Google spent $204 million on data center operational expenses in the second quarter of 2006, compared to $181 million in the first quarter. Here's an interesting wrinkle: because Google's income continued to improve, the percentage of revenue invested in data center capital expenditures remained steady at 8 percent.
So Google's data center investment is on track to exceed $800 million this year. In all likeihood, it will be even higher, according to Google CFO George Reyes. Here's an excerpt of an exchange between Reyes and Jeffries & Co. analyst Youssef Squali during yesterday's conference call:
George Reyes: On your question around CapEx, Eric just had a wonderful suggestion, which is to blame it on Larry and Sergey, all this CapEx that we're adding. But realistically, we think we can't put too much CapEx into the system. It's a really critical part of our competitive advantage and our infrastructure. It's something that you're going to continue to see quite substantial investment going forward over the next few quarters.
Squali: No acceleration or deceleration?
George Reyes: There will be definitely an acceleration.
Google isn't bashful about its growing capex investment because it's data center infrastructure plays such a critical role in the company's competitive advantage, Reyes said.
"Our investments in core technology historically have paid off very handsomely for us in terms of our results," said Reyes. "Rest assured that the decisions that we're making around capital expenditures are carefully thought out and prudently debated. Note that a significant portion of our CapEx spending is related to build versus lease decisions around data centers; and buy versus lease decisions around real estate. By making these investments upfront, we expect to reduce our overall cost of operations over the long run."
An unrelated but nonetheless interesting tidbit from the conference call came from Sergey Brin, and demonstrated Google's ever-expanding awareness of exactly what it users are doing online and when.
"In the quarter we made numerous changes to the way we serve ads that help us to show the best ad at the best time," Brin said. "For instance, one of our improvements involved showing fewer ads while users were looking for non-commercial information rather than seeking to purchase something, while showing more highly relevant ads when our users are searching for products and services."
Which means they probably have an algorithm that knows the difference. It's just one more way Google is leveraging its mother load of customer data - the "database of intentions" described by John Battelle.
