Practical Chip Design

Feb 1 2008 10:49AM | Permalink | Email this | Comments (2) |
Blog This! using:  Blogger.com | LiveJournal |
Digg This | Slashdot This | add to Del.icio.us

Aside from any personal or professional feelings about Microsoft, Yahoo, the former’s attempt this morning to purchase the latter could turn out to be a glint of good news in a very dark 2008 for the semiconductor industry in general, and for the SoC business in particular. This is for two reasons.

First, servers. Google, with their Google Earth, reference archives, and on-line applications, has defined the battle for dominance in on-line revenue in very simple terms: it’s a war of server farms. If you aren’t installing servers and storage arrays in increments of at least one American football field (hereby christening a new unit of server farms: one AFF), you are not playing. This will become even more the case as the focus in these areas starts to shift from shared applications to the creative use of user-generated data, such as incorporation of user-contributed still photographs into the Google Earth database.

But Yahoo, with it’s current financial structure and mediocre market capitalization, and with the current train wreck in the credit markets, is not going to be installing any AFFs of servers any time soon. It’s not clear that they could afford the power to operate and cool them, let alone the capital investment. And so the consumption of servers for these mega-farms would be limited to whatever Google feels is necessary to explore new applications.

If Microsoft succeeds in devouring Yahoo, all that changes. The people in Redmond understand this arms race perfectly well, and they have the cash to make it a real race. Competition between two server-laden giants would not only rapidly increase the installation of AFF after AFF of servers, but it would escalate the development of applications to absorb them. This translates instantly into good news for all the semiconductor suppliers that sell into the enterprise server market, as well as into a bonanza for enterprise-level switching gear. It would also stimulate the development of new ideas in server interconnect to improve latencies for distributed applications.

But there is a second effect as well. Such applications, as they draw in more user interactivity and push richer data streams back to users, will create greater demands on user terminals: especially the now-anemic mobile devices such as cameras, handsets, positioning devices, and mobile computers. This could justify a wholesale replacement of mobile electronics, giving a major boots to a market that appears to be saturating and drowning in useless feature creep. And that, in turn, would be a shot in the arm to developers of SoCs for both mobile applications processing and wireless connectivity.

So apart from anything specific about the nature or culture of either company, the mere fact that the acquisition of Yahoo would channel a huge amount of Microsoft’s wealth into direct competition against Google could be a much-needed gift to our little corner of the business. Not all unintended consequences are bad ones.