LOW IMPACT[80% scenario]Authored AI-hypothesis estimate of whether this top-level scenario materialises over the stated horizon. Anchored to historical base rates where applicable. Not a market-calibrated forecast. See /scenarios/methodology.→Efficiency breakthrough [15%]Conditional probability: GIVEN the parent scenario materialises, the share of that probability mass we estimate falls in this branch. Sums to 100% across siblings. See /scenarios/methodology.→[25% path]Conditional probability: GIVEN the parent scenario materialises, the share of that probability mass we estimate falls in this branch. Sums to 100% across siblings. See /scenarios/methodology.
New frontier model race consumes savings
Efficiency gains immediately redeployed into larger and more capable frontier models; Jevons paradox plays out; total compute demand stays high.
Trade lens —NVDA and ASML demand sustained on Jevons-paradox dynamics; TSM advanced-node demand holds; challengers still trail in frontier training. · small move · fast
Policy lens —The Commerce Department maintains strict export controls on H20-class and above GPUs as frontier training demand sustains; the White House AI Council reviews export-licence backlogs and issues strategic guidance on allied-country access; OECD updates frontier-AI governance recommendations to address Jevons-paradox compute scaling.
⚡ 288 supporting signals (7d)
Signal counts measure media attention over the last 7 days — not the likelihood of an outcome.
Scenario oddsAuthored AI-hypothesis estimate of whether this top-level scenario materialises over the stated horizon. Anchored to historical base rates where applicable. Not a market-calibrated forecast. See /scenarios/methodology.
80%
Generative AI
Y-path oddsJoint probability of this entire path = product of every step. Rounded to nearest 5%; "<5%" shown rather than 0% because rounding to zero overclaims certainty. See /scenarios/methodology.
~10%
Efficiency breakthrough
This pathJoint probability of this entire path = product of every step. Rounded to nearest 5%; "<5%" shown rather than 0% because rounding to zero overclaims certainty. See /scenarios/methodology.