I delivered three presentations at the "Data Centers and Utilities" session hosted by EUCI in Philadelphia late last month, and my opening session was significantly updated from the same conference held in San Francisco last year.
I must admit that I had to think on my feet for all three presentations, as the dozen-and-a-half participants were all from utilities (no data center operators or developers were there). And not only that, these weren't utility energy efficiency program managers, they were to a person involved in economic development activities.
Three utilities delivered presentations on the programs they have developed to try to attract data center development in their service areas:
- The Tennessee Valley Authority (a wholesale energy provider) described how they had worked with their utility customers and local authorities to identify and qualify about fifty prime sites in the seven-state region they cover. The information package they have for each site is really impressive, as well as the work they've done with electric planning engineers to identify available power capacity.
- American Electric Power (which has several utilities under it's corporate umbrella) has done about the same level of work, though they have identified only nine sites so far.
- And local utility PECO (an Exelon company) described how they were working with existing data centers in the greater Philadelphia region to help them expand in the area.
In my opening presentation, I noted that data center developers are looking for just a few key attributes when making a site selection: current and future power costs, power availability, and time to market (how long it takes to get the infrastructure in place).
I did point out that there are a handful of operators who are acutely interested in carbon content, but that most developers are interested in that topic only as a determinate of future power prices.
Perhaps the most interesting connection I made for the industry move to utility-scale data centers is that these facilities have several underlying cost advantages compared to enterprise facilities, with many of them representing a two-to-one or greater edge.
For example, these centers can be twice as efficient as legacy enterprise centers, and be located in areas where utilities are half as expensive as urban coastal areas. When you compound the inherent underlying cost advantages, you can predict an inexorable move to utility scale facilities offering hosting or cloud services.
Here's the presentation: