In firing AMD’s CEO Dirk Meyer, the company begs the question “what can be done better?” There are only so many computing technology places that AMD can go, let alone do well. Let’s explore those and see how AMD’s post-Meyer strategy may change, and create “significant growth, establish market leadership and generate superior financial returns”.
Setting the Playing Field
AMD’s January 10th press release quotes chairman Bruce Claflin as saying, “Dirk became CEO during difficult times. He successfully stabilized AMD while simultaneously concluding strategic initiatives including the launch of GLOBALFOUNDRIES, the successful settlement of our litigation with Intel and delivering Fusion APUs to the market.”
“However, the Board believes we have the opportunity to create increased shareholder value over time. This will require the company to have significant growth, establish market leadership and generate superior financial returns. We believe a change in leadership at this time will accelerate the company’s ability to accomplish these objectives.”
On January 24, AMD reported fourth quarter revenue of $1.65 billion was up 1.9% sequentially and flat year over year.
Computing Solutions was 74% of sales, down 0.6% sequentially and 0.1% from the year-ago quarter. Weaker demand in mobile, share losses in server and a steadier desktop business combined to generate these results. Average selling prices (ASPs) were down.
Graphics was 26% of sales, an increase of 8.7% sequentially and 0.7% from last year. The graphics business was up due to a double-digit volume increase, especially AMD’s Radeon HD 6800 and 6900 series chips.
The pro forma gross margin was 45.1%, down 62 basis points (bps) from the previous quarter and up 381 bps from the year-ago period. The sequential decline was on account of weaker pricing, lower utilization rates and market share losses in the high-margin server business.
Operating expenses of $602 million were up 1.2% sequentially. The operating margin dropped 35 bps sequentially to 8.6%. SG&A increased and R&D declined as a percentage of sales for the second straight quarter, which in combination with the lower gross margin resulted in the decline.
Multiplying gross margin percent times sales, AMD has about $740 million a quarter available to run the business, including new-market R&D and marketing.
Handicapping the AMD 2011 Line-up
In processors, 2010 was not a great year. The company is betting 2011 will be a better year. AMD is just moving to 32 nm process technology, a year behind arch-rival Intel. AMD is absolutely counting on a smooth transition to 32 nm process at GlobalFoundry in order for its new Fusion micro-architecture to be competitive in the netbook, laptop, desktop, and server markets.
AMD’s leap in 2011 is not without risk. In effect, AMD is bringing new designs to a new manufacturing process. In Intel’s terminology, this would be tick-tock in one iteration; instead of getting a new process well in hand (e.g., tick), followed by the new micro-architecture (e.g., tock). Lots can go wrong in a tick-tock, and when things go wrong, time-to-market is lost and the current product line has to stay in the market, losing competitive advantage monthly. The obvious solution for maintaining a less competitive position is to cut prices, and that is exactly what AMD has had to do in recent quarters.
Choosing the Strategy Ingredients
Here are my observations on AMD’s present markets along with how I see those markets accepting a greater AMD presence.
AMD has less than a 10% market share in highly profitable server-class microprocessors. The Bulldozer micro-architecture anticipated in 2H-2011 was originally expected in late 2009 on 45 nm. That program was scrapped, and Bulldozer is now on a 32 nm process road map. Without a new architecture, AMD’s Opteron server chips had tepid uptake by the server OEMs such as Dell, HP, and IBM in 2010.
Meanwhile, rival Intel’s 32 nm Westmere architecture in Xeon processors launched in 2010 have the best price-performance and performance-per-watt Intel has ever offered.
Because server chips provide hundreds of dollars in gross margin each, the highest margins in the microprocessor universe, AMD is compelled to compete aggressively for server market share; that’s where the money is.
The Bulldozer-based launch will likely coincide with Intel’s Sandy Bridge-EX release, probably this summer or early fall. In Olympic terms, AMD needs as close to a perfect 10.0 as possible for two reasons. First, it needs to recover customer mind-share and market-share lost while waiting for Bulldozer, and secondly, the company has a one-year window before Intel rolls out Ivy Bridge server processors at 22 nm in 2H-2012 that will likely use less energy to do more computing.
The 32 nm Brazos platform was rolled out this month in the form of the codename Zacate as competition for Intel’s Atom and various ARM-based competitors. It is the first of AMD’s Fusion processors that include a microprocessor and graphics processor on one silicon die.
Reviews of the Zacate-based machines put it in the netbook/nettop category: competitive with Intel’s current Pine Trail 45 nm versions of Atom, but not better than Intel’s 2010 Arrandale and Clarkdale mainstream processors.
I rate AMD as competitive with Intel and ARM in netbooks. Later this year, Intel rolls out its 32 nm Medfield version of Atom. The two companies will compete on price and performance.
Unfortunately for AMD, the netbook market momentum slowed dramatically in 2010 and into 2011 due to the tablet form-factor, and especially Apple’s iPad.
I don’t see any way for any vendor to halt the demand for tablets over netbooks. That makes AMD’s competitive position in netbooks worth relatively little as a growth driver, let alone contributing to superior financial returns. There are modest opportunities for netbooks in vertical markets such as education, but on the whole, netbooks are not going to make a huge difference to AMD’s performance.
Tablets and Smartphones
Tablets and smartphones are the siren song of 2011 growth. Neither AMD nor Intel are well equipped with X86 processors in 2011 to take on the entrenched ARM camp of Apple, Qualcomm, TI, and other ARM licensees. Yet the first street comments after Meyer’s departure were “he screwed up by not getting into the tablet and smartphone market.” Really? The smartphone/tablet market requires enormous investments over years to enter, including creating a software development ecosystem.
I’d like to hear the arguments for how AMD can make the smartphone/tablet investments without starving other current businesses, risking the whole. For the sake of that argument, let’s look at the options.
There are three ways for AMD to compete in the smartphone and tablet market for processors:
- Create a winning X86 microprocessor design with onboard graphics. This is the brass ring. Cost of entry is huge and the clock is ticking. As a soft fab, AMD gets margins no better than other soft fabs like Apple. In short, while the unit potential is huge, AMD is too small a piece of the puzzle to drive success for X86 in this market.
- Create a winning ARM microprocessor design with onboard graphics. This direction requires the greatest new R&D effort by AMD because it would be new to the ARM world. In the end, AMD would be competing not only with Apple’s designs, but also new competitors Microsoft and nVidia. Lots of ways to fail at great cost with this option, which means risks are high.
- License AMD’s graphics technology to ARM licensees. Rather than option 2, this option says “make some money by arming the ARM camp with our excellent graphics technology”. Requires some R&D, but lets the licensee duke it out for silicon contracts. Also, since nVidia will soon be out there with its graphics-rich design, AMD is a natural graphics competitor to nVidia to those ARM licensees who want better than average graphics, a differentiator. On balance, this option makes sense to pursue, as the graphics technology purchased in the ATI acquisition is one of AMD’s genuine jewels.
[Update Interim CEO Thomas Seifert told the Wall Street Journal February 17th that AMD would not pursue smartphones as the company lacks baseband radio technology to add to its systems on a chip. But the company will, like Intel, pursue the tablet markets in 2011 with even more competitive second-generation Fusion products in 2012. That strategy looks realistic to me. But the loss last week of both the chief strategy officer and chief operations and administrative officer are not good signs of a coordinated management team.]
Laptops and Desktops
Microprocessors for laptops and desktops are the bread and butter of AMD’s computing solutions business. The company has roughly a 20% market share overall. The landscape is one of constant trench warfare with arch-rival Intel, which has the other 80% of the PC business.
AMD expects a good 2011 based on market acceptance of its new 32 nm mainstream chips, starting with Llano in Q2. These will compete with Intel’s Sandy Bridge mainstream processors, which got thumbs-up reviews when launched earlier this month. By this time next year, Intel will have its 22 nm Ivy Bridge “tick” in the market, well ahead of AMD’s next process change. Ivy Bridge will further cut AMD’s lead in on-die graphics capabilities.
Moreover, Intel’s $15 billion in fab investments is forcing commensurate capital investments by AMD’s fab partners. This will require higher volumes by the AMD ecosystem to amortize these capital costs.
AMD has no choice but to continue to invest heavily to win in the laptop and desktop market because that’s where most of the company’s business is today. The company has done best in the past couple of years in the value and transactional PC segments. That is, the low-price PCs. As a result, AMD’s gross margins are more than 2,000 bps lower than Intel’s.
Emerging Technology Markets
If the AMD board is serious about the need to create “significant growth, establish market leadership and generate superior financial returns” — and I do not dispute that need — then emerging technology markets which may never pan out, or take years to mature, are not particularly attractive to a company that needs a fix in the next year or so.
Some of the emerging technology markets that AMD is likely to evaluate include:
- Fusion supercomputers: China’s new #1 supercomputer is a hybrid of nVidia graphics chips (with lots of scientific calculating power) managed by Intel Xeon servers. AMD can do this too.
- Microservers: the concept is to offload low-impact loads onto low-cost, low-power servers that could be densely packaged. Products are just reaching the market. IT is barely aware of the concept, let alone sold on the investment. Will compete with virtualized workloads on traditional enterprise servers.
- Cloud computing: Lots of servers share and pass around workloads. There’s some beef here, but right now it’s mostly smoke and mirrors as a server market.
- Virtual Reality: anything that requires more graphics processor is good for AMD. There are lots of reality-based activities going on (and have been for more than a decade) such as 3D. However, virtual reality is like artificial intelligence: always on the horizon.
If Dirk Meyer is sitting at home wondering exactly what he should have done differently, he has my sympathy. There are no obvious go-to strategies for AMD to take which have low risk and the high financial reward that the board says it wants:
- The under-utilized jewel of AMD is the graphics technology it acquired from ATI in 2006.
- The server market is where the gross margin dollars are. Share is worth fighting for.
- Netbooks are not a big enough market to propel AMD to the next level.
- Tablet and smartphones are a dog-eat-dog market where the big dogs are fighting for large-volume market share. AMD’s best-fit is to license its graphics technology and to stick to its X86 knitting. [Update Walking away from the smartphone market is painful but the right strategy. There will be some opportunities in tablets as that market will look far different in three years than the all-Apple conditions today.]
- Laptops and desktops are AMD’s core business. AMD must continue to innovate or be left in the dust by Intel. Increased market share in a growing global market worth 350 million processors in 2010 is AMD’s most obvious lever.
- Emerging technology markets are not attractive to a company which needs to improve its performance over the next year or so. Success starts too small and takes to long to deliver.
My unsolicited advice to AMD is to execute better, not chase after new, and potentially ephemeral, markets.
[Disclosure: The author has no financial position in AMD.]