In a sense, "Microsoft is becoming a fabless semiconductor design firm," said Peter Glaskowsky, editor in chief of The Microprocessor Report. These companies--without their own chip-fabrication factories, or "fabs"--design their processors but outsource manufacturing to foundry companies... Commercially, Microsoft will differ from other fabless companies, such as Transmeta, in that its main customer will be itself. Still, Microsoft could conceivably leverage its investment by using the chips in many products."Not only will Microsoft co-design the CPU, they are partnering with ATI to design a custom graphics chips, and with SIS to design the chip set. This is a big difference from their experience with the first Xbox. For that, MS bought off the shelf OEM components. As another story today on CNet reports, the stakes in the set top box war are going up. Both Sony and Microsoft are determined to win at all costs. Microsoft has never been shy about using control of one technology to advance another.
Which brings us to AMD. Microsoft moving into the CPU business doesn't seem as big a stretch anymore, does it? I have no idea where this story originated. It may be a leak, it may be wishful thinking, it may be a trial balloon, or it may be rank speculation. Another Inquirer article by Phil Trent discusses the advantages of such a move. I think this article as mostly correct but Trent under estimates the potential antitrust trouble. There might also be significant resistance in the Far-East. Not enough to spoil the deal, but enough to make it difficult. It is no secret that Computer OEMs and consumer electronic companies love Linux. They have learned from 25 years of Sony taxes and the hated VHS royalty to support open standards whenever possible. And they don't trust Microsoft, but then they don't trust anybody. Apparently with good reason. The disastrous Xbox experience in Japan and Korea may also indicate intrinsic market resistance. Or, may simply indicate the Xbox's older demographic focus doesn't work in those markets. But it can't be ignored.
AMD is a very good buy at the moment. The financial markets have yet to digest the magnitude of their 64 bit victory. Microsoft also has piles of cash and very little that they can buy. Since the Intuit purchase of 1995 fell through because of antitrust concerns, Microsoft has been constrained in its purchases.
I tend to think AMD would be more trouble than they're worth to Microsoft. I also don't see a master plan in play. Except for initiatives like Longhorn and the Xbox, Microsoft tends to try lots of things and then reinforce success. All of these potential initiatives carry significant risks. Microsoft has historically been poor of dealing with strategic problems, but has effectively offset that failing with brilliant tactics. And Ballmer must be looking for an opportunity to put his own mark on the company. Not to boost his ego, he seems to have that under control, but to cement his regime.
Linux is clearly gaining momentum in the marketplace. In particular Sun's Java Desktop appears poised for some big wins. It must be clear in Redmond that margins in the operating system business are in decline. The delay of Longhorn puts their traditional means of boosting income out of immediate reach. The need for new sources of revenue is about to become acute. If they intend to move decisively into the silicon business we can expect that to happen quickly. However, the monetization of patents is going to move ahead smartly. I'd be very surprised if we don't see almost weekly announcements on this front. This is likely to become a significant revenue source for Microsoft, and a major cost of doing business for the rest of industry.
- "Toll Road, Page 1"
- "Toll Road, Page 2"
