AMD's Close Future Analyzed: What Tomorrow Brings?
4/27/2012 by: Marcus Pollice
Last week AMD presented their first quarter 2012 earnings numbers. This is an in-depth analysis of what was said during the call.
Rory Reed made a great display of making very elaborate statements that lacked actual content when examined carefully. He constantly uses a number of keywords to basically say "what we do is the right thing". As always he also emphasized the continued focus on execution.
Manufacturing
Supply issues on the 32nm node can now be considered a thing of the past. The executives explained that there have been constant improvements in yields allowing them to meet customer demand. While not giving detailed guidance about the improvements being made, the CFO remarked that AMD "we are not any longer supply limited on that node". For the coming quarter, incremental yield improvements are expected.
During the call it was revealed, that during the second half of 2012, tape-outs and also the production of 28nm APU products will commence at GlobalFoundries, having a slightly negative impact on gross margin. Some of these chips will be sold for revenue even in 2012. This is done in preparation for launches in early 2013.
AMD Roadmap for desktop and mobile APUs and a single CPU silicon. Steamroller is a follow-up to Piledriver performance core, while Jaguar is a successor to low-power Bobcat architecture.
The executives didn't provide any more detail as to which 28nm products will start to be produced in 2012. Earlier this year we unveiled that AMD will transition their entire 28nm lineup to bulk silicon, with the successor to Trinity called Kaveri being manufactured at GlobalFoundries, planned for 2013 ramp. However, it could also be any of the Kabini or Temash APUs, which are planned to replace the Bobcat-based 40nm chips for the Brazos / Brazos 2.0 platform. In that case, we would expect that mass production begins during third quarter 2012, in order to have the chips available for the launch on CES 2013, two years after the debut of the original Brazos platform with Ontario and Zambezi APUs.
Rory Read explained that the ramp of 28nm GPU products is proceeding successfully. Even though there is a constant stream of reports about capacity constraints at TSMC, the foundry is able to meet all of its commitments to AMD. They are still seeing strong demand for the Radeon HD 7000 series products which is expected to continue. The CEO admitted that the constrained environment limits their possibilities of expanding this business.
Notebook / Client
In the notebook segment, AMD made great progress. This can almost solely attributed to the success of the Ontario, Zacate and Llano APU products and their respective chipsets. Overall AMDs shipments in this space are nearly 100% APUs and up 30% compared to last year. The company is particularly successful in the $400 to $700 mainstream price band, which makes up about 50% of the market. Brazos continues to be very successful too, especially in emerging markets.
Going forward, AMD sees further growth opportunities, especially with the upcoming Trinity APU, improving performance per watt almost twofold. Most of this increases come from the GPU, however CPU performance is going to increase as well. Compared to Intel it will still not hold up, but AMD believes the types of software used by their consumers benefit more from a faster GPU. At least in terms of design wins, AMD can already tout a huge success.
During the conference call there was also a lot of spin about Brazos 2.0. From a technical perspective, it is not very exciting since it's basically the same as Brazos with slightly higher clock speeds and minor power optimizations. The chip is still the same manufactured at 40nm at TSMC. Nevertheless AMD seems to be able to sell these like hotcakes. Since its introduction in January 2011 AMD was able to sell over 30 million Brazos units, i.e. selling at least 60 million AMD chips (APU+Chipset).
With their 17W models of Trinity AMD also wants to tackle what they call ultrathin notebooks, as they can't use the Ultrabook trademark which is owned by Intel. AMD takes a slightly different approach and attacks the mainstream price segment. Rory Read on numerous occasions criticized Intel for only catering to the premium segment. AMD wants to bring ultrathins to the average customer who is unlikely to spend a grand on such a device.
To put it with the words of the CEO: "Ultrathins are for everyone," and continued: "...ultrathins and thin and light are going to be a core product across the industry. This is nothing new. Thin and light has been going on for the past 15 years. We're taking it to the next level and we're going to bring it to everyone."
Our take is that there is a place for both types of systems. It remains to be seen however which one will make more profit. Intel tries to do it with margins, AMD with volume. This is already comparable to what happens in the traditional notebook space with AMD grabbing a lot of share in the mainstream segment with Llano.
Desktop
The CEO acknowledged that on the desktop AMD took a hit due to the initial problems with the 32nm SOI chips. Most interestingly he noted that 45nm CPUs (i.e. Phenom II) are still selling quite well, but 32nm offerings are now ramping up as the supply situation on 32nm improves. What he didn't mention is that for example Bulldozer is not that attractive given it's price point and performance.
Server
Regarding the server segment, analysts were concerned about ASP erosion. They apparently failed to understand that due to its performance characteristics, Bulldozer had to be priced competitively and couldn't help increase ASPs. Bulldozer-based CPUs now make up over 50% of AMDs server shipments. According to the CEO the company has worked out a strategy that allows them to grow their share without going all out on a low price strategy to capture more share. He especially highlighted the strategic acquisition of SeaMicro that will help them to that end. Of course for now there is no product based on the SeaMicro fabric from AMD, so this can be considered another way of saying "Shhhh move along, there is nothing to see here". But AMDs executives see this more as a long term strategy, so we will see how it plays out in the future.
Analysts tried to get a comment on the competitive situation numerous times, but Rory Read almost always dodged them. He explained how AMD is greatly positioned for the entry and mainstream server market and implied that is enough. He is confident that AMD has what it takes to tackle the changing market environment in the future. Rory strained that AMD is positioned in the server space for the long term.
Graphics
Due to the launch of various Radeon HD 7000 series products AMD was able to raise the ASPs in the graphics segment and thus improve revenues in a quarter that is usually down because of seasonal effects. On the other hand their game console licensing revenues were down a bit which can be attributed to seasonality. The Radeon HD 7000 series products were particularly successful in the channel.
In the second quarter AMD plans to expand the series into the notebook segment. Regarding this we'd like to note that while AMD was first to market with the 28nm GPUs on the desktop, in the mobile space NVIDIA will beat them to market. A lot of the upcoming Ivy Bridge based notebooks with discrete graphics will feature NVIDIA GPUs. AMD believes that as the market is transitioning to APUs, discrete graphics will become less important in the notebook space. Also AMD might not want to support Intel too much with their discrete options but rather get them bundled with Trinity later in the year for the Dual Graphics option.
Microsoft Windows 8
The CEO believes that Windows 8 will be mostly centered around the traditional x86 platform, despite all the buzz about Windows on ARM. Similar to Intel executives he touts backward compatibility as the main advantage of x86 based platforms. ARM competitors would first have to build up an ecosystem to be successful. He also acknowledges the need for low power devices but boasts that AMD is very well positioned in this regard.
PC market / HDD shortage
Regarding the HDD shortage following the flood catastrophe in Thailand, Rory Read doesn't believe it had significant impact on AMD. He said the supply chains are very resilient and the situation should be a thing of the past since the end of January for any company in the PC market. The only effect he recognizes are slightly higher prices that will continue into the next one or two quarters.
When it comes to the development of the PC market in general, AMD believes that in 2012 the outlook is very positive. The company is poised to take some share of that growth. When looking at the different geographies and the related business environments, the picture AMD paints is very much comparable to Intels view. Due to macroeconomic effects, Europe is considered to be a difficult environment.
On the other hand Rory Read claims that North America performed very well, when in fact any reputable market research outfit reports similarly bleak numbers. The truth of the matter is that for a considerable amount of people computing got a commodity in the developed Western countries, which prolonged upgrade cycles. The old machines are simply good enough as long as they don't break. As usual, anyone playing demanding games or doing serious business on their PCs is an exception to this rule.
Tags:
AMD, Advanced Micro Devices, AMD Q1, AMD 1Q, Revenue, Net Loss, Non-GAAP, GAAP, GlobalFoundries, 28nm, APU, Wafer Supply Agreement, SeaMicro, Rory Read, Brazos, Brazos 2.0, Trinity, 32nm, SOI, Bulk, Dell, Tyan, MSI, Fujitsu, Opteron, Opteron 3200, Fusion Center of Innovation, Nuvixa, AMD Ventures, gesture-based, video communication, GPU, CPU,
© 2009 - 2011 Bright Side Of News*, All rights reserved.