35%

grid and power economist · openai/gpt-5.6-sol

Only 24 days remain, and the qualifying bar is unusually strict: a primary hyperscaler source or utility filing must explicitly connect more than 500 MW to power-first development or behind-the-meter generation. Large campus, land, or load announcements without that linkage fail. Quarter-end can concentrate disclosures, and strong generator/equipment backlogs support continued deal activity, but September is generally outside hyperscaler earnings season. Long interconnection timelines also favor exploratory announcements while making specific, documentable power commitments scarce. The small power-lens reference class is outweighed by the short remaining window and ambiguity resolving against the forecast.

semiconductor supply chain analyst · openai/gpt-5.6-sol

Only 24 days remain, sharply reducing the opportunity relative to the original forecast window. Strong utility backlogs, generation capex, and equipment orders support continued hyperscaler power procurement, but they do not imply an imminent qualifying disclosure. The rule is narrow: a hyperscaler primary source or utility filing must explicitly state more than 500MW and characterize it as power-first or behind-the-meter generation. Quarter-end can prompt announcements or filings, but ambiguous campus capacity, future land, or aggregated power plans would miss. The forecaster’s 0.60 confidence band has also historically overpredicted outcomes.

https://www.sec.gov/edgar/search/

capital markets credit analyst · openai/gpt-5.6-sol

Only 24 days remain, and no qualifying announcement appears in the supplied digest. Converting the original 0.6 over roughly 116 days to a constant-hazard probability for the final 24 days gives about 0.17. Strong power-sector RPO growth and active hyperscaler infrastructure spending raise the chance somewhat, but the rule requires an explicit >500MW figure plus power-first or behind-the-meter language in a hyperscaler primary source or utility filing. General campus, land, energy-procurement, or future-generation announcements would miss under the ambiguity policy.

superforecaster statistician · openai/gpt-5.6-sol

Base rate: 0.50 for disclosure questions (n=29); the 0.70 power-lens rate has only four observations. With just 24 days remaining, time-scaling the full-window base rate implies roughly 0.13. Strong supplier RPO and capex growth supports continued AI-power investment, modestly raising announcement risk. However, resolution requires a primary hyperscaler or utility source, an explicit capacity above 500MW, and power-first or behind-the-meter generation—not merely land or load plans. Those conjunctive disclosure requirements and ambiguity-against policy offset the positive sector signal. The forecaster’s 0.60–0.69 bin realized 0.556, confirming mild overconfidence.

geopolitics regulation analyst · openai/gpt-5.6-sol

Only 24 days remain from an originally roughly four-month window, with no qualifying disclosure supplied. A constant-hazard update from the 0.4 long-horizon base rate implies about 0.17 for the remaining period. Quarter-end announcement clustering and strong generation-equipment backlogs modestly raise that estimate. However, resolution requires a hyperscaler primary source or utility filing that explicitly exceeds 500MW and characterizes the project as power-first or behind-the-meter; generic campus, land, capacity, or power-procurement announcements fail. The lens:power sample is favorable but too small (n=4) to dominate.

direct · undated

SEC.gov | EDGAR Full Text Search

Cited by semiconductor-supply-chain-analyst: Only 24 days remain, sharply reducing the opportunity relative to the original forecast window. Strong utility backlogs, generation capex, and equipment orders support continued hyperscaler power procurement, but they do not imply an imminent qualifying d