Is AMD now ready to impact the enterprise space?

IT IS PRETTY clear (see part two) that AMD has fully demonstrated the exacting requirements that the business world demands. Market perception of AMD's supposed inferior platform quality has taken all this time to erode. HP's decision to offer AMD supported (Compaq branded) desktop and notebook business class PCs should be all the assurance that one would need that AMD is now ready to aggressively compete in this space. Many system builders and second tier OEMs have been successfully selling AMD powered business systems for many years. HP's decision is a reflection of what has become a market reality.

Will the enterprise momentum continue with Hammer?
The majority of the IT world were caught sleeping at the wheel when companies like Rackspace and Cobalt Networks exploited the server world attributes of AMD's K6-2. Many of the major industry players have now woken up and are now successfully using AMD Athlon-based systems in varied commercial environments. But with the promise of what AMD's Hammer will bring to the table next year, this momentum looks set to continue.

The immutable laws of the PC industry for transition to occur
With AMD's Hammer due to debut next year, and Intel's FUD campaign now in full swing, let's remind ourselves of what these laws are.

Jerry Sanders, chairman of AMD, talking about the Athlon to Hammer transition, explained what these laws are:
- New product must cost less while providing equal functionality/performance to previous generation, or,
- New product must provide greater functionality/performance at a cost comparable to previous generation, and,
- New product must maintain legacy compatibility

Hammer will provide all the above:
- For all customer segments (commercial and consumer)
- For all product segments (desktop, mobile, server and workstation)
- Common architecture in both hardware and software
- Next generation 32-bit performance
- Evolutionary 64-bit capability

Pat Gelsinger, Intel VP and CTO, co-authored with John Crawford, chief architect of the 80386, "Programming the 80386." In the introduction of their book, they clearly understood the importance of legacy compatibility, which is in total harmony with AMD's position today.

"An important question to answer is: Why should you be reading this book? Why will your understanding and programming of the 80386 benefit you for the next decade or two of your programming career? The answer is in the tremendous cumulative investment in the 86 family. To design and use a computer, investments are continually made. These investments are by those designing computers, operating systems, programming languages, application programs, and additional hardware. The investments also include programs you may write yourself and, of course, your time to learn. Thus, computer families, such as the 86, evolve and share compatibility from one generation to the next. This compatibility allows the use and leverage of massive investments already made into a computer family."

If this was true in 1987, and four more x86 generations have since been added, the cumulative investment has now grown to be many more times the massive investments of 1987. History has taught us that we ignore it at our peril. So why did Intel run roughshod over these laws and develop Itanium to be totally alien to them?

What is ironic, is that Pat Gelsinger, as CTO, has prime responsibility for steering Intel's technological path, which encompasses Itanium. The same can be said of John Crawford, who is an Intel Fellow, Enterprise Platforms Group, and also Director, Itanium Architecture. It would be interesting to listen to how they reconcile what they said in 1987 with Itanium's lack of legacy (ignoring emulation mode which nobody would seriously use) compatibility and its total contradiction of these laws.

What will it take for AMD to reach critical mass?
When Hammer debuts next year, AMD's product offerings will grow to a level where it should be easier to sell into the commercial mainstream. AMD is now preparing the ground with its Assured Program, which should convince the corporate buyer of the seriousness and commitment that AMD and its partners have to supporting this important market.

It should be clear to the independent observer that AMD produces the best price/performance processors in the industry. But what will it take for AMD to significantly improve its market share position? Four scenarios - technological, financial, Yamhill, and Dell; be it any one or any combination, could bring about this change.

If Intel is not able to develop the P4 and Itanium, in the same way that AMD's Hammer is able to build on Athlon, Hammer could prove to be the technological straw that breaks the camel's back and takes AMD to critical mass and beyond. If this should happen, Intel's market dominance and control will be curtailed. The next three years may see this happen.

At a past technology conference, Jerry Sanders shared an interesting corollary that Alan Greenspan, chairman of the Federal Reserve, made concerning companies that hold dominant market positions. He said, "While it may be possible for a competitor to dominate a market for an extended period of time in free competitive markets, that is possible only if dominance is maintained through cost efficiencies and low prices that competitors have difficulty matching."

It is clear that Intel's processor pricing model is still very profitable, but its cost efficiencies are not reflected in low prices. If AMD is able to make significant market inroads at Intel's expense, and the chip giant is unable to compete at the price points of its competitor, then this may prove to be the financial noose that destroys its dominant position.

Jerry Sanders has said that his biggest fear would be an Intel processor that is based on AMD's x86-64 instruction set, which Intel has every right to do. But Intel's Paul Otellini is reported to have said that Intel won't do this or even release its own (Yamhill) hybrid version. Instead, he said Intel's future is firmly fixed in the Itanium camp. When one reflects on this and the immutable laws that were discussed earlier, and IBM's decision to offer its own PowerPC-based 64/32-bit hybrid, Intel's decision looks to be way out of step with industry thinking. If Intel is immovable from this decision, which I think it is not, this could turn out to be the biggest own goal that Intel has inflicted on itself and its shareholders. This one decision could put AMD on a growth curve that could substantially increase its market share position.

Getting Dell to sign up to Hammer would be AMD's biggest coup of its design win history. Currently the number one OEM in the world, having just pushed HP off its pedestal, Dell's continued growth continues to amaze. If Hammer is adopted, then Dell with Hammer would be a frightening prospect to compete against.

Of course, to bring this all to pass, AMD has to execute its roadmap and also bring UMC, its foundry partner, up to speed as well. If this all happens to the current timetable, and any one of the scenarios pan out, then the future for AMD and its processor business will look very rosy indeed.

The choice IT buyers have to make
History can be very instructive and beneficial if one acts in a way that accounts for the failures and misdeeds of the past. Microsoft is a classic example where illegal means were used to destroy competition and then reap the benefits of a monopolistic position. We all know the results of those acts: Limited choice, high prices, enforced upgrades, and pricing schemes that don't take into account the needs of the customer. Companies have reacted by moving to Linux and other technologies. But how does all this tie up with hardware? It's simple, competition. You've seen and have probably experienced the results of Microsoft's abuse. Now think about this from a hardware perspective.

No one ever got fired for buying IBM, but the majority of buyers today buy Intel-based systems. If two competing PC products, Intel and AMD, meet the requirements of a purchase order, which one should the IT manager buy? Hopefully, any IT buyer worth his salt will be fully abreast of the products that can be used to leverage the best possible deal. IT buyers not only have the responsibility to purchase the best PC product that meets the need at the best price, they also have to think about the long term impact that today's buying decision will have on future PC costs. If there is no competition, or the competition is weak, it is not difficult to surmise the impact that this will have on pricing.

The current share of the x86 processor market is probably around 80/20, give or take a few points, in Intel's favor. Coming back to the question, if IT buyers want to benefit long term from continued competitive pricing, it seems obvious that they would have to buy AMD. How long should one endorse such a buying policy? As long as it takes for AMD to reach market parity with Intel. Let's not forget VIA. It sells some very low cost PC products which could prove to be very successful in its target markets. At the end of the day what's best for everyone is competition. If market share becomes more evenly distributed, then it's a win-win for the buyer. In their own way, the IT buyer can influence the PC market for the common good.

Going the extra mile
It is sad to think that even with all the evidence that is before us, AMD still had to go the extra mile to convince business users that they can provide solutions that are not only stable, but have passed a broad range of validation tests which gain its assured status approval. As Jerry Sanders said not so long ago, "We're getting the nose of the camel into the tent." That is most definitely true, but what an effort it has taken just to get this far. If AMD can build upon this progress with Hammer, then the camel may be poking its head out of the other end. If this should happen, the change this would cause to the enterprise landscape would not look too pretty from Santa Clara.

The HP business desktop and notebook design wins will complement AMD's Assured Program and the other areas of its enterprise business. If businesses really want to see some real competition, which results in bottom dollar pricing on systems that they purchase, then they know what they have to do. As Jerry Sanders is so fond of saying: (paraphrased) AMD stands for competition. Competition is good, and we're good competition for Intel. We help our customers work smarter, and our customer's success is our success.

Fulfilling the AMD dream
Anyone who has listened to past presentations by AMD, will be familiar with the descriptive language that Jerry Sanders has used to explain his company's position. The AMD dream is one such example. The one-thousand-day-journeys is another that he's used to differentiate the different AMD periods. This began back in 1997, which means we're now in the early part of the third dispensation. It would be fitting for the chipmaker, if the AMD dream that Jerry Sanders has so eloquently espoused over the years, would become a reality in the current era. Part of that dream is to effectively compete with Intel top to bottom in all markets. With Hammer debuting next year, this will soon become a reality.

At the last AMD shareholder's meeting, which was the last that Jerry Sanders chaired as CEO and chairman, responding to a question, Sanders said, "In my view the 7th generation, the Athlon, was designed by the finest micro processor design team on the face of the earth." Recalling Bill Gates' comments on the original Athlon - "Jerry, we think Athlon is a home run" - Sanders went on to say, "If that's the case, Hammer Opteron is a grand slam." When the AMD dream becomes a reality, will you be jumping ship to switch to AMD? µ

See also:
Part 1: Serious about servers
Part 2: Would you switch to AMD?.