The past year has seen some mighty struggles in the marketplace, with some big winners and some very big losers. VIA Technologies recently snatched both Cyrix and IDT from certain extinction, though it appears likely only the intellectual property will remain after the smoke has cleared. This essentially leaves only two main combatants locked in a fierce struggle, with possibly the same fate in store for the loser.
Just as we thought the battle would be engaged in full fury, it appears that it the decision may be a technical knockout because of a no-show. After one competitor has made steady gains in the low-end over the past year, while gearing up for a bloody battle at the high-end, it seems that another product delay may determine the outcome.
That’s right, rumors have it that Coppermine will be late…
On Monday, independent benchmark numbers were released for the AMD Athlon processor, showing a significant speed advantage over the Pentium III. It was also announced that AMD will release a 650MHz version of the chip in a move to become the leader in MHz as well as performance. In the meantime, Intel has delayed the introduction of their .18 micron Coppermine processor due to ‘shrinkage’ problems, forcing them to push the .25 micron core to 600MHz just to stay competitive. Now the rumors say that the 600MHz Coppermine may not be released until Q1 2000, with a 500MHz version debuting in November.
Of course, this shouldn’t be too surprising considering Intels history of missed targets…
What??? Intel has a history of missing their targets? Isn’t it AMD that has this reputation? Actually, it is AMD that has the reputation, but apparently because the Intel PR machine is exceptionally good at what they do.
Just to prove the point, let’s look back at a report from Sept 19, 1996 that indicates the Merced processor was to ship in late 1998. Today the target is late 2001, and even that is questionable. Even better is a report from Dec 27, 1996 that gives the anticipated target dates for Klamath, Deschutes, Katmai, Willamette and Merced. Granted, these dates were actually pulled from a report by Michael Slater of MicroDesign Resources, a well respected processor technology analyst, but his info is generally regarded the next best thing to official Intel statements.
Using the information provided in the December report, here is a table showing the original dates, and the actual (or projected) dates:
Klamath | Q2 1997 | May 7, 1997 | Hit Target |
Deschutes | Q4 1997 | Jan 26, 1998 | 1 Qtr Late |
Katmai | 1H 1998 | Feb 26, 1999 | 1 Yr. Late |
Willamette | 2H 1998 | 1H 2001 (proj) | 2 Yrs+ Late |
Merced | 1H 1999 | 2H 2000 (proj) | 1.5 Yrs Late |
This trend is not restricted only to processors. The i810 chipset was originally scheduled to be released in April of this year, but only now are motherboards using the chipset available. This is due to at least two delays caused by problems with the chipset. Of course, the i820 chipset was delayed two months also.
Strangely, Intel doesn’t seem to get beat up over these delays like their competitors do when they are late. It seems as if the Intel PR department is exceptionally good at spinning a tale that dazzles the media and makes them forget that Intel once again missed the boat.
In 1997, AMD grabbed significant market share at the low end with their K6 processors – a market that Intel had ignored. The threat was significant enough that Intel introduced their own low-end processor, the Celeron, in May 1998. This first attempt was a dismal failure, but Intel followed up in August with the Mendocino, which was simply a Deschutes core with 128K of on-die L2 cache. Though Intel was actually playing catch up with AMD, they have since been hailed as the ‘patron saint’ of the low-end systems by some because of their aggressive move to take over that market.
If one looks objectively at the facts, it becomes obvious that Intel has been caught flat footed once again. For the past year, Intel has been pooh-poohing PC133 SDRAM but now find that they must rush to support it. Now we find that they have no immediate answer for the Athlon and may be forced to drastically cut prices and fight to keep dominance at the low end – a position that AMD was in only a few months ago. In fact, Intel just announced an unscheduled price cut for August 22 on Pentium II and Pentium III processors.
Obviously, it is not reasonable to count Intel out and they certainly have the assets to survive a very long war of attrition. They also have built up substantial brand-name loyalty amongst small, mid-sized and large businesses, which is where the real revenues are generated. The fact remains, however, that Intel has been flanked more than once this past year, and have suffered some surprising loses in the marketplace. It looks like this trend will continue for at least another 6 months or longer.
With today’s fight already apparently lost, Intel must now focus on winning the rematch in late 2000, when Willamette appears. That is, as long as they don’t have yet another product delay…
Be the first to discuss this article!