did not think it worth tooling up to produce. Those legacy processors will still be sold by Intel to Marvell.

The margins were good on the Marvell processors, but now those revenues will go away. So the best guidance Intel can give is that Q1 revenues will be weaker than is the norm for the season.

At the same time, Intel said their inventories are thin. And with demand in India, China, and Russia charging ahead, I think they are being cautious in their predictions. Technology is still a growth area, and computers, even if they morph into cell phones, are the heart of modern technology.

According to the Intel page at NASDAQ, the company has a PE ratio of 16.86, which inverts to earnings on stock of 5.9%. With Intel expected to grow revenues 10% in 2008, and net income (profit) growing faster than revenues, that is a bargain price.

Along the same lines, I think that NVIDIA stock is a bargain today. I'll write about NVIDIA again after I hear their what they have to say at their analyst conference.

By way of disclaimer, I own AMD stock, but not NVIDIA or Intel stock.

Keep diverse!

See also www.intel.com

Wednesday, November 21, 2007

AMD Valuation

Last year AMD bought ATI for $5.4 billion. Today AMD is valued by the market at $5.9 billion. What gives? What sort of valuation should be put on AMD stock?

AMD, of course, makes microprocessors, or CPU's, the cores of computers. It traditionally competed only in the PC space until it introduced its Opteron processor for servers in 2003, and usually had less than 10% of the market, with Intel having about 90%. From 2003 until 2005 AMD had a good run, gaining market share and hitting a stock price over $40 in January 2006. Then Intel struck back. By slashing its prices and promising a new generation of chips for later in 2006, Intel first cut into AMD's profit margins and then began to regain market share. The new chips introduced in 2006 had their faults but were overall on par with AMDs. Then came the ATI acquisition, which saddled AMD with a lot of debt. To add to this toxic brew, AMD's quad-core chips, code named Barcelona, were delayed in 2007. For more AMD background and history check out my AMD page.

Because of the necessity of competing on price and the poor sales of ATI chips after the acquisition (competitor NVIDIA introduced new graphics chips that put ATI behind), AMD has had a very bad year. The good news is that the stock is now astonishingly cheap, if you are willing to pick up on the risk. Last week a wealthy shiek bought some $600 million's worth, so you would not be alone in your gamble. [I own AMD stock and also owned ATI stock before the merger. I also once owned NVIDIA stock but foolishly sold it.]

Lets use the numbers from AMD's latest quarter, Q3 2007, as a proxy for its current condition [See also my AMD October 18, 2007 Analyst Conference Summary]. Revenues were $1.63 billion, up 18% from Q2 2007 and up 22% from year-earlier.

Which would seem to say that AMD is accelerating out of its 2006 slump, at least on the revenue side.

Q3 2007 net loss was $226 million, which tells you the core story of AMD's low stock price. While it was better than the net loss of $457 million in Q2, it was a reversal of a net income of $121 million in Q3 2006. Q3 2006 in itself was not a good quarter for profits.

The main reason I think AMD is underpriced is because in Q3 2007 it reached positive cash flow. That is, much of the net loss was for non-cash expenses, including ATI acquisition costs. AMD spent the money on ATI last year, but under generally accepted accounting principles (GAAP) some of that expense is written off over time.

Also, AMD shipped a record number of processors in Q3. Despite predictions by Intel fans that AMD would sink back into the obscurity from whence it came, and claims that AMD notebook chips sales would be down, AMD's processors for notebook computers had 68% better unit shipments than the year-earlier quarter.

After some real problems marketing graphics chips after the ATI acquisition, that segment seems to be getting on track with a 29% sequential increase in graphics revenue.

Intel has always led AMD in process technology (the ability to put more logic gates in the same area of a chip). Just a few years ago AMD was typically two years behind Intel. This month Intel introduced chips using 45nm technology, though they don't seem to be actually available in any quantity yet. AMD said it 45nm chips will ramp in Q2 2008. My guess is that means a May release, so AMD has closed that gap down to 7 months. But AMD has a lead in putting 4 processor cores on a single semiconductor chip, and in design architecture.

If AMD were a recent startup investors would probably consider it a tiger, given all these trends. Intel is a heavy competitor, of course, so heavy that many nations are looking into Intel's reputation for illegal anti-competitive practices and AMD had to file a lawsuit to force Intel to allow computer makers to by AMD chips without retribution.

While Intel is AMD's main danger, a secondary risk would be the possibility of a slowing of the expansion of the computer market. Given the booms in India, China, Russia and elsewhere, I'd prefer to bet on a strong global market even if the U.S. market weakens.

One place AMD shines is giving its customers (the computer makers) what they want, in contrast to Intel's take-it-or-leave-it approach.

Another downside concern is the slow pace of introduction of quad-core Opterons. If AMD does not get its act together on this soon, it could lose all the gains it made since 2002 in the relatively high profit margin AMD market.

On balance, I see AMD growing in 2008, with positive cash flow and progress probably positive GAAP net income by the second half. AMD's stock price probably won't move above $15 per share until this current run of market turmoil ends and AMD shows 2 consecutive quarters of profitability.

But with a current revenue annual run rate of $6.5 billion, and apparently accelerating, I think a rational market would put the market cap at no lower than $7 billion. Which works out to $12.65 per share.

And if indeed AMD continues to gain revenues as quickly as it has lately (possible with its new Phenom processor, quad-core Opteron, and Spider graphics system), it could go a lot higher a lot sooner than most analysts and investors think.

AMD believed it would have a long-term competitive advantage if it became the only company with high-end graphics and general microprocessor capability. It might still be proven right.

Strangely, while I've been writing this its AMD's stock price has risen to $11.05, from $10.55 or so when I started writing.

More data:

www.amd.com
competitors:
www.intel.com
www.nvidia.com

Monday, October 29, 2007

AMD Virtualization Wars

For years now AMD (AMD) and Intel (INTC) have been slugging it out in the microprocessor market. The next arena for this slugfest is virtualization, the process in which a single computer can run multiple instances of an operating system, or even entirely different operating systems (typically Linux and Microsoft).

Intel has been the dominant player in the microprocessor industry since the 1980s and still dwarfs AMD. But AMD gained a great deal of market share earlier in the new millennium with its Opteron and Athlon processors. Unfortunately it did not gain enough to prevent Intel from having a comeback in 2006 that has extended up until at least now. The main cause of AMD's insufficient gains during a period of time when it processors outperformed Intel's on any basis was human nature. Most businesses had standardized on Intel processors; many did not even consider taking advantage of the AMD proposition. I think a lot of this was influenced by the broad ownership of Intel stock among key decision makers, but I can't prove that without doing some costly research.

The question for those owning AMD stock (including me) and those rare people considering buying it should be: what will happen in Q4 2007 and in 2008. This will be highly dependent on the new quad core Opteron processors (code named Barcelona) and their virtualization features.

There are plenty of technical articles being written on virtualization and on AMD's and Intel's versions of it. The reality is, however, that until there have been a number of large-scale commercial installations that use virtualization 24 hours a day for long periods of time, we won't know if one design is superior to the other.

Virtualization seems like no big deal on a desktop or notebook computer. It is basically making possible in software what used to be done, in a clunky way, with a hard disk partitioned between two operating systems. Now using VMware or other virtualization software both operating systems can run at the same time, sharing the microprocessor, memory, and network bandwidth.

But according to Rackable Systems, making a server farm work well with virtualization is not so trivial of a task. Server farms already have issues with interconnecting all of their computers. There is an issue called load balancing: if you are, say Amazon (a Rackable client) and have thousands (or millions) of people searching your gigantic Web site all the time, and the Web site has to be coordinated with a database and a shipping system and a content management system, well, in non-tech speak it gets hairy. A virtual operating system sitting idle is just a waste of expensive equipment.

Rackable has a new line of servers ready for introduction that are engineered from the ground up for virtualized server farms. Probably IBM, Dell, Sun and HP are working on this issue too and will make their claims.

It will sure be interesting to see if this virtualization-ready line is available in only one flavor, AMD or Intel, or if it available for both.

Another area of interest in the AMD versus Intel battle is processors for notebook computers. In this area I see pundits constantly reporting that AMD cannot possibly compete. Yet it seems like every quarter AMD picks up a bit of market share in notebook computers. Partly this is because Intel started with so much market share and does have good notebook technology. Intel likes to charge its customers through the nose when it has no competition (hence accusations of monopoly pricing). Surprise, most people don't feel like they can pay $2000 for a notebook computer that will be broken on obsolete in 2 years. People really like computers to be in the $500 to $700 price range, and to make a profit in that market many notebook computer manufacturers have turned to AMD. With AMD readying a new line of notebook-optimized microprocessors, motherboard chips and graphics chips, it is possible that it may gain even more market share in 2008.

I consider investing in AMD to be a risky business. Investors should always balance the risks in their portfolios by diversification.

More data:

AMD home page
Intel home page
virtualization
My AMD page with links to summaries of AMD analyst conferences
VMware virtualization

Monday, August 6, 2007

Microchip, Housing, and Semiconductor Demand

Microchip (MCHP) reported Q2 revenues last Tuesday that were down from April guidance but slightly better than they had warned about in June. Revenues were $264.1 million, up 2.3% sequentially from $258.2 million and up 0.6% from $262.6 million year-earlier. That is not exactly high-growth but it is better than many semiconductor companies reported, notably Motorola (MOT).

The slump in the semiconductor chip industry, which appeared to have have hit bottom for most companies in Q1, was mainly caused by inventory restructuring at end customers and distributors. Given that, while unit shipments increased in many cases, pricing was weak and the result were drops in revenues. Of course the semiconductor industry is complex. Many companies reported weak demand for telecommunication infrastructure chips as a result of industry consolidation. Motorola also lost market share in the low-end of the cell phone market to Nokia.

Microchip had predicted a 5% sequential rise in revenues. At their analyst conference (See my summary) they accounted for their 2.7% miss of that prediction partly because of the housing industry slump. You might think that houses don't have microcontrollers and other semiconductor chips in them, but you would be wrong. Microchip supplies parts that go into five things that can go into new home construction: garage door openers, security systems, thermostats, irrigation equipment, and airconditioning. Yet despite the rather broad slump in the housing market, this did not cause Microchip revenues to fall; instead they failed to grow as quickly as expected by about 1%. It's stock price has held up well because it pays a hefty dividend and has good profit margins even in weak periods.

Slow housing construction hits specific industries like lumber and concrete pretty hard, but many economic sectors are only affected by a drop in housing construction if that is accompanied by lower consumer demand. In addition, most U.S. semiconductor companies do more than half their sales outside the U.S.

Microchip's management, humbled by missing guidance, now only expects revenues to be flat to up 2% in Q3. Usually demand in Europe is slow in Q3 because the whole subcontinent goes on those nice vacations only the elite are allowed in the U.S. But Q3 is when Asian manufacturers have to gear up for Q4 sales in the U.S., and while demand in India & China is not quite as seasonal, it is growing rapidly.

While not wanting to minimize the pain of individuals who took out adjustable rate mortgages, or became the ultimate owners of those mortgages, after the Federal Reserve had (belatedly) started raising interest rates, I don't think that particular situation says much about the economy at present. Housing demand always returns when prices get low enough to entice cautious people back into buying. Some paper profits will dissappear, but very little actual money will be lost on net by anyone.

It is just one more reminder that if you can, buy at the bottom of the cycle when others are panicking. Don't be the greater fool. Refrain from buying when prices lose touch with reality.

Note: I own Microchip stock

More data: Microchip investor relations page

Saturday, July 21, 2007

AMD Ambitions Still Burn

A humbler but unbowed AMD management reported Q2 2007 results to analysts on July 19 and answered questions. There is no doubt that AMD is on the ropes and Intel, just 3 years ago a complacent, fat, slumbering giant, is wide, awake, mad, and pounding the hell out of them. But that does not mean that AMD investors are out of the money in the long run. At least there is a referee now; Intel's mafia-goon tactics from the 1990's are being watched by courts and regulatory agencies around the world.

Despite Intel's bluster at its own analyst conference on Tuesday (see my detailed summary), it appears that AMD actually gained overall market share again in Q2. We'll have to wait until independent results are released to be sure.

But there was a lot of bad news at the AMD conference (see my detailed summary). Earnings were highly negative at a $600 million loss or $1.09 per share. Even if you pull out $130 million in one-time charges, they lost $470 million in the quarter on revenues of $1,380 million.

They also wrote off $30 million in obsolete processor inventory. What I believe happened is that they sold a lot more obsolete stock that went into low-end desktop and notebook computers. Given the price war between AMD and Intel, even up-to-date stock may have been sold at less than the cost of production.

So are the folk at AMD committing suicide? Maybe not, but they are taking a gamble. Intel has immense depth in resources; they are far more able to play the price war game than AMD. In 2005 that was okay, because AMD processors were acknowledged as superior in the industry. But in 2006 Intel brought out a new generation of processors that are roughly on par with AMD, so pricing became a serious issue. Intel in Q2 2007 had almost as much net income as AMD have revenue, making it seem like back in 1997 when AMD was allowed to live only so that Intel could not be accused of being a monopoly.

Now the good news for consumers and maybe even AMD investors: blood still flows in AMD veins. Management still plans to beat Intel in the long run, and it is still a credible story. You see this in three areas: the release of Barcelona processors due this quarter; the minimization of layoffs; and the coming push at the graphics market.

Intel now has a long history of spending a lot of money to convince people that its processors are wonderful, but doing inelegant engineering that does not really serve the end customers well. AMD continues to take a different approach: they are relatively quiet about their upcoming processors. Because, I believe, secrecy serves them well: if they announce what they are doing, the engineers at Intel will start copying them. Like they did with the transition to 32/64, with dual core, with energy efficient chips, and with on-chip memory controllers.

Lots of IT departments and consumers still have not bought anything with an AMD chip in it. Some people are just conservative, or own Intel stock. They did not by Opteron server processors or Athlon desktop chips when they had a clear lead on performance, performance per price, and performance per energy use. Some people are just going to stick with Intel.

But Toshiba, one of the last holdouts, is now offering computers based on AMD processors. AMD would not say exactly how much impact this had on Q2 revenues; I suspect almost none. Most Toshiba notebook and desktop pcs will still run on Intel chips, but it is still a gain for AMD and a loss for Intel.

In fact the PC makers like having two sources to play off against each other.

Intel laid off 12,000 employees in the last year; AMD has laid off some 500. Intel cut fat, which is good for everyone except the fired employees; AMD can only cut lean, so it is trying to do no cutting at all.

AMD is also in a tough fight with NVIDIA. AMD bought NVIDIA's traditional rival, ATI, which was doing poorly in many ways, and which has been a major factor contributing to AMDs recent string of losses. But again, there is reason behind this short-term madness.

AMD only made computer chips (and memory, but it divested that) in the past. There can be benefits to specialization, but we are in an era of consolidating more function into less silicon. ATI makes "chip sets" as well as graphics chips. Chips sets are the glue on computer motherboards. Intel was already making processors, chipsets, motherboards, and low end-graphic chips. Now AMD has a full set of solutions for computer assemblers. At the same time AMD has made it clear that it will keep its solutions "open" so that they will work with chips from NVIDIA and other makers. Computer manufacturers can make the best design for their customers.

If you don't own any AMD stock I would not advise buying any at this point unless you like a very high risk factor. I own AMD stock. But take a good look at Barcelona when it comes out; that is the real key to the short-term value of AMD stock.

Wednesday, July 18, 2007

Intel An Indicator for AMD?

Yesterday Intel (INTC) reported results for Q2 2007. Tomorrow rival AMD (AMD) will report results. What was only a few years ago a battle between a David and a Goliath has settled into a a long war between two giant companies (with Intel still being several times larger than AMD by any measure).

Despite the enormous progress AMD made between 2002 and 2006, becoming the clear technology leader and gradually eating into Intel's overwhelming market share, AMD investors are little better off than they were a few years ago. Long-term relationships are hard to sever; just when AMD looked like it had done the job in 2005, Intel finally was nearly ready with new microprocessor technology that had similar capacities to AMD's (and Intel fans claim its chips are superior). And let's face it: AMD got cocky too early. They bought troubled graphics maker ATI at about the same time that Intel introduced its Core Duo processors; they failed to maintain a clear lead in technology.

Intel's not very secret weapon was, and is, being ahead in fabrication technology. This is usually referred to by the scale of the lithography used, with smaller being better. Thus when Intel had 65 nm (nanometer) chips, AMD was still using 90 nm. Now that AMD is using 65 nm, Intel is ready to introduce 45 nm, which AMD won't be able to match until late in 2008. Smaller scale means more transistors (or gates) per chip.

All this meant that the inefficient processor design that Intel had evolved over two decades was able to keep roughly on par with AMD's better design (notably its on-chip memory controller). In Intel's defense it has to be said that rather than staying with the legacy designs it had planned to move to the all 64 bit Itanium processor, which is a great piece of technology that never caught on because of the enormous task of re-writing computer programs to work on it. AMD's scheme to evolve towards 64 bit processing (using chips that can run both 32-bit and 64-bit software) eventually had to be copied by Intel.

In 2006 and into 2007 profits for both companies plunged as they competed on price as well as technology. AMD has to gain market share or it will not be able to make the investments it needs to stay ahead of Intel in this race. Intel needs to defend its market share. As a result PC makers and end customers have gotten great deals on processors.

Yesterday Intel's revenues were down sequentially, but well up from its disastrous Q2 2006. It has been doing well selling pseudo quad-core server chips. AMD made the mistake of waiting until it could bring out true-quad core chips, which should happen in Q3.

The mid term future (6 months to a year) value of AMD versus Intel depends heavily on how good AMDs new chips are. As with all complex technologies, there will be room for differences of opinion. As has been the case in the past, Intel's vast advertising budget will mean that most reviewers will give Intel at least the benefit of the doubt on any issue. Also, the Intel chorus will sing "Don't buy AMD, wait until the next generation of Intel" loud and sweetly, a ploy that has worked well thus far.

You can read my detailed summary of the Intel results and analyst conference at Intel July 17, 2007.

It is generally believed that tomorrow's AMD results will a continuation of a disaster story, if possibly not quite as bad as Q1. You might want to read up on my summary of AMD's Q1 results and conference.

It is possible, given Intel's report, that AMD did better than expected in sales of low-end notebook and desktop chips. This is not necessarily a good thing. It keeps AMD market share while AMD waits to introduce new high-end products, but it may mean bad margins and lower than expected earnings (already expected to be deeply negative).

Intel did well with its quad-core server chips; that might really hurt AMD. Its dual-core Opteron servers had been the mainstay of its profitability.

Another thing to look for tomorrow is how well its graphics chip division (formerly ATI) did. NVIDIA has been kicking ATI around for a couple of years now. AMD has promised a turn around, but I doubt they were able to accomplish that in Q2.

If you don't own Intel or AMD already (I own AMD) neither stock looks like a solid bet right now. AMD's low stock price may be attractive, but as long as the processor price wars continue, it is the stockholders of Sun, HP, Lenovo and Dell that will benefit.

More data:

My Intel page
Intel Investor relations page
My AMD page
AMD Investor relations page

Tuesday, July 10, 2007

Marvell Releases 10-Q with Fiscal Q1 2008 Results

Marvell released a 10-Q giving its results for the quarter ending April 28, 2007 (Fiscal Q1 2008) yesterday. In the press release of Q1 results, and at the analyst conference (See my summary), only the revenues had been reported due to the ongoing stock option accounting restatement. The filing of the 10-Q confirms that the process is complete and accounting issues have been resolved.

If you have been following the Marvell story the 10-Q confirms what has been broadly guessed. Revenues were $635 million, up 2% sequentially from $622 million in Q4 2007 and up 22% from $521.2 million for fiscal Q1 2007. This annual increase of revenues was largely due to two major acquisitions, Intel's XScale division and Avago, a specialist in chips for printers.

Also mainly due to the acquisitions earnings swung from a healthy profit of $77.6 million year-earlier to a substantial loss of $52.9 million. This was expected because the larger acquisition, XScale, was losing money when acquired and is not expected to become accretive to earnings until calendar 2008.

Cash from operations was $54 million, up from $47 million year-earlier. Contrasting this with the GAAP earnings shows that many of the charges against earnings were non-cash.

The real question is, what does the future look like? The XScale processors are meant to compete in the high-end cell phone market. They were not chosen by Apple for the iPhone, though at least one chip by Marvell was. However, even under Intel the revenues from XScale were growing at a good pace. By combining the XScale processor with other highly-regarded Marvell analog and digital components, it is possible (but not guaranteed) that traction will be good. By increasing XScale revenues and decreasing production costs through outsourcing, Marvell expects to make the XScale business profitable some time next year.

XScale is not the only trick Marvell has to pull off. It did so well in creating chips that go in disk drives that it seems unlikely that it can gain much more market share there, and that market itself is not growing quickly. So Marvell is pressed by the need to innovate, winning share in other markets and entering new markets. One area of optimism is chips to improve performance of high-definition displays.

The main danger for Marvell, other than a macroeconomic downturn, is failure to expand XScale revenue. For now Intel is producing the processors and Marvell is obligated to buy them even if it finds itself unable to sell them.

More data:

Marvell investor relations page
My Marvell (MRVL) page

Tuesday, July 3, 2007

Marvell's New Clothes

Marvell (MRVL) is one of the many technology companies that has been under a cloud due to stock option dating issues. The cloud for investors, in most cases, has not been so much the risk of an SEC investigation or related train-wreck scenarios. Rather, while the companies do their accounting restatement investigations they do not release information that make informed investing possible. They do not release earnings; they don't file 10-Q's or 10-K's with the SEC.

On July 2nd Marvell made some of its tardy filings with the SEC. They had already announced the completion of their restatement investigation back in May. The filings included a 10-K covering fiscal 2007 through January 30, 2007. The most recent 10-Q was for fiscal Q3 2007. What we really want to see, of course, are the numbers for the quarter (Q1 2008) that ended April 28, 2007 (See my summary of the May 17 analyst conference for what we know so far).

Still, there is a lot of grist for analysis in the 10-K. For starters, for all practical purposes Marvell made no profit for its year. Out of annual revenues of $2.24 billion (well up from $1.67 billion in fiscal 2006), it lost $12 million, or $0.02 per share. In 2006 it had profits of $199 million, or $0.32 per share.

Time to flee the stock? Sorry, too late. The share price topped out at over $38 back at the beginning of 2006; it ended today at $18.32, and has been as low as $15.25 this year.

Let's try another reality check, since revenues have risen rapidly while earnings have fallen. What happened to cash?

In fiscal 2007 net cash provided by operating activities fell, but not as precipitously as earnings. It was $402 million in 2006, then $337 million in 2007. There was a serious cash outflow, but it was for acquisitions, not for operations.

Marvell made two big acquisitions in 2007: Avago (chips for printers) for $262 million in May and Intel's cell-phone (XScale) microprocessor division for $600 million in November. These acquisitions had serious one-time costs and the Intel division in particular had major ongoing net losses when it was transferred.

Marvell also acquired UTStarcom in February 2006 for about $40 million. According to a neat little chart on page 126 of the 10-K, the acquired businesses, if they had been bought at the beginning of fiscal 2007, would have created an overall loss of $475 million for the year, while bumping up overall revenue to $2.6 billion.

Are these guys stupid? Did the people who created Marvell out of some analog chip designs and built it to a $1 billion business in a few years lose their minds? I don't think so.

They did load up heavily on risk. A good way to get a sense of risk, which is mainly from the Intel purchase, is to use your find function on the 10-K with "Intel" as the search string.

To get back to being heroes for their stock holders Marvell employees have to turn the acquired companies to profitability while continuing to grow their more established business lines. The big bear here is the XScale application processor. Under Intel its revenues were growing, but R&D costs were high compared to revenues, resulting in division losses. Marvell intends to move production from Intel's fabs to its fabless (outsourced) model by mid-2008. That will allow some cost reductions. But mainly Marvell has to rapidly increase sales while keeping prices firm.

These XScale processors go mainly into high-end (3G) cell phones and combine well with Marvell's analog technologies. So Marvell has bought into the cell phone market in a major way. But many other firms, like Texas Instruments and Motorola, are also gunning for the high-margin, high end of that market. The market is expected to grow rapidly, spurred on by the introduction of the iPhone.

Do I know if Marvell can succeed in this market? No, I just don't know. They have a great track record. If they can ramp the XScale business and decrease costs, today's stock price may look way low a year from now. But if XScale tanks, which is a distinct but hard-to-quantify possibility, then today's stock price will seem to have been based on optimism, not solid value.

I own Marvell stock, but I acknowledge the future is hard to see. It would really help if management would break out revenues and costs by division at the next analyst conference.

More data:

Marvell investor relations page
my Marvell page

Wednesday, December 13, 2006

AMD Leaps Over Intel

I wrote recently about Intel's attempt to leapfrog AMD by selling a quad-core microprocessor well before one became available from AMD. I explained that the Intel processor did not really have four cores on one chip. Rather it has two dual-core chips mounted into one pin package. Despite some other shortcomings the Intel chip packs a lot of computer power into one package.

Now it appears AMD has leapfrogged Intel, and only a few weeks later. According to Tom Yager, Infoworld's chief technologist at its test center, he has been given a pre-release AMD Quad FX (aka four by four). To get four cores on a board it has two sockets each accepting a dual core chip.
What is the difference? It is all about architecture and the future road map.

As usual the Intel processors need a separate chip for memory management. For years AMD has had its memory management on its processor chips, a considerable advantage. Why did Intel not just fit a memory manager onto its chips? Because of other inefficiencies in the Intel design they don't have the real estate available on the chip.

The AMD four by four chips have two complete CPUs on the chip, with independent level 2 cache, independent memory controllers, and independent buses. This makes the chips much more scalable than Intel's and better able to handle multiple programs running simultaneously.

Also, the current socket AMD is using allows for upgrades to true, single-chip quad-core processors that are due out in mid 2007. So in a few months if you have the bucks and need the high-end experience (say you are a 3-D graphic designer, or just like to watch video while you run your stock market or engineering applications) you can be running eight highly efficient cores all at once. This will work particularly well with Windows Vista.

Certainly with its tremendous marketing muscle and army of talented engineers, Intel will be unleashing its own dream processors as time goes by. But for performance, for bang to the buck, and for power savings, AMD is now the safe bet.