In my work with utilities who are trying to attract utility-scale data center development to their service area, I make it clear that power price is the ultimate differentiator, with availability/time to serve and options for clean power secondary criteria.
The economic development representatives I've met with seem to grasp the concept pretty well, but when you test them for their power price "story", the tale is usually pretty hazy!
Let's start with the answer to this question: "What is your power price for a multi-MW data center?".
Most utilities I've worked with answer that question with "X.X cents per kWh", but in fact they are quoting an average industrial cost that ignores several key features of loads of this type.
Where does the utility rep get their price per kWh quote? Likely from the Rates Department (or equivalent), where an analyst takes all of the revenue from a given class of customers and divides it by the total kWh usage of the class.
So, for multi-MW loads, you quote the "average industrial rate" - the total revenue collected from all industrial customers (generally classified as those with peak loads of 1 MW or greater) divided by the energy use by those customers.
There are however pretty wide disparities in the rates charged to those customers, with load factor/shape and service level chief among them.
Customers with "peaky" load profiles, low load factors, and who take service at secondary or primary voltages pay lots more than customers with flat load profiles/high load factors who take service at transmission voltage.
Here's what I would do to derive a more accurate power cost quote for utility-scale data centers: I'd take a 1 MW load, assume a unity load factor, and populate a spreadsheet with 1 MW of demand and 1 MWhr of usage across 8760 hours per year.
Then I'd take my rate schedule, and I'd plot these charges in the spreadsheet:
- Customer Charge (Monthly)
- Meter Charge (Monthly)
- Power Factor Adjustment (Applied Monthly)
- Peak Demand Charge (Monthly)
- Maximum Demand Charge (Monthly, set by highest demand in previous 12 months)
- Summer On-Peak Energy Charge
- Summer Partial-Peak Energy Charge
- Summer Off-Peak Energy Charge
- Winter Partial Peak Energy Charge
- Winter Off-Peak Energy Charge
- (Set all demand and energy charges based on service voltage of secondary/primary/transmission)
If your rate structure mirrors all or many of these features, I bet the average rate you derive from this analysis is lower than your "average industrial class rate".
Now, to complicate matters for some utilities, they now have peak-period pricing reflecting actual market rates, moving down the road to real-time pricing. At the other extreme, you might find some utilities that have very simple industrial rates, though a straight per-kWh rate or even declining block rates are relics of the past for most of us.
I'll cover another aspect of the power price "story" that utilities need to be able to tell when courting utility-scale data center developers, but quoting your rates accurately for the type of load you are attracting is key.
You need the opening of your story to be: "For your multi-megawatt load, assuming you take transmission voltage service and have a unity load factor, your average rate will be X.X cents per kWh (quite a bit lower than the average rate for our customers in the industrial class)."
Recent Comments