At the recent Hot Chips Conference, a symposium for high performance CPU designs, Advanced Micro Devices (AMD) finally showed its teeth by disclosing details of its new CPU micro-architectures: for the performance space (notebook, desktop, server) the company discussed its upcoming "Steamroller" micro-architecture and for the low power (notebook, netbook, tablet) segments, its "Jaguar" micro-architecture. These positive developments, coupled with an already diminished valuation, lead me to believe that AMD has upside to $8/share in the medium term.
"Steamroller" Targets Notebooks
It is no secret that PC industry is shifting to "thin and light" form factors with a strong focus on battery life. AMD's upcoming "Steamroller" architecture will bring significant performance per watt improvements to AMD's "Bulldozer" and "Piledriver" cores through a number of innovations detailed here. These improvements, coupled with a move to a 28nm manufacturing process, should allow AMD to continue the momentum into the "thin and light" space that was started with the firm's "Piledriver" based "Trinity" APU.
With a more competitive part on the mobile side than its current offerings, coupled with the firm's history of aggressive pricing, AMD could very well be in a position to take share from rival Intel (INTC) in the notebook space.
In addition, the "Steamroller" core will be more efficient, so AMD will be able to produce more power efficient and higher performing "Opteron" server CPUs to compete with Intel's "Xeon" lineup. AMD is unlikely to take the raw performance or performance per watt crown from Intel on the CPU side of things. However, improved efficiency on the CPU side and continued strength on the integrated GPU front (heterogenous computing could make a big splash in the server/HPC market), will make AMD a formidable competitor in this segment, especially as AMD has no problems competing on price.
"Jaguar": Netbooks, Notebooks and Microservers
Quite possibly the more exciting development is the upcoming "Jaguar" core, whose micro-architecture was also revealed at the conference. The chip will pack four cores, be built on a 28nm process, and have many features that will improve battery life significantly, making SoCs aimed at tablets from this micro-architecture viable, opening up a new revenue stream for AMD.
In the netbook and low end notebook space, such a chip coupled with AMD's very competitive integrated graphics should allow for this offering to be very attractive to OEMs, especially as they aim to find ways to improve their gross margins while still offering products that perform to specification.
Finally, AMD's purchase of SeaMicro could prove to be particularly smart, as this gives AMD a perfect venue with which to achieve market penetration in the "microserver" arena. With high performance, low power x86 cores, AMD's products here could prove effective at staving off the micro-server threats from Intel and ARM Holdings (ARMH)
Stock Has Significant Upside
While the strong product pipeline is nice, it merely serves to highlight the absurdity of AMD's current valuation. The company's shares currently trade at 0.43x sales against the median semiconductor P/S of 0.96 and mean of 4.99. To get an idea of how AMD stacks up against its direct competitors, check out this chart:
NVDA Price/Sales Ratio data by YCharts
Intel and Nvidia (NVDA), AMD's two major competitors, trade for 5x AMD's P/S. While Intel is dominant in the x86 space, it does not have the discrete graphics business that AMD has. And while Nvidia is dominant in the discrete graphics space (although by a much smaller margin than Intel is over AMD in the x86 space), it lacks an x86 license, which gives AMD a much larger TAM for their CPUs that includes notebooks, desktops and the data center. While AMD is notably behind Intel in x86 CPU performance, the gap will close with AMD's next generation architectures, and in the mean time can still compete on price in an already high margin arena.
Speaking of margins, it's worth examining AMD's gross margin profile compared to the two aforementioned competitors to see if perhaps low margins could be causing the relatively low price-to-sales valuation:
AMD Gross Profit Margin data by YCharts
While AMD's gross margins trail Nvidia's and Intel's, it does not seem that this rationally accounts for the large price to sales disparities.
While AMD's currently experiencing macro headwinds from diminished PC demand as well as competitive pressure from Intel and Nvidia in their CPU and discrete GPU segments, respectively, I do not believe that the only other major x86 player as well as the only other major high performance GPU player should be trading for less than 1/2 sales and for only 8.81x estimated future earnings.
Even priced at a conservative 1x sales, this would represent upside to $8/share, which I would say is a reasonable target as AMD demonstrates over the following quarters that it has fixed the issues that were unearthed at the most recent quarterly report.
AMD's graphics IP and high performance CPU design skills are also valuable and are likely not baked properly into the price. At such low valuations, AMD is a very clear potential takeover target for Qualcomm (QCOM), Nvidia, and even Apple (AAPL), all of whom are attempting to build higher performing, more power efficient SoCs for use in more than just tablets and smartphones. In fact, Apple, a well-known patent troll, could very likely attempt to acquire AMD for the rights to a whole slew of patents relating to high performance CPU and graphics processor design.
The Inevitable Short Squeeze
The short interest on AMD is actually a fairly staggering 14%. Interestingly enough, even though AMD is still profitable and strong enough sales to support a much higher market capitalization, short sellers seem to be standing firm in their positions, likely betting on bankruptcy. This doesn't seem to be a particularly safe bet (in fact, it seems downright reckless) at these levels or given the company's current state. AMD was both profitable and cash flow positive in the most recent quarter, despite seeing soft sales (note that until that quarter, AMD had been very consistently beating estimates).
The downside is now fairly limited and the upside potential is staggeringly huge. Any number of the following catalysts could send the stock soaring as shorts attempt to cover:
- An in-line or better earnings report [since the Dell (DELL) and HP (HPQ) reports have now convinced Wall Street that both Intel and AMD will warn/miss]
- More competitive products in the mobile, desktop, and server markets utilizing the firm's updated "Piledriver" core. Success here will allow AMD room to increase ASPs (improving gross margins and the bottom line) as well as take share from Intel
- An acquisition (although I find this scenario unlikely, the mere hint of one could cause a significant move in the stock price as it did on August 8)
- Design wins in upcoming Windows 8 tablets
- Further penetration into "ultra-thin" notebooks
- Increased discrete graphics market share
Another point to consider is that that even though the short interest represents a fairly large bet on bankruptcy, the likelihood of this is very low. Over the last five years, AMD has managed to cut its debt load by 62%:
AMD Long Term Debt data by YCharts
While cutting the debt, AMD has also been able to, through consistent profitability, keep its cash and short term investments intact (AMD has nearly enough to wipe away all of its debt):
AMD Cash and ST Investments data by YCharts
It doesn't seem that AMD is in particular danger of insolvency, especially as the company continues to be profitable, isn't burning through cash, and is making significant strides in lowering its long term debt over time.
The Bearish Case
While AMD's technology is strong and its valuation attractive, there are a number of issues that significantly weigh down the stock's price.
First, the company recently issued $500M worth of senior notes due in 2022, with the possible uses of the proceeds of the issuance as follows:
- The repayment or repurchase of some or all of its outstanding 5.75% Convertible Senior Notes due in 2012
- The repayment or repurchase of some or all of its outstanding 6.00% Convertible Senior Notes due in 2015
- Cash payments to Global Foundries related to the 28nm product limited waiver of exclusivity
- Potential strategic transactions
To me, this seems perfectly reasonable: take advantage of low interest rates today to borrow to pay off debt instead of draining cash reserves. However, the short sellers likely saw the "need" to do this at all as a sign of weakness, fueling their positions.
Further, AMD admitted that last quarter's execution windfall on the desktop channel front was due primarily to their own mismanagement. Shorting a stock because management seems to be incapable is not unreasonable, but I contend that a single poorly executed quarter isn't necessarily the sign of death. In this case, desktops anywhere near as important to either CPU vendor on a long term basis as mobile is, so AMD's continued strength in the mobile segment with "Llano" and "Trinity" is encouraging.
Additionally, competitive pressure from Intel on both the micro-architecture front as well as the process technology front has caused a decrease in the ASPs of AMD's high end desktop and server CPUs. However, it seems that AMD has given up trying to fight this losing battle and is instead looking for profitable niches (hence the big focus on general purpose GPU/laptops) where their products will be able to provide meaningful price/performance advantages over Intel's products while keeping ARM chips from consuming their business in low end notebooks and netbooks.
Finally, while not a huge loss in terms of revenue, AMD lost the discrete graphics spot in the higher end MacBook Pro computers to competitor Nvidia. However, as gamers (the people who need discrete GPUs) will overwhelmingly choose Windows PCs, I do not see this as particularly material to AMD's long term future.
Hating the PC is rather in vogue right now, especially as the media attempts to portray the PC as "dying." While tablets are interesting supplementary media consumption devices, there is still room for notebook and desktop growth, albeit at a slower pace than previously seen.
Make no mistake: AMD is a much riskier play than Intel in the PC space. No dividend, much lower PC market share, and a less favorable debt/equity ratio make AMD a fundamentally weaker company. However, in terms of valuation, AMD is significantly more attractive than Intel (which is already incredibly attractively valued), and has products (HPC accelerators, discrete GPUs) that even Intel doesn't yet have.
For the investor with high risk tolerance, who believes that AMD can overcome its execution difficulties, and who believes that the PC market isn't going to go "poof" in the long term, AMD represents a solid bargain at these levels.