US vs EU Interconnection Queue: Scale, Composition, AI Collision Risk

The US queue is roughly five times larger than the EU by project count, but that scale is distributed unevenly. (Source: LBNL Queued Up 2025.) A majority of AI-adjacent interconnection pressure is concentrated in just two ISOs: MISO and SPP. Concentration of this kind amplifies collision risk beyond what aggregate numbers suggest. A single ISO under saturation pressure can block more MW of AI infrastructure than a distributed EU-scale permitting backlog, because the bottleneck is localized and cannot easily be rerouted.

Texas (ERCOT) and the Midwest (MISO/SPP) are early saturation zones. MISO alone accounted for approximately 20% of total US AI-adjacent interconnection requests in 2024. (Source: LBNL Queued Up 2025; MISO queue filings, AI-adjacent classification.) This geographic concentration means that capital allocation decisions that look balanced at a national level may be highly exposed when mapped to ISO-level queue pressure. Withdrawal velocity is rising in these clusters – a leading indicator that projects are discovering grid constraints after initial capital deployment, not before it.

The EU does not have a single interconnection authority below the ENTSO-E level. This creates a structural asymmetry: EU permitting averages approximately 7.4 years versus roughly 4.2 years in the US, and withdrawal data is not consolidated across member states. (Source: ENTSO-E Statistical Factsheet 2024; LBNL Queued Up 2025.) The UK adds a further complication post-Brexit – limited continental interconnector fallback creates structural scarcity that is not visible in ENTSO-E headline figures. Both markets show withdrawal velocity rising in AI-adjacent technology clusters, but the EU signal is harder to read because the queue data does not exist in a normalized form.