Skip to content

AI Market · Forecast desk · Live · software

25%

At least one Fortune 500 enterprise discloses a production deployment greater than $10M annualized of a specialist video / embodied multimodal model (Perceptron Mk1, MolmoAct 2 successors, or analogous) as a primary tier rather than a general LMM.

Probability history

0%25%50%75%100%05-1709-06deadline

Called at 60% on May 17, 2026. Repriced 1 time; now 25%. 24 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)60%
Reference class (n=29)50%
Recalibrated53%
Feature model18%
Council17%
Final25%

Drivers

  • Reference class qtype:disclosure runs 0.50 (n=29) vs stated 0.60.
  • Hierarchical calibration moves stated 0.60 to 0.53 (lens software, kappa_stated=24.0, kappa_lens=200.0).
  • Feature model at 0.18: conjuncts=1, named disclosure=True, horizon=136d.
  • Council of 5 at 0.17.

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.

18%

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

Only 24 days remain, and the supplied digest contains no qualifying named-customer disclosure. The rule is unusually strict: a Fortune 500 customer, explicit production status, specialist multimodality, and a disclosed dollar or capacity figure proving more than $25M annualized. Enterprise AI deployments are increasingly common, but customers and vendors rarely publish contract economics, while announcements often describe pilots or omit measurable scale. September also offers fewer routine earnings disclosures than quarter-end reporting months. The software and disclosure base rates are materially higher, but they do not reflect this late-stage conditional window or the conjunction of documentary requirements. Ambiguity resolving against the forecaster further lowers the chance of a marginal announcement qualifying.

16%

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

Only 24 days remain, and the digest contains no qualifying named-customer disclosure. Rapid model releases and hyperscaler commitments support growing multimodal adoption, but they do not satisfy the strict resolution test. Enterprise announcements commonly describe pilots, production workflows, users, or broad vendor commitments while withholding annualized deployment value. Requiring a Fortune 500 name, explicit production status, specialist-multimodal scope, and a figure above $25M creates substantial conjunction and ambiguity risk. The software-ledger base rate of 0.452 is therefore less relevant after most of the forecast window has elapsed without a documented hit. This is materially below the May forecast of 0.6.

18%

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

The underlying adoption trend is strong, but the disclosure hurdle is unusually strict: a named Fortune 500 customer, explicit production status, specialist multimodality, and a dollar or capacity figure clearly exceeding $25M annualized. Enterprises and vendors rarely publish deployment economics at that granularity. Only 24 days remain, and most calendar-Q3 10-Qs will arrive after September 30. Rapid hyperscaler capex and RPO growth support large deployments generally but do not establish this specific customer-level trigger. Ambiguity resolving against the forecast further lowers the odds versus the software-disclosure base rate.

14%

superforecaster statistician · openai/gpt-5.6-sol

Base rate: 0.452 for software questions, with disclosure questions at 0.500 and original >90-day questions at 0.400. I reduce sharply because only 24 days remain and the digest provides no qualifying named-customer evidence. Resolution requires one disclosure to identify a Fortune 500 customer, state production rather than pilot status, establish specialist multimodality, and quantify over $25M annualized or qualifying capacity. Such customer announcements commonly omit spending and use ambiguous partnership language, which resolves as failure. Broad “at least one” scope and strong AI investment help, but model cadence and vendor capex do not verify customer-level economics. The forecaster’s 0.60–0.69 band realized 0.556, supporting a modest overconfidence correction.

17%

geopolitics regulation analyst · openai/gpt-5.6-sol

The broad software and disclosure base rates near 0.45–0.50 materially overstate this unusually strict event. Resolution simultaneously requires a Fortune 500 name, explicit production status, specialist multimodality, and a disclosed dollar or capacity figure establishing more than $25M annualized. Enterprise AI growth and rapid model releases support the underlying deployment trend, but companies rarely publish customer-level economics. Only 24 days remain, with limited major reporting or conference catalysts before quarter-end. The supplied digest contains no qualifying named-customer evidence. Ambiguity resolving against the forecast further lowers the chance below the forecaster’s 0.6.