 |
|
|
|
|
2002 Annual Shareholders' Meeting
W.J. Sanders III, AMD Chairman and CEO
What’s Past is Prologue
For the worldwide semiconductor industry, the year 2001 gave new meaning to the word cyclicality!
After soaring to more than $204 billion in 2000, worldwide sales of semiconductors fell precipitously to $139 billion in 2001.
It was a very difficult year for chipmakers as a 32 percent decline in industry sales -- the sharpest drop in the history of the microchip industry -- erased all of the gains of the prior year, and more. Spending on information technology declined dramatically as the worldwide economy, already in a lethargic state, was further weakened by uncertainty in the wake of the horrific events of September 11. The “dot-com” implosion had a devastating effect on the communications and networking sector, which had experienced extraordinary growth the prior year.
[Slide #1: AMD sales vs. industry sales, 1992-2001]
In this grim landscape, AMD continued to outperform the worldwide semiconductor industry. With total sales of $3.9 billion in 2001, our annual sales declined by 16 percent year-to-year – half of the rate of decline of the worldwide microchip industry.
[Slide #2: Top 15 semiconductor manufacturers]
Our performance propelled AMD to a tie for eleventh place in the ranks of worldwide semiconductor manufacturers – the highest ranking in our history to date!
[Slide #3: AMD sales vs. industry sales, 1975-2001]
For more than a quarter of century – for as long as the Semiconductor Industry Association has compiled worldwide sales data – AMD has outpaced the industry in revenue growth by more than 50 percent. During this period, AMD sales grew at a compound annual rate of 21.2 percent compared to 13.8 percent for the worldwide semiconductor industry.
[Slide: AMD logo]
AMD sales declined in the just-completed quarter, reflecting normal seasonal patterns coupled with continuing weakness in demand for flash memory in the communications and networking sector, which remained moribund.
[Slide #4: AMD cumulative PC processor unit sales]
In the first quarter of 2002, AMD shipped its 200 millionth PC processor!
Continued strong demand for the AMD Athlon XP processor enabled us to achieve record unit sales of 8 million PC processors in the just completed quarter. Total PC processor revenues increased by 3 percent to $684 million from the $661 million reported for the comparable quarter last year. PC processor revenues declined slightly from the record level reported for the strong holiday selling season in the fourth quarter of 2001. We believe we gained unit market share in PC processors during the quarter.
We began shipping AMD Athlon XP processors produced on 130-nanometer technology for notebooks during the first quarter, and will commence shipping 130-nanometer desktop versions during the current quarter. Our conversion to 130-nanometer technology is proceeding extremely well, and we now expect to complete the transition to 130-nanometer technology ahead of our previously announced timetable.
[Slide #5: Industry flash memory sales, 1997-2001]
Memory product sales, principally flash memory devices, in the first quarter of 2002 declined by more than 60 percent from the like period of 2001.
The industry “flash crash” of 2001 was by far the dominant factor in our year-on-year decline in total sales. The market sectors where AMD was strongest – communications and networking – were hardest hit by the dot-com and telecom collapse. Recovery in the overall communications sector is lagging the general economy.
[Slide: AMD logo]
While sales of flash memory products continued to decline in the first quarter of 2002, we believe the flash crash is now behind us. For the second consecutive quarter, our total unit and bit shipments increased. Demand was strongest for our higher-density devices, driven by a nascent recovery in the cellular phone market, where we are gaining market share at the market leaders.
This quarter we have begun sampling flash memory products based on our proprietary MirrorBit architecture, and we expect that these products will contribute substantially to revenues in the second half of this year. MirrorBit technology is a breakthrough in flash memory architecture that enables each memory cell to store twice as much data with no compromise in performance or reliability.
[Slide #6 – MirrorBit cost advantage]
The flash memory market is extremely competitive, with multiple producers. As in all markets with multiple producers, there is an overwhelming competitive advantage to being the low-cost supplier. We believe that our MirrorBit architecture will give AMD a significant cost advantage over the competition at comparable technology nodes. Our MirrorBit architecture has strengthened our position with leading customers.
We are presently converting Fab 25 in Austin to produce leading-edge flash memory products, and we are redoubling our technology development efforts to achieve industry leadership in cost and die size. We are also pursuing an aggressive technology roadmap in our Spansion JV3 fab in Aizu-Wakamatsu, Japan.
The flash memory market will continue to be a major growth opportunity. Flash memory is a ubiquitous enabling technology for a very broad range of electronic products. We are confident that demand for flash memory will recover and return to strong growth patterns. We are well positioned to grow sales sharply when demand recovers.
The Internet continues to be the primary driving force of new developments in computation and communications. While we have heretofore focused all of our efforts on improving our position in PC processors for notebook, desktop, and server applications, we did not have products to exploit the newest track for Internet access – personal connectivity devices such as personal digital assistants, web tablets, and portable and wired Internet access devices and gateways.
[Slide #7: Headline on Alchemy acquisition]
Our recent acquisition of Alchemy Semiconductor earlier this year will enable AMD to address this high-growth and promising market segment.
With the Alchemy acquisition, AMD gained an extremely talented design team, including several of the true pioneers in high-performance, low-power processor design. Alchemy system-on-a-chip products will contribute to AMD revenues by year-end.
[Slide: AMD logo]
Despite difficult industry conditions in 2001 -- absent recovery in demand from our customers in the networking and communications sector and facing sluggish demand in the PC market -- AMD has nevertheless strengthened its position in terms of product portfolio, process technology, and production capacity. R&D spending, at $170 million in the first quarter of 2002, was at record levels.
We have built and maintained strong relationships with key customers, entered into strategic alliances in pursuit of our “virtual gorilla” strategy, and continued to focus on innovation as the cornerstone of our competitive strategy.
I am both confident in and excited about the future of AMD!
[Slide #8 – Key microprocessor families]
This graphic depicts Gartner-Dataquest’s breakdown of semiconductor sales by electronic market segment for 2001 and their five-year outlook through 2006. Please note first that Data Processing Electronics, i.e., computing, was by far the largest market segment last year at $59 billion and is expected to experience the largest absolute growth of $36 billion over the next five years to $95 billion, or 37 percent of the total. This is AMD’s primary market of opportunity. Please also note that key microprocessor families currently in that segment include IA-32 architecture-based products, that is, X86 processors, e.g. Intel’s Pentiums, where we compete directly and successfully with our Athlon product line. We currently offer no alternative to Intel’s Itanium. As you will soon see, this is about to change.
[Slide: AMD logo]
As William Shakespeare wrote in The Tempest, “What’s past is prologue.” I believe it will be helpful to look briefly at where we have been and the challenges we have overcome in order to understand the journey ahead.
The “modern history” of AMD began in March of 1994, when AMD prevailed in critical litigation with Intel Corporation concerning intellectual property rights essential to our participation in the x86 PC processor market. With our victory in that litigation and the subsequent signing of a new, long-term intellectual property cross-license agreement, AMD gained a measure of control over its own destiny.
The first product of our independent PC processor development efforts was the AMD-K5 processor, introduced in the first quarter of 1996. The introduction of the AMD-K5 processor was a watershed event for AMD: it demonstrated that we had the capability of designing, developing and producing an independently engineered x86 instruction-set PC processor fully compatible with the industry-standard Microsoft Windows operating system and all other widely used operating systems. In short, it signaled that AMD could offer an alternative to an Intel monopoly in the PC processor arena.
Even more important, by pursuing an independent processor development path, we were free to innovate within the Microsoft Windows standard without waiting for Intel to define the PC processor. AMD at last was able to shed the “clone” label.
Concurrent with these developments, other monumental changes were forever altering the terrain on which we would compete.
In 1995 Microsoft introduced its new Windows 95 operating system, bringing point-and-click computing to mainstream desktop computing. This development coincided with the “overnight” ascendancy of the Internet and the worldwide web.
Together, these developments literally brought a new world of information and entertainment to the fingertips of millions of computer users around the world. They also fueled a demand for greatly enhanced multimedia capability as a standard feature of a personal computer.
The advent of visual computing offered AMD an opportunity to gain a competitive advantage by providing innovative, differentiated features in our PC processors. Our proprietary 3DNow! Technology, with instruction-set extensions fully supported by Microsoft, provided a greatly enhanced visual computing experience, differentiated our processors, and helped AMD solidify its position in this new competitive environment.
[Slide #9: AMD Worldwide PC processor market share, 1997-2001]
We began the new era of AMD facing an uphill struggle against a much larger, well-entrenched adversary. Our PC processor unit market share in 1997 was less than 10 percent. In 2001, our market share surpassed 20 percent. During the year, we gained four points of unit market share in PC processors despite the first year-to-year decline in total PC units sold since the mid-1980s.
There is a simple explanation for our success: we deliver products that enable our customers to gain a competitive advantage in the marketplace. We live by a belief that “Our customers’ success is our success.”
[Slide #10: Upside cover]
The arena in which we compete is littered with the bodies of those who have attempted to challenge Intel’s monopoly and failed.
Upside magazine recently reported, “… AMD is the last remaining microprocessor manufacturer among the 15 companies that Intel licensed as second sources for the Intel 8080 at the beginning of the PC boom in the 1980s and the only real challenger to Intel’s total domination of the microprocessor industry.”
The cornerstone of our competitive strategy at AMD is, and has always been, innovation – delivering the technology and solutions that work best for our customers.
What makes AMD different?
What makes AMD better?
Why is AMD the last man standing?
Not long ago, while reading The Oxford History of the World, I was struck by the authors’ analysis of what drives human progress and the parallels with our competitive strategy. It is not “the possession of certain faculties or physical characteristics” of our species that has enabled human progress, say the authors. Rather, it is the “remarkably intense level of activity and creativity” and the “cumulative capacity to create change.”
Thirty-three years ago, when AMD was founded, I wanted above all to build a company that would be relevant. I was not interested in simply creating a start-up company that would burst on the scene with merely a new product or a new technology only to disappear or be acquired by a larger competitor. My dream was to build a culture that would spawn innovations that could change the face of the industry.
[Slide 11: virtual gorilla]
The Oxford History authors captured the essence of our “virtual gorilla” strategy when they wrote: “The rapidity with which humanity has achieved so much can be accounted for quite simply: there are many more of us upon whose talents humanity can draw and more important still, human achievements are essentially cumulative.”
At AMD we have sought to become a nucleating force for change based on competition and choice. We recognized early on that we could not change the world by ourselves. We understood that progress is cumulative, and it is the combination of great ideas from many sources that drives advances in the human experience.
AMD has built a culture that is cumulative and progressive – an environment where innovation and change can thrive. Equally important, we have recognized the importance of working with the cumulative and aggregate intellectual and creative capabilities of an industry as opposed to a closed system of proprietary and exclusionary products and practices.
We believe in partnerships, and AMD is a great partner.
Through the years, we have engaged with premier information technology companies, including Siemens, Sony, Hewlett Packard, Fujitsu, Motorola, and most recently, UMC in successful collaborative agreements covering product and technology development and high-volume manufacturing. In addition, we have dozens of infrastructure partners with whom we are working to deliver innovative solutions that our customers need and want.
[Slide #12: AMD U.S. patents issued, 1998-2001]
Patents issued by the United States Patent and Trademark Office are a widely accepted metric of both innovation and productivity in research and development. By this metric, AMD is an extraordinary success story.
In 1998, AMD received 560 new U.S. patents and ranked for the first time in the top 25 in the world in U.S. patents issued. We were justly proud of this achievement, especially in view of the fact that our size and our total R&D spending clearly did not rank AMD among the top 25 companies in the world. Since then, we have steadily added to our intellectual property portfolio. In 1999, we received 825 new patents, and in 2000, for the first time ever, we received more than 1,000 new patents. In 2001, AMD received 1,090 new U.S. patents, which placed us at number 14 in the world in new patents issued.
[Slide #13: AMD processor innovations – series of bullets]
AMD is a leader in implementing new technologies in order to gain a competitive advantage.
- Our AMD-K5 processor was the first superscalar RISC (reduced instruction-set computing) implementation of the X86 instruction set. RISC technology enabled AMD to create a much more efficient PC processor architecture that delivers superior applications performance by executing more instructions per clock cycle than other architectures.
- We were the first to implement “flip-chip” technology in a PC processor, substantially improving the use of silicon real estate with attendant performance enhancement.
- We were the first to incorporate on-chip L2 cache memory, an important feature in improving processor performance.
- Our 3DNow! technology, the first independent enhancement of the x86 instruction set supported by Microsoft, delivered major improvements in the visual computing platform and provided a clear demonstration of our ability to offer a superior, differentiated solution.
- AMD was the first to implement copper interconnect technology in a PC processor. Every processor produced in Fab 30 since start-up has featured copper interconnects, which enhanced the performance and reduced the cost. Our experience with copper interconnects has given us a generation head start in the use of a technology essential in the production of future technology nodes.
- Our PowerNow! technology is the first dynamic power management system for a PC processor. This technology scales the PC processor performance to the requirements of the application in use at any given time, providing longer battery life in notebook applications.
- Our HyperTransport technology is a new high-speed bus protocol and physical interface for point-to-point data transfer, with transfer rates up to 12.8 gigabytes per second. HyperTransport technology has been widely adopted by leading equipment manufacturers and is an important feature of future AMD products.
- AMD is leading the way in the implementation of SOI (silicon-on-insulator) technology in our eighth-generation Hammer family of PC processors. SOI technology delivers important advantages in both performance and power dissipation.
Our early adoption of new ideas and advanced technologies has placed AMD in a position of pre-eminence.
[Slide #14: The incredible shrinking die]
This graphic shows the significant benefits we derive from our innovative design and architecture. Note that we have a substantial advantage in die size compared to Intel at every technology node.
A small die size is important in several ways.
First and most obvious, a smaller die uses silicon wafer area more efficiently by providing more gross die per wafer, and secondarily, more effectively by significantly improving yields. This phenomenon reduces the number of wafer starts required to meet a given unit volume. This is a compelling cost benefit because it reduces the total silicon area required and thereby reduces dramatically the capital costs for physical facility.
Of equal importance in reducing costs is driving to ever-finer feature sizes by rapid migration to new technology nodes.
[Slide #15: Fab 30 Technology Migration]
This graphic shows the effects of our aggressive technology migration plans. AMD has consistently demonstrated an ability to execute a very rapid transition between process technology generations.
As I noted earlier, we have accelerated our timetable for the transition in Fab 30 to leading edge technology. By the end of this year, 100 percent of the PC processors shipped from Fab 30 will be manufactured on 130-nanometer technology and we expect to be sampling 90 nanometer devices.
Typically, we complete the transition three quarters after our initial production starts on a new technology. The rapid transition to more advanced technologies has the effect of dramatically increasing production capacity in an existing wafer fabrication facility. For example, the migration from our current 180-nanometer production to 90-nanometer production will effectively more than quadruple our production capacity in Fab 30 by the end of 2004.
[Slide #16: Fab 30]
Our collaboration with the State of Saxony and the Federal Republic of Germany enabled AMD to build and have 100 percent ownership of the most advanced processor fab in the world with significantly less equity investment than otherwise required. Fab 30 today is an unqualified success. The Dresden facility is running near peak capacity in terms of wafer starts with industry-leading performance metrics. I believe Fab 30 has been the most successful new wafer fabrication facility in the history of the semiconductor industry.
[Slide #17: AMD and UMC logos]
In January, we announced another major innovative initiative to expand our wafer production capacity and sustain our technological leadership in a manner designed to optimize return on capital. We entered into a comprehensive alliance with UMC, one of the world’s premier semiconductor manufacturing companies. There are three major elements to the AMD-UMC alliance.
First, we have concluded a foundry arrangement on favorable terms under which UMC will produce PC processors to augment our Fab 30 production capacity for devices produced on 130-nanometer and 90-nanometer geometry technologies.
Second, AMD and UMC will collaborate on advanced process technology development for logic products, beginning with the 65-nanometer node.
Third, the two companies have established a 50-50 joint venture that will own and operate a 300-mm wafer fabrication facility planned to begin production in 2005 on that 65-nanometer technology. AMD will benefit from UMC’s considerable experience as a leader in manufacturing on 300-mm wafers.
The net effect of our design and architectural leadership and resultant small die sizes, rapid adoption of advanced process technology, and innovative manufacturing strategy is that we can achieve substantial growth in unit output with measured growth in physical facilities and substantial capital savings. AMD will not need to expend billions and billions of dollars on incremental production facilities to achieve our market share objectives. As a result, we expect to have an industry-leading cost structure for AMD PC processors.
[Slide #18 – Product lineup]
AMD’s paradigm for small die sizes and rapid technology migration is applicable across the entire breadth of computing, enabling AMD’s outstanding products to mount a challenge for leadership in computation and information and media access.
[Slide #19 – modified Slide #8]
Here is today’s line-up. But what about the future? This modified version of the Gartner-Dataquest graphic shown earlier depicts the market opportunities in computing for our eighth-generation processor family, code-named “Hammer.”
What portion of the computing opportunity is available to Hammer? In short, all of it! The future of computing is Hammer. The desktop and notebook versions of the family, internally codenamed “Clawhammer,” will build on the established brand equity of the AMD Athlon name.
Production shipments of AMD’s eighth-generation AMD Athlon processor for desktops, as previously noted, are slated to begin by the end of the year. They incorporate AMD’s x86-64 technology, while running 32-bit Windows applications faster than any processor on the planet. They will be 64-bit capable, enabling Windows users to migrate seamlessly from 32-bit to 64-bit computing in the same system while maximizing performance for both.
The enterprise class server and workstation version of the family, internally codenamed “Sledgehammer,” will be marketed under the brand name Opteron.
[Slide #20 – Opteron slide]
Sledgehammer is defined as a simple powerful tool. Opteron means simply the best. It’s what information technology customers in the enterprise want. It preserves their enormous investment in software. It puts them in control of the timing of their migration to 64-bit computing.
[Slide #21 – Opteron features]
AMD Opteron processors incorporate AMD x86-64 technology and feature three HyperTransport interconnects which can provide up to 19.2 gigabytes per second aggregate bandwidth for server systems. The AMD Opteron processor will compete effectively against high-end Xeon and Itanium processors from Intel.
Yesterday we announced with Microsoft their endorsement of the AMD x86-64 technology and their agreement to develop a 64-bit version of Windows for AMD’s eighth-generation Athlon and Opteron processors. This development has been underway for several years and is the result of the efforts of hundreds of Microsoft and AMD employees.
Microsoft’s Dave Cutler, the father of Windows NT, says it best: “Over the past couple years, I’ve been working with AMD on their next-generation K8 processor. What’s really exciting about the K8 is that it has both 32-bit and 64-bit capabilities. Furthermore the 64-bit systems will run the existing 32-bit applications. So that will protect customers’ investments in software and hardware. Currently we have the 32-bit Windows XP and Windows 2000 Server systems running on the K8 first silicon, which has proven to be very stable.
[Slide #22 – Cutler quote]
“We also have developmental 64-bit version (s) of Windows XP and Windows.NET Server running on the very same hardware system. I’m really excited about this chip.”
So am I. Indeed, here today we have several demos including our Melody dual Opteron processor system, i.e., Sledgehammer, running 64-bit Windows advanced servers operating system along with 64-bit and 32-bit applications.
[Slide #23 – Sledgehammer demo]
In this networked system, the client is running on an eighth- generation Athlon processor (i.e., Clawhammer) under the standard off-the-shelf 32-bit Windows XP professional OS.
[Slide #24 – Clawhammer demo]
Competition is good, and AMD is good competition. X86-64 technology is the best solution for increased computing power for the information technology industry going forward. This approach provides investment protection, allowing users to continue to use their existing 32-bit applications in which they have already invested, and gradually transition software to 64-bits as needed. All Windows users can benefit from 64-bit technology. Again, Microsoft’s Cutler says it best. “Over the last ten years, the applications we’ve put on PCs have grown, they’ve grown in size and computational demands, and 32 bits of address space just isn’t enough anymore. The size of databases has grown to the point where we just can’t get the performance out of the 32-bit address space that we need to get to continue to support these applications. Over the past few years we’ve added a few features to extend the life of the 32-bit system, but it’s not enough and we need to move to 64 bits to continue to support these large databases and high end desktop applications.” AMD will lead the migration to 64-bit computing. AMD will be first with 64 bits to the desktop.
With the advent of this eighth-generation family of processors, with AMD’s X86-64 technology fully supported by Microsoft, we can create a fundamental change in the market. We can change the face of the industry. We can realize the AMD dream.
[Slide: AMD logo]
AMD has always been a team effort. For 33 years I have been honored to be the leader of that team. Today I turn over that leadership to Hector Ruiz. I am happy to do so in that Hector and I have a common philosophy; we play to win. Defeat is not an option. We have a common vision for AMD’s future: to be the number one supplier of flash memory through a reinvigorated joint venture with our good partner Fujitsu and to be the number one supplier of microprocessors for computation and information and media access through architectural and design leadership. We have set an ambitious goal to achieve these objectives by the end of this decade. Taking AMD to this next level will continue to be a team effort. Hector has a very strong supporting team, with just the right mix of seasoned experience and youthful energy in every key position.
Here are some of them.
[Slide #25 – AMD Leadership Team]
Rob Herb, 40, is executive vice president and chief sales and marketing officer.
Dirk Meyer, 40 is group vice president of our Computation products Group.
Bertrand Cambou, 46 is group vice president of our Memory Group.
Bob Rivet, 48, is senior vice president and chief financial officer.
Tom McCoy, 51, is senior vice president and general counsel.
A very strong and deep bench backs up these senior executives.
[Slide – AMD logo]
As I step aside as CEO after 33 years and fashion a new role as Chairman of AMD, I view my transition as a passage, not journey’s end. The adventure continues.
AMD has a strong balance sheet with nearly $1.3 billion in cash. We generated a positive cash flow from operations of $160 million in the just-completed quarter. In the face of severely depressed revenues resulting from the worst industry downturn in history, we are flirting with profitability while expending record sums on R&D. We generated $160 million in positive cash flow in the just-completed quarter. I believe as the industry recovers, Hector will lead this team to produce truly astonishing results. The AMD dream will become reality.
Thank you and God bless.
|
|
|
|
|
