Power Demand/Supply Refresh: More Announcements Than Electrons

研报英文原文证据摘录

Power Demand/Supply Refresh: More Announcements Than Electrons

neration may continue to slip. Even after risk-adjusting the pipeline, data center

load growth remains large enough to reshape the grid.

Challenges to developing data centers is resulting in three distinct strategies:

1) Pivot towards Behind-the-Meter and Hybrid models to enable expedited development. We

increasingly call this 'Hybrid' as data center operators will take 'all they can get' from the grid first,

and then supplement with BTM strategies. Question remains in many instances what is realistically

possible from gas pipeline supply, and how adjacent are gas resources. Co-locating to gas pipeline

access is critical (if not multiple gas pipeline options). Hybrid sites are the calling for the moment.

2) Larger data centers at existing sites. We see a 'doubling-down' effect at existing data center sites.

Meta's recent experience in Louisiana appears to be followed by others. This is not about expanding

the duration of the ramp phase but rather adding more in parallel to ramp faster at permitted sites

with adjacent data centers. We see this as being the trend for the balance of 2026. Watch Google in

particular around this strategy in IA, MO, and MI among other sites. A key debate is how data center

development trends as the model shifts from training focused to inference use cases. This in theory means

less development in markets with high latency and a pivot towards tier one markets again.

3) Maximizing what transmission deliverability can be via storage. Substantial focus on curtailment

of late, but really it is about T&D deferral with utility partnerships to enable more transmission without

overloading grid substations on peak days. This concept appears poised to grow meaningfully. Expect