Analysis: AMD struggles to survive

Advanced Micro Devices Inc.'s place in the annals of the semiconductor industry is now more or less fixed. The battle for dominance in the microprocessor market has been won and lost: The crown belongs to Intel.

AMD's future isn't clouded in any great mystery. The company is burdened with debt, is losing market share and appears destined to become an even smaller player in its core MPU market. Recent strategic moves at AMD are not making a dent in Intel's leadership: The acquisition of graphics IC vendor ATI Technologies hasn't lived up to expectations and becoming a fabless IC vendor isn't a game changer either.

But executives at AMD haven't yet come to terms with their second-place status. On the contrary, they continue to pursue actions they believe would level the playing field and make it possible for them to beat Intel.

No match for Intel
The bottom line is that AMD cannot beat Intel, not now and perhaps not ever. The company's top executives, including outgoing executive chairman Hector Ruiz, may believe they can continue to trade punches with Intel in this unequal slugfest but that strategy is, to put it bluntly, delusional. It has only brought misery to AMD and pushed the company into a spending contest it has no hope of winning.

While AMD continues its grandstanding, the rest of the industry can glean significant lessons from its experience. As observers ponder AMD's future, they must focus not merely on how the company arrived at its current disadvantageous position or even what it could do to reverse Intel's overwhelming edge, but on what they can learn from one of the most fascinating developments to emerge from the sector.

The AMD-Intel multi-decade saga offers a great opportunity to examine the dynamics of competition in the industry; in particular, what roles marketing, R&D, product development, manufacturing efficiencies, financial resources and managerial expertise play into how effectively companies differentiate themselves in the marketplace. What makes one company a winner and another a distant second in a market clearly large enough to sustain at least two major competitors? Are process technologies and manufacturing efficiencies huge differentiators? Do varying administrative strategies have a significant impact? Finally, what role, if any, do marketing, public relations and mindset manipulation play?

AMD has been outspent, outgunned and outmaneuvered for decades. The foundry joint venture it set up in October with Abu Dhabi-based Advanced Technology Investment Co. (ATIC) won't remove the humiliation AMD executives must feel, knowing they fired the company's last and most potent weapon and still left their competition largely unscathed.

As it turned out, Intel was more concerned about the disruptive effects of the ongoing credit crunch than about the balance sheet improvement actions of rival AMD. "Our future operations involve a number of risks and uncertainties—in particular, current economic uncertainty, including the tightening of credit markets, as well as future economic conditions," Intel noted in its Q3 regulatory filing.

Industry analysts hold a similar view of AMD's situation. "AMD's market position remains substantially unchanged," said Lucy Patricola, an analyst at Standard & Poor's. "The increased liquidity [from the JV] extends the time during which AMD must address substantial marketplace pressures, including its second-tier position behind industry leader Intel."

Other analyses confirm AMD must undertake a tough task to match Intel's spending power. Bill McClean, principal analyst at IC Insights, made it clear that AMD executives had poured scarce cash into an unwinnable war with an opponent whose resources were vastly superior. AMD, McClean noted in a report, spent $6.4 billion on capital equipment from 2004 to 2008, compared with Intel's budget of $25.6 billion during the same period. "Intel's capital expenditures over the next five years will total about $30 billion, which would be five times [AMD's] Foundry Company's most optimistic plan of investing $6 billion over that time," McClean said.

The conclusion, McClean said, is that "AMD's new strategy and business model keeps them in the MPU game—however, it will not shift the balance of power in the MPU or foundry segments of the IC industry over the next five years."

Hope in multicore programming
Yet, attempting to shift the dynamics of the MPU market in favor of AMD has been the chief pursuit of the company's executives for decades. They continued deploying resources in furtherance of this goal even as it became obvious the competition was pulling away. AMD's blind desire to beat Intel in process technology and manufacturing efficiencies pushed its competitor to regularly revamp its product line, a strategy Intel honed years ago and which its executives believe has always worked for the company.

Nobody is suggesting AMD throw in the towel and let Intel establish a monopoly in the MPU market. But recognizing the limits of its resources would go a long way toward making the company a more successful IC vendor and bolstering its competitiveness in whichever segments it chooses to participate, commented Louis Savain, an independent software developer/researcher, in an EE Times Web forum.

"Everybody should know by now that AMD's only hope of competing successfully in the processor marketplace is to come up with a new technology that solves the multicore programming problem," Savain wrote. "It must either forge a new market or fall by the wayside. Unfortunately, AMD has been playing a 'monkey-see, monkey-do' game with Intel. The stark reality is that nobody can beat Intel at its own game."

But it wasn't always Intel's game. That's why AMD's Ruiz believes AMD can still trounce Intel, if it can successfully pay down debt—through the sale of business units, if necessary—transition into a more cost-efficient fabless manufacturer and get the courts to declare that Intel violates antitrust laws, thereby forcing the market into a more competitive balance.

Reality check
Let's quickly rule out the possibility any of these approaches will dramatically alter AMD's fortune.

AMD expects the new wafer foundry JV with ATIC to help it regain lost ground but even this is highly doubtful. The JV promises to be disruptive, but mainly for AMD, because it still cannot match Intel's capital expenditure or R&D budget.

Paul Otellini, Intel president and CEO, said his company is not concerned about AMD's new fabless strategy. "From my perspective, nothing has changed. You still have someone that has to invest in capital at one end of the food chain. You have to put capex in the ground and build factories and develop technology, then optimize that to products and get a return on that capital and then build those products and now sell it to someone else for some price who then sells it to the end customer. So the food chain really hasn't changed in terms of the ecosystem, except that there's one more person perhaps looking for a return in the equation."

Moreover, giving up manufacturing might have helped temporarily secure financial stability for AMD, but it would still have to fund its share of the foundry JV or give up ownership—which would further constrain its operation, as it could no longer dictate terms of engagement to the new company.

Intel, on the other hand, considers control of the process development and manufacturing environment a major competitive advantage, according to chief financial officer Stacy Smith. "We certainly believe, for our business, the ability to integrate design and manufacturing is a core asset," Smith said.

Having failed to trump Intel in the marketplace, AMD opened another battlefront in the courts, where it accused its rival of antitrust practices. While the company may have a strong case, most legal experts say it takes years, sometimes decades, for courts to resolve business disputes. The impact on market share and the competitive environment is limited.

Never-ending rivalry
Perhaps AMD's most significant failure was in allowing Intel to break away and establish a huge revenue lead. Once established, a near-monopoly is not easily dislodged, and Intel took every step to ensure its dominance would never be challenged again.

The two companies began their fierce competition back in the early 1990s, when AMD's revenue was closer to Intel's.

In 1992, AMD recorded revenue of approximately $1.5 billion compared with $5.8 billion for Intel. The revenue gap between the two companies began widening dramatically the next year and the gulf has expanded over the past decade. Intel's revenue, for instance, doubled to $11.5 billion by 1994 while AMD's grew a respectable but still smaller 41 percent.

By the end of 2008, analysts project Intel's revenue will increase to $39.8 billion, dwarfing AMD's estimated annual sales of $6.1 billion. Based on these numbers, Intel's sales will have grown 586 percent in 16 years while AMD's increased a comparatively meager 307 percent.

The revenue gap between Intel and AMD is significant in several ways. As Intel's sales grew faster than its competitor's, so did the company's spending power and ability to pour more financial resources into critical areas like employee recruitment, R&D, capital expenditure, marketing and manufacturing process improvement. In the first six months of this year, for instance, Intel generated more than $5 billion in cash from operating activities compared with AMD's negative $226 million in the same period.

More critically, the lopsided spending patterns between the two companies fed OEM customer and consumer perceptions, creating a negative, also-ran view of AMD and a generally positive impression for Intel.

AMD's new asset-lite manufacturing strategy ensures the company lives to fight another day, but the best it can hope for is to not get knocked out. That shouldn't be too difficult; Intel needs AMD around to ward off allegations it's a monopoly.

- Bolaji Ojo
EE Times