Intel's Desktop Forecast
Desktop CPU Forecast Chart
Now we're cooking
OK, this is what everyone's undoubtedly interested in - desktop CPU pricing. This is where all the new technology surfaces first. The trends here don't necessarily follow those for either the mobile or server/workstation market, and the desktop field is where you'll see Intel react to any potential threat. Notice how desktop prices are separated and adjusted monthly.
So what can we expect?
Let's take a quick look at Pentium III first. The Pentium III 550 has just recently been released, and it looks like this is going to be one of the last .25um CPUs. Starting in August and September, we're going to see the P3 600 make one last encore at .25/512KB, and while 100MHz FSB will stay around for quite a little bit after that (mainly to flesh out the big jumps between 133MHz multipliers), everything after that is going to be Coppermine-core .18um, with integrated 256KB on-die cache.
Of course, 133MHz will have to be introduced concurrently with the "Camino" 820 chipset, which is being held back due to availability and pricing of Rambus Direct RAM. Intel has no intentions of supporting the current initiative towards PC133 or DDR SDRAM, but we'll see where the market pushes on that issue. Depending on 820 delays, Intel may decide to release Coppermine P3s at 100MHz FSB SECC2 first, in order to quell the rising tides, and stall for 820.
Is that the REAL reason?
Those keeping up with the news may have noticed that Intel has already begun shipping .18um CPUs, primarily in the form of mobile P2 and Celeron processors. A lot of people believe that this is high-volume "test run" to warm the fab plants up for the impending arrival of the AMD K7. While experience tells us that coincidence is usually the rule rather than the exception, we have no doubt that Intel has the capability to push out Coppermine Pentium IIIs at 600Mhz and up, right here and now, and definitely by the end of June.
While Intel has no reason to immediately fear AMD's K7 impact, they're not blind to the long-term implications. The fact is, their only option at this point is to take on AMD in raw MHz - put out faster rated chips for cheaper. This can be an annoyance for Intel, and heck it may hurt their quarterly earnings reports, but it's a practice they've perfected against AMD, a company that is depending on K7 to achieve profitability.