A large load can connect years before the transmission it would otherwise need, if it accepts curtailment. Nobody can price that curtailment, so nobody accepts it. We measure it from public data.
FERC has directed all six grid operators to offer flexible service to large loads. The question of whether it exists is settled. How anyone measures what accepting it costs is not — and a developer who cannot put a number on curtailment risk cannot finance the site.
We publish that number for 44 interconnection-relevant nodes in Northern California, built entirely from CAISO settlement data. No utility data agreement, and nothing a recipient cannot recompute.
Nothing in any published interconnection metric tells a developer which is which.
And the hours are forecastable a month out: 31% median error against 65% for the standard benchmark, better at 39 of 45 sites. Walk-forward, never fit on its own future.
A congestion price is a market outcome. An instruction to curtail comes out of a utility's contingency analysis, from an N−1 event, not from an economic price. Those can diverge in exactly the hours that matter.
This is a screening layer. It tells you which sites deserve a power flow study. It does not replace one, and we will not claim it does.
The console is live. Every figure recomputes from public CAISO data, and the failures are published beside the results.