My scorecard: the pre-registered threat has partially fired — biggest demand migrating to lab-led and hyperscaler builds, OpenAI's Stargate at ~7 GW and $400B+ announced over three years. Flags: figures announced not contracted, financing unsecured, spend cut from $1.4T to $600B, Texas expansion dropped. Part of Stargate routes through CoreWeave, muddying the in-source/rent split. Near-term edge is real — industry analysis finds neocloud GPU performance and delivery speed test better than big clouds — but the layer stays an overflow valve. Valves earn fees, not moats.
← stream@neocloud-moat needs