Skip to content

AI Market · Forecast desk · Live · capital

27%

A definitive agreement of at least $100B in vendor-guaranteed AI data-center financing is publicly documented in a filing or company release by December 31, 2026.

Probability history

0%25%50%75%100%08-0109-06deadline

Called at 34% on August 1, 2026. Repriced 1 time; now 27%. 116 days to deadline.

How the number was built

Final method: stacked-fixed-weights (resolved n=57 < 80). Each rung is an independent estimate; the final is a stacked combination.

Stated (issue date)34%
Reference class (n=29)50%
Recalibrated52%
Feature model41%
Council13%
Final27%

Drivers

  • Reference class qtype:disclosure runs 0.50 (n=29) vs stated 0.34.
  • Hierarchical calibration moves stated 0.34 to 0.52 (lens capital, kappa_stated=24.0, kappa_lens=200.0).
  • Feature model at 0.41: conjuncts=1, named disclosure=True, horizon=152d.
  • Council of 5 at 0.13.

Cautions

  • all 5 votes came from one vendor (openai); extremization disabled, persona diversity only

Council

5 personas on one model (openai/gpt-5.6-sol); other vendors were unavailable this run, so read this as one model's persona-averaged view, aggregated by log odds mean without extremization.

14%

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

I am below the forecaster’s 0.34. Large and rapidly growing RPOs support extraordinary AI infrastructure demand, while low broad high-yield spreads and subdued volatility make financing feasible. But a single signed financing of at least $100B with explicit vendor-guarantee language would be unprecedented in scale, create major contingent-liability and concentration issues, and likely require extensive syndication and approvals. Transactions are more likely to be staged across facilities, projects, or counterparties. The strict requirement for a definitive 8-K or equivalent—and rejection of MOUs—materially lowers the four-month hit probability.

21%

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

Capital markets are supportive: investment-grade spreads and volatility are low, while public-company RPO and capex figures confirm exceptional AI infrastructure demand. But a single $100B financing is extraordinarily large, and explicit vendor-guarantee language would require complex credit approval, documentation, and likely material disclosure. Large customer commitments or vendor investments would not qualify. With only 116 days remaining, an announced framework or MOU could easily miss the definitive-agreement test. The strict 8-K-or-equivalent requirement and adverse ambiguity policy put this below the forecaster’s 0.34.

12%

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

Capital demand is plainly large: the digest shows multiple public issuers with RPO above $100B and rapid capex growth, while broad high-yield spreads remain benign. But the threshold concerns one signed financing—not projected spending or RPO—and requires explicit vendor-guarantee language. A $100B guarantee would create extraordinary contingent exposure, requiring extensive approvals, covenants, and material disclosure. Only 116 days remain; a project announcement, customer commitment, consortium, term sheet, or MOU would fail. Single-B OAS at 10.51% also complicates execution for weaker borrowers. The exact-document requirement and adverse ambiguity policy make late or vague disclosure a miss. This is materially less likely than the forecaster’s 0.34.

8%

superforecaster statistician · openai/gpt-5.6-sol

I use 0.40, the ledger’s >90-day base rate, rather than the 0.708 capital-lens rate because that class mixes much easier outcomes. This proposition has several necessary gates: exceptional scale, AI-data-center purpose, explicit vendor guarantee, signature rather than an MOU, and qualifying disclosure within 116 days. Large RPOs and accelerating capex show financing demand; low broad credit spreads and VIX aid execution. Conversely, the 4.77 10-year yield and 10.51 single-B spread make giant risk-bearing guarantees harder, and the digest identifies no pending definitive agreement. Hard-date disclosure further reduces odds. Low-confidence calibration has only one observation, providing no credible uplift.

14%

geopolitics regulation analyst · openai/gpt-5.6-sol

The capital-spending and RPO figures show extraordinary AI infrastructure demand, but they do not evidence a qualifying financing agreement. A single $100B financing with an explicit vendor guarantee would be unusually large, require substantial underwriting and risk allocation, and likely generate conspicuous issuer disclosures. Only 116 days remain. Financing conditions are mixed: broad high-yield spreads are tight, but the 10-year Treasury yield is 4.77% and single-B spreads are elevated. Most importantly, the rule excludes MOUs and resolves ambiguity against a hit; vendor investment, purchase commitments, backstops, or several smaller facilities would not suffice without signed guarantee language in a definitive filing.

Evidence the engine used

direct · undated

SEC.gov | EDGAR Full Text Search

Cited by grid-and-power-economist: I am below the forecaster’s 0.34. Large and rapidly growing RPOs support extraordinary AI infrastructure demand, while low broad high-yield spreads and subdued volatility make financing feasible. But a single signed financing of at least $100B with explicit vendor-

direct · undated

CBOE Volatility Index: VIX (VIXCLS) | FRED | St. Louis Fed

Cited by grid-and-power-economist: I am below the forecaster’s 0.34. Large and rapidly growing RPOs support extraordinary AI infrastructure demand, while low broad high-yield spreads and subdued volatility make financing feasible. But a single signed financing of at least $100B with explicit vendor-