Children: Sit down and be quiet.
Maybe it's lunar phases or whatever, but this thread, which tends to emit more heat than light anyway, has deteriorated below its already weak norms. I monitor the postings to see if there are some technical nuggets in what is often low-grade ore and am now depressed because time and energy are being wasted on Accounting 101.
A couple of "facts":
1) Intel has never (to my knowledge) released MPU units or ASPs, so the numbers bandied about are estimates. Let's assume that something north of 25 million units per quarter at $220-225 per unit gets us in the ballpark.
2) Intel's gross margin of 59% is a corporate average. It includes motherboards, NICs, ProShares, flash memory and every other product and service that generates revenue, just as AMD's gross margin includes flash, communication chips and, up until now, PLDs.
...Since MPUs represent a much less competitive market than many of Intel's other activities, it's probably safe to assume that the gross margins on Intel's MPUs run higher than 59%. But given the large percentage of the total revenues generated by MPUs, the MPU gross margin is unlikely to be radically higher than the corporate total (everyone can make their own guess, but does the math allow something much higher than 70%? you make the call). Within MPUs, it's hardly absurd to assume that profitability for the parts with no competition is significantly higher than those facing AMD (and Cyrix, RIP). Despite the claim that Intel "suffers" high costs associated with the Xeon (and I'll be staying awake at nights worrying about the Xeon's ability to cover costs - not), it might be that the part has gross margins of 85% in the larger cache configurations (just an informed guess on my part). Of course, the Xeon volumes are much smaller, so their gross margins are pulled down when blended into the whole (however, the notion of parts having gross margins in excess of 100% is an arithmetic impossibility, so let's jettison that from the discussion immediately).
...Given the volumes of the various parts, couldn't the Celeron line have gross margins of about 55-60%, PII/III about 65-70% and Xeon about 75-80%? Even if these numbers are not exact, the concept shouldn't be too far off. But the notion of the high-end making all the money and the low end losing it seems unlikely.
3) As mentioned by another poster, R&D is expensed, so by the time a product is developed, its revenues are funding the development of the next generation.
Now, an opinion - nothing that Intel has done strikes this observer as illegal or unwarranted. If anything, the reaction to the incursion of Cyrix and AMD in the low end was probably later than it should have been. But after the initial Celeron misfiring, the response has probably been as well executed as possible (within the context of a four-segment strategy, which is where I have misgivings, since I continue to believe that the desktop is but one segment). AMD's problems have been a combination of difficulties in responding in rapid-enough fashion to an impressive array of steps taken by Intel. It's misleading to attribute AMD's recent failures entirely to internal shortcomings or to Intel's aggressive measures, since they are invariably intertwined. I seriously doubt that there are any accounting issues involved.
I know that the class is restlessly awaiting recess ("it's XXX weeks until June"), but some of this behavior should really be saved for the playground. - Tad LaFountain |
