The LBNL/Brattle finding about the spread of utilization costs is the part most coverage misses — the beach house analogy lands.
Quick question on Part 2: PJM's last two auctions cleared at a FERC-approved cap of around $333/MW-day, and PJM estimated the uncapped price would have been closer to $530. When the cap expires, do you think state-level rate-class solutions like Virginia's are sufficient, or does the cost-allocation problem need to be solved at the RTO level?
There is bad load growth and there is good load growth. Environmental advocates decry bad load growth associated with data centers because it affects affordability. They ignore the fact that their quest to decarbonize the economy will also increase load growth but that is all in a good cause.
Your description of THE STACK raises another unacknowledged environmental advocacy issue. The use of a magical cap-and-invest program that can ensure compliance while simultaneously raising money to pay for the transition assumes that the only costs are associated with the purchase of allowances. Those proceeds fund all sorts of favored efforts and the cost of transition if there is anything left over. What they do not realize is that cap-and-invest in the electric sector raises generator costs and affects THE STACK payouts so consumers are paying for that aspect too. I think this could easily double the cost for the Regional Greenhouse Gas Initiative program but all that money goes to the generators including those that have no compliance obligations.
Really mitigate the cost by having the data center generate their own electricity. Near constant demand right? So build the appropriate size plant and don’t impact the grid. Alleviates all of the mental and theoretical exercises.
Yes, that is one of the approaches I will be discussing in Part Two of my post. There are some practical problems involved that make me think this won’t work for all new data centers, but it certainly is a viable option for some of them
I would be curious to hear your thoughts on the overall impact of data center demand on costs in the supply chain of the energy industry. For example, the cost of gas turbines has increased, the cost of transformers have increased, labor costs for construction have increased. Timelines for receiving these critical pieces of equipment have increased, which generally increases costs. While DC’s may pay for these directly for their facilities, their increased demand is driving all of these other cost increases at the moment. It seems to be more of a societal impact rather than the impact of a specific project.
One additional comment, while the buildout of DC’s at this new increased rate is recent, I do know of specific examples where DC facilities came on line early in this cycle without the level of scrutiny they now receive and were able to get there energy based on “slack” or over capacity in both the gas and electric systems. I think in most areas of the country that is no longer the case, and that may be why you saw rates go down initially in areas with high DC growth because they more efficiently utilized overbuilt systems. Those days are over.
Thanks for those excellent comments. I am aware of the supply chain issues but haven’t really focused on them so far. Maybe that would be a good subject for a future post. I also agree with your point about early entry data centers using up available capacity without affecting prices. That was the point of my beach house analogy.
Thanks Matt, I understand and appreciate the beach house analogy. I would offer a slightly different perspective. The ratepayers paid for the slack in the system, it provided resilience. Having the extra guests at the beach house made the rent cheaper for a while. I think it remains to be seen whether or not they’ll be paying for the renovation of the house that’s now being required of our energy infrastructure grid. It’s challenging, and I truly appreciate your thoughtful approach to analyze this.
The LBNL/Brattle finding about the spread of utilization costs is the part most coverage misses — the beach house analogy lands.
Quick question on Part 2: PJM's last two auctions cleared at a FERC-approved cap of around $333/MW-day, and PJM estimated the uncapped price would have been closer to $530. When the cap expires, do you think state-level rate-class solutions like Virginia's are sufficient, or does the cost-allocation problem need to be solved at the RTO level?
Thank you for this article. I have two comments.
There is bad load growth and there is good load growth. Environmental advocates decry bad load growth associated with data centers because it affects affordability. They ignore the fact that their quest to decarbonize the economy will also increase load growth but that is all in a good cause.
Your description of THE STACK raises another unacknowledged environmental advocacy issue. The use of a magical cap-and-invest program that can ensure compliance while simultaneously raising money to pay for the transition assumes that the only costs are associated with the purchase of allowances. Those proceeds fund all sorts of favored efforts and the cost of transition if there is anything left over. What they do not realize is that cap-and-invest in the electric sector raises generator costs and affects THE STACK payouts so consumers are paying for that aspect too. I think this could easily double the cost for the Regional Greenhouse Gas Initiative program but all that money goes to the generators including those that have no compliance obligations.
Really mitigate the cost by having the data center generate their own electricity. Near constant demand right? So build the appropriate size plant and don’t impact the grid. Alleviates all of the mental and theoretical exercises.
Yes, that is one of the approaches I will be discussing in Part Two of my post. There are some practical problems involved that make me think this won’t work for all new data centers, but it certainly is a viable option for some of them
I would be curious to hear your thoughts on the overall impact of data center demand on costs in the supply chain of the energy industry. For example, the cost of gas turbines has increased, the cost of transformers have increased, labor costs for construction have increased. Timelines for receiving these critical pieces of equipment have increased, which generally increases costs. While DC’s may pay for these directly for their facilities, their increased demand is driving all of these other cost increases at the moment. It seems to be more of a societal impact rather than the impact of a specific project.
One additional comment, while the buildout of DC’s at this new increased rate is recent, I do know of specific examples where DC facilities came on line early in this cycle without the level of scrutiny they now receive and were able to get there energy based on “slack” or over capacity in both the gas and electric systems. I think in most areas of the country that is no longer the case, and that may be why you saw rates go down initially in areas with high DC growth because they more efficiently utilized overbuilt systems. Those days are over.
Thanks for those excellent comments. I am aware of the supply chain issues but haven’t really focused on them so far. Maybe that would be a good subject for a future post. I also agree with your point about early entry data centers using up available capacity without affecting prices. That was the point of my beach house analogy.
Thanks Matt, I understand and appreciate the beach house analogy. I would offer a slightly different perspective. The ratepayers paid for the slack in the system, it provided resilience. Having the extra guests at the beach house made the rent cheaper for a while. I think it remains to be seen whether or not they’ll be paying for the renovation of the house that’s now being required of our energy infrastructure grid. It’s challenging, and I truly appreciate your thoughtful approach to analyze this.