Is this an attempt to get into some Texas-Crypto demand response bounties? If the premium offered to reduce demand exceeds the costs of (risk of) breaking SLA for the datacenter operator they'll gladly take the money and shift the load to a neighboring region.
ICYMI - PGE in Oregon has proposed a large load tariff that has some interesting provisions on flexibility - basically allowing more than a pro-rata share of available capacity if customers commit to dispatchable flexible demand programs: https://edocs.puc.state.or.us/efdocs/UAA/ue430uaa333701025.pdf
For those not familiar with cloud capacity, there has long been a huge discount for interruptable loads, in the form of "spot" instances. Would be amazing to see this extended all the way out to power consumption.
I think it really depends on what compute workloads predominate the data center. Batch workloads like training AI models should be pretty interruptible.
Comments
A bad outcome for consumers.