(original title: NVIDIA: Counting Chickens Before They Hatch)
As an analyst and investor, one of the main things I do is count chickens before they hatch. Today graphics chip specialist NVIDIA has a high price-to-earnings ratio (non-GAAP 62x trailing, 26x leading), indicating some investors are counting on a lot more chicken profits in the future than they have been seeing lately. Is this a smart assumption?
NVIDIA will have a profit boost for the next six years from its licensing deal with Intel. For the most recent quarter ending January 30, 2011 it booked $57 million for the litigation settlement portion of the deal (booked as a negative operating expense, not revenue). For each quarter it will book about $60 million in royalties.
Other than that, the last couple of years have been hard on NVIDIA. Q4 fiscal 2011 (the most recently reported quarter) revenues were $886.4 million, down 10% from $982.5 million in the year-earlier quarter. That does not sound like a high-growth company that deserves high PEs on its stock price. [for a fuller report on Q4, see NVIDIA Q4 fiscal 2011 analyst conference call summary]
What speculators are speculating on is a chip called Tegra. The first version was interesting but did not generate much revenue. The second version is available now in a few tablet and smart phone devices. It produced little revenue in Q4, but is supposed to contribute substantially to Q1. Therefore, instead of a normally seasonally down Q1, guidance is for a 6% to 8% sequential revenue ramp. That is impressive, if it happens.
In addition, the third version of Tegra, called Kal-El for now, is already sampling and is supposed to be in devices for sale for holiday 2011 shopping. Reviews of Tegra 2 are generally positive.
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 ever servers.
Before you pay a premium for all them chickens, you might want to think about all the other chickens that will be on the market. NVIDIA's Tegra CPU unit is based on the ARM architecture, which anyone can license. The NVDA advantage is in graphics, but there are a number of companies that license graphics capabilities that were designed specifically to work with ARM. NVIDIA's were designed to work with the 8086 architecture of Intel and AMD.
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 phones than NVIDIA does; so does TI. Intel and AMD want to get into the game. AMD gained a lot of market share against NVIDIA in discrete graphics cards for computers during 2009 and 2010; their fusion chips offer some extreme advantages, especially for tablet computing. 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.
None of the other contenders have PE ratios as high as NVIDIA's. I have owned NVIDIA stock in the past, and if it had a low PE I might scoop up those chickens right now. I have always admired NVIDIA's technological skills, and the Tegra 3 is promissing. I don't see how it can be all that much better than competitors, however. Everyone has promissing designs, everyone is hustling to squeeze usability out of the same limitations of silicon.
I have done well recently with some hatchlings at Dot Hill, Dendreon, and TTM. On the other hand, I like to keep in mind Anesiva, where I correctly predicted its Zingo product would hatch (get FDA approval), only to watch it die a horrible death and then take the entire company into bankruptcy with it. The secret is to see them chickens before momentum investors drive up the stock prices. If you bought NVDA a year ago at $8.65 per share, congratulations. If you are thinking about buying it today at $25.48, you might want to consider that it could take NVIDIA a couple of years of outstanding growth to justify this price.
Showing posts with label smart phones. Show all posts
Showing posts with label smart phones. Show all posts
Thursday, February 17, 2011
Wednesday, January 12, 2011
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
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
Thursday, August 26, 2010
Marvell Technology (MRVL) Mobile Invasion
Marvell Technology Group had a mixed fiscal second quarter (Q2) 2011 ending July 31, 2010. Revenue fell below prior guidance. On the other hand it was was $896.5 million, up 5% sequentially from $856 million and up 40% from $640.6 million in the year-earlier quarter.
Wall Street was not suprised at the short fall because the PC supply chain was believed to have hit the pause button in July. Marvell's biggest product segment is semiconductor chips that go into hard drives (HDDs) for data storage. HDD revenue was 15% down sequentially from Q1. There had been supply constraints over the last year, and as a result HDD manufacturers had overbuilt inventory preparing for back to school sales. This was largely from the Europe economic effect, not so much an actual downturn in demand in Europe as cautious inventory control by spooked CEOs. Marvell management believes consumption of excess inventory is mainly behind, as orders rose again towards the end of July.
Making Q2 not such a bad quarter was the long-promised ramping of Marvell products for the mobile/wireless markets. This segment saw an over 50% sequential jump (and up 140% y/y) as silicon was shipped for a number of new products that consumers (mostly in Asia) will see in Q3. About 15% of sequential growth was due to to an Armada based product, with a single customer getting ready for a game product introduction. 30% of sequential growth was due to a cell phone processor for a customer ramping a new product. 55% of sequential growth was for embedded Wi-Fi, which is a market the Marvell has been successful in for some time, and is still building on that success. Management believes it will see 15% to 20% sequential growth in Q3 in the mobile and wireless segment.
Earlier this decade Marvell made very large commitments to R&D and also made some acquisitions that analysts questioned at the time [See Marvell's Huge Research and Development Budget August 24, 2007]. Now we are seeing fear in the hearts of competitors in a slew of fields that Marvell did not even play in five or ten years ago.
Take one known example from the mobile device market. Hawang has been producing eReaders for the Chinese market since 2008. The Marvell based model is due out by September. According to Hawang it offers "better performance at a better price ... true mass market pricing." This is possible because of Marvell's intellectual property and ability to put many functions on a single chip. Notably this single chip solution (SoC - System on a Chip) includes an integrated e-Paper Display (EPD) controller. It can display standard pdf documents and requires very little power. I also expect we will see OEM announcements of tablet computers that are based on Marvell chips as we progress through the year. These may be in Asian markets, but some should be available in the U.S. Keep in mind that this is a very competitive market. In addition to Apple's internally developed A4 chips (which almost certainly won't be sold to competing OEMs), top competitors include Broadcom, Qualcomm (Snapdragon chips), Intel, TI, and Freescale.
With its low-power application processors, cell phone modems, wi-fi and bluetooth technologies, all possibly on a single chip, Marvell has apparently won a leading position in China's new Ophone (smartphones for new cellular signal standard) production, which should begin ramping around the beginning of 2011. That, in itself, will be a huge business.
Another space that Marvell is already competive in and may dominate sooner than you think is Ethernet switching. This has become increasingly important with the proliferation of datacenters and cloud computing.
Most intriguing is the idea that Marvell is going to use products intially developed partly using the Intel division it acquired to go after the server market. Marvell is not the only company that can see using ARM architecture to design servers that sip power, but it does bring a formidable amount of talent to the table. Let's see: really fast, very low power consumption processors, networking chips, and storage chips. Sounds like the makings of a cloud datacenter to me.
Probably my favorite thing about Marvell managment is that they think long term. They have proven they can take a concept, get out a product a few years later, and dominate a market a few years to a decade after that. You might want to take notice of Marvell now, because by 2012 everyone will be noticing.
Resources:
Wall Street was not suprised at the short fall because the PC supply chain was believed to have hit the pause button in July. Marvell's biggest product segment is semiconductor chips that go into hard drives (HDDs) for data storage. HDD revenue was 15% down sequentially from Q1. There had been supply constraints over the last year, and as a result HDD manufacturers had overbuilt inventory preparing for back to school sales. This was largely from the Europe economic effect, not so much an actual downturn in demand in Europe as cautious inventory control by spooked CEOs. Marvell management believes consumption of excess inventory is mainly behind, as orders rose again towards the end of July.
Making Q2 not such a bad quarter was the long-promised ramping of Marvell products for the mobile/wireless markets. This segment saw an over 50% sequential jump (and up 140% y/y) as silicon was shipped for a number of new products that consumers (mostly in Asia) will see in Q3. About 15% of sequential growth was due to to an Armada based product, with a single customer getting ready for a game product introduction. 30% of sequential growth was due to a cell phone processor for a customer ramping a new product. 55% of sequential growth was for embedded Wi-Fi, which is a market the Marvell has been successful in for some time, and is still building on that success. Management believes it will see 15% to 20% sequential growth in Q3 in the mobile and wireless segment.
Earlier this decade Marvell made very large commitments to R&D and also made some acquisitions that analysts questioned at the time [See Marvell's Huge Research and Development Budget August 24, 2007]. Now we are seeing fear in the hearts of competitors in a slew of fields that Marvell did not even play in five or ten years ago.
Take one known example from the mobile device market. Hawang has been producing eReaders for the Chinese market since 2008. The Marvell based model is due out by September. According to Hawang it offers "better performance at a better price ... true mass market pricing." This is possible because of Marvell's intellectual property and ability to put many functions on a single chip. Notably this single chip solution (SoC - System on a Chip) includes an integrated e-Paper Display (EPD) controller. It can display standard pdf documents and requires very little power. I also expect we will see OEM announcements of tablet computers that are based on Marvell chips as we progress through the year. These may be in Asian markets, but some should be available in the U.S. Keep in mind that this is a very competitive market. In addition to Apple's internally developed A4 chips (which almost certainly won't be sold to competing OEMs), top competitors include Broadcom, Qualcomm (Snapdragon chips), Intel, TI, and Freescale.
With its low-power application processors, cell phone modems, wi-fi and bluetooth technologies, all possibly on a single chip, Marvell has apparently won a leading position in China's new Ophone (smartphones for new cellular signal standard) production, which should begin ramping around the beginning of 2011. That, in itself, will be a huge business.
Another space that Marvell is already competive in and may dominate sooner than you think is Ethernet switching. This has become increasingly important with the proliferation of datacenters and cloud computing.
Most intriguing is the idea that Marvell is going to use products intially developed partly using the Intel division it acquired to go after the server market. Marvell is not the only company that can see using ARM architecture to design servers that sip power, but it does bring a formidable amount of talent to the table. Let's see: really fast, very low power consumption processors, networking chips, and storage chips. Sounds like the makings of a cloud datacenter to me.
Probably my favorite thing about Marvell managment is that they think long term. They have proven they can take a concept, get out a product a few years later, and dominate a market a few years to a decade after that. You might want to take notice of Marvell now, because by 2012 everyone will be noticing.
Resources:
Marvell fiscal Q2 2011 analyst call summary
My Marvell analysis page
Marvell and Hawang E-Reader [Marvell release June 1, 2010]
Marvell Moby Technology site
Monday, June 7, 2010
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
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
Monday, March 8, 2010
Marvell Sees Inflection Point
Marvell Technology Group, Ltd. (MRVL) is led by Sehat Sutardja, who built the company on his invention of a better chip to enable hard disk drives. Marvell has been the leading (by market share) supplier of hard drive controller chips for years, and this segment still represented 50% of revenue for the latest quarter. Marvell has also branched out to a variety of newer areas, based on its ability to combine analog functions and digital functions on a single chip (called SoC, for System on Chip).
Marvell was hit hard by the recession: in fiscal Q4 2009, ending January 30, 2009, it had $512 million in revenues, down from a peak of $843 million in the quarter ending August 2, 2008.
Last week Marvell reported for fiscal Q4 2010, ending January 30, 2010, revenues of $842 million. GAAP net income was an amazing $205 million; cash flow from operations was $281 million.
But the future, while speculative, could be even more amazing. Sehat referred to a coming inflection point. An inflection point on a curve can be from going down to up or vice versa, but in this case he is talking about a dramatic increase is Marvell's revenue growth rate.
As I have pointed out in the past, Marvell spent a lot of money on research and development in the earlier part of this decade, and only cut back a little during the recession. This year new products that were sampled in 2009 are ramping into volume production, and new products sampling this year are going to ramp in 2011. What are those new products?
Most notable are "communication processors" for smart phones, particularly the oPhone beginning to be sold in China. In order to get a price point low enough for mass marketing in China, more features than ever before had to be integrated into a single semiconductor chip. Marvell is unique in being able to offer general application and signal processing, Wi-Fi, Bluetooth, graphics, and cellular modem. Over 90% of oPhone models (there are several companies building them) use Marvell chips.
This alone could be sufficient to cause the inflection point. While the oPhones are far less expensive than models sold in the U.S., Marvell gets very good profit margins on the chips it makes for them. Start multiplying these per-chip margins by the 100s of millions of likely oPhone buyers in China, and you can see the potential.
But there is more. Marvell has become an increasingly prominent player in high-speed Internet switching chips. The competition there is intense, but intense competition has not impeded Marvell for very long in other fields it has chosen to enter.
Many of the newer products will be based on a microprocessor line called Armada. The thing about Marvell is that unlike Intel, they usually don't just make a microprocessor and sell it to customers who put it in a socket. Marvell works with manufacturers to integrate exactly what they need for their products on a single chip that incorporates the microprocessor. This gives faster execution and communication times and reduces costs, while leaving Marvell an ample profit margin.
Expect new products to be announced based on Marvell technology all though 2010 and 2011 (and likely well beyond).
While competition is intense and there are many pitfalls, it looks to me like Marvell is going to become the very center of the semiconductor chip industry during this decade. Of course others have already entered the SoC business, and stand-alone chips will continue to be manufactured for equipment with volumes insufficient to justify an SoC design. I'm not saying the engineers at Intel, AMD, NVIDIA, Broadcom, Microchip, etc. are not very good, and all these companies have plenty of financial ability to compete however they like. But Marvell seems to have found how to
I have owned Marvell stock since January 2005.
For detailed results from Marvell's latest reported quarter, see my
Marvell Technology (MRVL) Q4 2010 analyst conference summary
See also www.marvel.com
Marvell was hit hard by the recession: in fiscal Q4 2009, ending January 30, 2009, it had $512 million in revenues, down from a peak of $843 million in the quarter ending August 2, 2008.
Last week Marvell reported for fiscal Q4 2010, ending January 30, 2010, revenues of $842 million. GAAP net income was an amazing $205 million; cash flow from operations was $281 million.
But the future, while speculative, could be even more amazing. Sehat referred to a coming inflection point. An inflection point on a curve can be from going down to up or vice versa, but in this case he is talking about a dramatic increase is Marvell's revenue growth rate.
As I have pointed out in the past, Marvell spent a lot of money on research and development in the earlier part of this decade, and only cut back a little during the recession. This year new products that were sampled in 2009 are ramping into volume production, and new products sampling this year are going to ramp in 2011. What are those new products?
Most notable are "communication processors" for smart phones, particularly the oPhone beginning to be sold in China. In order to get a price point low enough for mass marketing in China, more features than ever before had to be integrated into a single semiconductor chip. Marvell is unique in being able to offer general application and signal processing, Wi-Fi, Bluetooth, graphics, and cellular modem. Over 90% of oPhone models (there are several companies building them) use Marvell chips.
This alone could be sufficient to cause the inflection point. While the oPhones are far less expensive than models sold in the U.S., Marvell gets very good profit margins on the chips it makes for them. Start multiplying these per-chip margins by the 100s of millions of likely oPhone buyers in China, and you can see the potential.
But there is more. Marvell has become an increasingly prominent player in high-speed Internet switching chips. The competition there is intense, but intense competition has not impeded Marvell for very long in other fields it has chosen to enter.
Many of the newer products will be based on a microprocessor line called Armada. The thing about Marvell is that unlike Intel, they usually don't just make a microprocessor and sell it to customers who put it in a socket. Marvell works with manufacturers to integrate exactly what they need for their products on a single chip that incorporates the microprocessor. This gives faster execution and communication times and reduces costs, while leaving Marvell an ample profit margin.
Expect new products to be announced based on Marvell technology all though 2010 and 2011 (and likely well beyond).
While competition is intense and there are many pitfalls, it looks to me like Marvell is going to become the very center of the semiconductor chip industry during this decade. Of course others have already entered the SoC business, and stand-alone chips will continue to be manufactured for equipment with volumes insufficient to justify an SoC design. I'm not saying the engineers at Intel, AMD, NVIDIA, Broadcom, Microchip, etc. are not very good, and all these companies have plenty of financial ability to compete however they like. But Marvell seems to have found how to
I have owned Marvell stock since January 2005.
For detailed results from Marvell's latest reported quarter, see my
Marvell Technology (MRVL) Q4 2010 analyst conference summary
See also www.marvel.com
Subscribe to:
Posts (Atom)
