Without further ado, we are finally able to provide our in-depth analysis of AMD’s second quarter 2011 earnings report. We already covered the financial side of the release right after the announcement was made and will now dive deeper into what AMD told to investors and how the company plans to evolve in the future.
With a great part of their product lineup is now made up of Fusion products, it’s not surprising there was a lot of talk surrounding these products. Especially in the notebook segment those chips have been a big hit, as they already make up for 70% of their mobile shipments and help drive the ASP for those products upwards. Their A-series APU codenamed Llano, which has been launched towards the end of the quarter, was in strong demand by customers. AMD already shipped a million units of those chips and generally expects it to ramp even faster in the second half of the year than Brazos did following its launch in January. This is quite impressive, since AMD considers Brazos one of the fastest driving processors in their history.
Interestingly enough, AMD even specifically mentioned the Lenovo X120e notebook, which was one of the top selling commercial notebooks powered by a APU. This is somewhat ironic, considering that it is not even available all over the world, most notably not at all in Western Europe but also other parts of the world. Lenovo’s rationale for this is that they carefully analyze markets and decided not to offer it in certain markets. One might guess, this statement was not considered very helpful by a lot of people expressing the desire to purchase this notebook on their Facebook page.
AMD’s Roadmap Calls for Zambezi to be the last CPU, followed by Komodo APU.
For the coming year, AMD already has an aggressive strategy laid out, with successors of Brazos and Llano planned to launch in the first half of 2012. The mainstream APU Trinity, which has already been publicly demonstrated at Computex, is also poised to get a low power SKU with less than half the TDP of the current 35W Llano offerings. This chip is targeted at performance-oriented ultra-thin notebooks and could give Intel’s Ultrabooks a run for their money.
AMD also plans to target tablets with their APU products. Due to AMDs rather late entry into the netbook market, they haven’t seen a lot of cannibalization by tablets. Rick Bergman also pointed out, that only a fraction of their Brazos sales are going into actual netbooks, the rest being notebooks, desktops and all-in-ones. With their Z-series APU already launched, they show commitment towards this product segment, which AMD similar to Intel considers an additive computing device. The company is confident that their technology matches very well with the upcoming Windows 8 operating system, which is seen as a critical inflection point.
Regarding ARM-based SoCs, for the next year AMD remains confident that they won’t affect their bottom line in 2012. In general AMD considers the tablet market still very fragmented and outside of the iPad also rather small. Similar to Intel, AMD considers backwards compatibility of their x86 products as one of their key advantages and thus feels to be in a great position with their x86 and GPU cores.
AMD expects to increase their market share with their APU products, which should already materialize in the second half of the year. Thomas Seifert, CFO and current interim CEO at the company, strained, that at this point AMD primarily focuses on increasing profitability than share. This basically means that we shouldn’t expect drastic price slashes, unless the competitive situation changes dramatically.
The interim-CEO said, that Interlagos received very good customer acceptance, with numerous platforms said to launch later this year. The key to the product becoming a success is more performance in a given power envelope, according to AMD executives. Speaking about performance, we have already highlighted the lowering of performance guidance of Interlagos in a separate breakout article.
When asked how it will affect AMDs rather low market share in the server space, Mr. Seifert explains that while in this space changes don’t happen that quickly, he expects a material impact in the fourth quarter of the year. Regarding the launch schedule, AMDs Rick Bergman leaves no doubt, that the company is set to ship the Bulldozer-based products finally in August:
"There’s Interlagos, which is our server-based products, which will begin shipping initial production next month. Likewise, there is a desktop version called the Zambezi, which is geared more towards the desktop enthusiast market. That will also begin initial production and shipments next month."
According to AMD, companies in the HPC space will embrace Interlagos first. This comes as no surprise, as companies like Cray have already committed to initial shipments of Interlagos CPUs, which are drop-in compatible with their current platforms.
Turning to their graphics business, AMD sees the upcoming 28nm GPUs as an opportunity to increase market share in this segment. Up until this point, no more specific launch date than "later this year" was given. Obviously AMD happily reported that Nintendo chose a Radeon HD graphics solution for the Wii U, the successor to the highly popular Wii. According to AMD, currently about 2/3 of game consoles contain a GPU from them. Overall about 130 million consoles are said to ship over the course of the next three years, which provides a good business opportunity for AMD.
Something that wasn’t mentioned before is that Mr. Seifert repeatedly talked about the possibility of customized APU silicon targeted at markets adjacent to game consoles. At this point, no further details have been provided.
Due to the 11% sequential decline of revenue of the graphics division, there was a lot of talk about GPU cannibalization due to APU products and CPUs with integrated graphics from Intel. AMD executives don’t think they lost a lot of share in the graphics market in the second quarter and blame the loss mostly on seasonality, even though their report clearly states reduced sales of discrete mobile GPUs. As a side note, AMD also mentioned that they had record unit shipments and revenue in the professional graphics segment, which can be portrayed as winning on a low level.
AMD is quick to point out, that Brazos, their entry-level APU platform, certainly doesn’t cannibalize discrete GPU sales, as in the target segment, discrete GPUs rarely have been used much. Regarding Llano, which is ironically marketed with a discrete level graphics capability, the executives note that there is a discrete GPU option. They admit though, that at this point it can’t be said how that plays out.
Regarding the Intel graphics solutions, AMD thinks that it is still required to employ discrete graphics for it to be considered modern. Rick Bergman concludes from this, that GPU sales will be flat. Mr. Seifert then concedes, that in the long term low end GPUs will be replaced with Fusion-type products, which in the end is good for AMD as it moves low margin revenue to high margin revenue.
While AMD is now a fabless company, it has a good working relationship with their partners GlobalFoundries and TSMC. Due to the fact that all their 28nm are targeted for manufacturing on bulk processes, they can flexibly allocate production to those two partners. According to Mr. Seifert they "made good progress" on the 28nm process, especially on graphics. AMD already got working silicon in their hand and plans to lead the transition to 28nm GPUs. Another interesting fact about the manufacturing related side of business is, that AMD managed to reduce unnecessary inventory buildup. This is attributed to both being a fabless company now and improved planning processes.
The CEO search
Regarding the ongoing CEO search of the company, Harry Wolin, Senior Vice President, general Counsel and Corporate Secretary, briefly filled in the investor community that the board is happy with the quality of previously interviewed candidates. He declined to give a timeline, since finding the right candidate has a priority over quickly finding someone to fill this position. The board wants a person with the right vision, experience and track record and hopes to regain share value over uncertainties, that might have driven off investors once that person is found.
Alluding to the present situation at AMD, Mr. Seifert said slightly amused "I told you I’m not sitting here warming a chair." While Mr. Seifert has been pretty clear that he doesn’t intend to hold on to that position any longer than necessary, he might get pressed by the board to take the job permanently if the search cannot be completed within a reasonable timeframe.
According to Peter Felix, president of the Association of Executive Search Consultants, whenever a CEO search takes that long, it either means the board searches for someone who cannot be easily found or the company is having severe problems making the position undesirable, reported the Washington journal.