How Intel Wrecked Itanium

Microsoft and AMD have been cozy of late. Microsoft is devoting significant resources to support AMD's x86-64 extensions. Microsoft could have chosen to ignore AMD. Apparently it cares enough to create a Windows port for unproven technology from a company with an uneven record in the marketplace.

It is understandable that Microsoft does not want Intel to be dominant in the PC processor market. Keeping competition alive in the processor market keeps processor prices low, increases PC sales, and consequently increases Microsoft revenue. The price war between Intel and AMD has greatly benefited Microsoft. During the recent tech meltdown, Intel and AMD were both in the red while Microsoft was reporting revenue growth and jacking up licensing fees.

These numbers from Intel, AMD and Microsoft income statements clearly show the trend:

Revenue in Millions of Dollars
Year AMD Intel Microsoft
2000 4,644 33,726 22,956
2001 3,891 26,539 25,296
2002 2,697 26,764 28,365

Without the price war Intel would have kept processor prices high to keep its profits margins up and preserve the pricing structure of the processor market. As a consequence, a significant decline in PC sales would have occurred. Microsoft would have seen a leveling off or a decline in its revenue. Under this scenario Microsoft had no course of action to offset the decline in revenue. A steep increase in licensing fees would have led to renewed calls for Microsoft breakup.

Increasing licensing fees is not a viable long-term option for Microsoft anyway. Every increase in software prices makes Linux more viable as an alternative. Increasing the volume of PC shipments is a better goal, as it increases revenue with fewer side effects. The price war between Intel and AMD is ensuring that processor prices are low and demand for them is high.

To keep the good times rolling Microsoft needs to ensure that both AMD and Intel permanently stay engaged in a price war. An upper hand gained by either will not be good news for Microsoft. The Itanium strategy followed by Intel is very dangerous for Microsoft. Success with Itanium will restore Intel hold over the processor market. Intel will once again be in a position to set higher processor prices.

Microsoft is forced to support Itanium as it needs a 64 bit OS that can compete in the high-end server market. To counteract the influence of Intel, Microsoft is backing AMD strongly. Microsoft's long term objective will be to have Itanium as a server platform and x86 as a consumer PC platform.

With Microsoft squarely behind AMD, it is reasonable to expect x86-64 to be a resounding success. Intel will be forced to introduce its own 64 bit extensions. Of course Microsoft will quickly oblige by supporting Intel extensions as well. Microsoft certainly doesn't want AMD to start playing the role of Intel. Itanium probably will have to settle into a niche market.

Intel did have an alternative to all of this. It could have traded its monopoly status for a duopoly status. Industries dominated by two suppliers sometimes settle into duopolies. The suppliers don't need to meet (which is highly illegal) to set prices. Bargaining to share the spoils is done using signaling. A classic case of such signaling is documented in the article "The Bidding Game."

There was a chance for the x86 processor market to settle into a duopoly at the start of the processor price war. During 2000 AMD was publicly proclaiming a target of 30 percent market share to be achieved by the end of 2001. That might have been a signal to Intel. AMD at that time had a market share of around 20 percent. A target of 50 percent increase in market share in a very short time is very ambitious. Even if this announcement was not a signal, it suggested AMD had strong technology and no problems ramping up production. AMD's Fab in Dresden was coming online and its production had to be sold somehow.

Intel should have read the signal and responded by conceding some market share to AMD. Instead, Intel unwisely and arrogantly chose to start a price war. If Intel had allowed AMD to take 25-27 percent market share, the industry might have settled into a duopoly. AMD did not want a price war and would have settled at that.

Creating a duopoly is a much tougher proposition now. By competing with AMD, Intel has destroyed the old favorable pricing structure of the processor business. Consumers have gotten used to lower prices for PCs and will respond unfavorably to price hikes. Palm and Handspring during the tech recession were faced with a similar situation but smartly avoided a price war to preserve the pricing structure of the PDA market. Intel should have realized what it was doing but it seems Intel managers are mere technologists and lack business sense.

Establishing a duopoly in the processor business might not have rescued Itanium but it would have increased its chances of success. Processor prices would have stayed high under a duopoly and Microsoft would have had less of an incentive to interfere in the processor business. Microsoft would have smelled blood but not tasted it.

More importantly, as a reward for conceding market share, Intel could have received AMD's cooperation on Itanium. Intel's smartest move before the price war would have been to license Itanium technology to AMD on terms favorable to itself. Licensing Itanium to AMD at that time would have confused Microsoft, and probably forestalled support for x86-64. Itanium has proven to be a dog and AMD is unlikely to buy into it now.

The recent AMD Fab36 announcement has taken the last breath out of Intel's strategy to muscle AMD out of the microprocessor business. Intel still has a window of opportunity to make up with AMD. Cooperation will be the key to profitability and future success for both AMD and Intel.

by Usman Latif  [Dec 01, 2003]


Microsoft & x86-64
The Bidding Game
Palm cuts prices to deal with product glut
Linus Torvalds, Itanium "threw out all the good parts of the x86"