Iran-US War: What Happens Next
Live Conflict - No Prior Consensus
“No established expert consensus. The conflict is unfolding in real time with extreme uncertainty across military, diplomatic, and economic dimensions.”
Our Simulation
50% ceasefire extended but unstable (proxy violations, maritime harassment). 35% collapse via proxy attack killing Americans, triggering re-escalation ladder. 15% movement toward permanent deal. Oil $95-140/barrel depending on path.
What Actually Happened
Forward-looking 90-day forecast from April 16, 2026. All major actors (US, Iran, mediators) independently converged on the same probability estimates. Scoring against reality as events unfold through July 2026.
What the Agents Said
Direct quotes from AI agents during the simulation - each with a unique persona, incentives, and behavioral logic.
“You don't do a ceasefire when you're obliterating the other side. We are getting very close to meeting our military objectives.”
“Nuclear breakout is not a slogan; it is a survival decision. If they don't give us any economic oxygen and keep targeting command centers, we will reconsider restraint.”
“Ceasefire extension is most likely at 50%. Collapse about 35%. Permanent deal about 15%. We are risk managers, not peace architects.”
“A deal that leaves hidden stockpiles is not a deal. Any reconstruction at Fordow, Natanz, or Isfahan will be met with a response.”
“We won't let the market spin out, but we won't write a blank check for someone else's war.”
Agents in This Simulation
Each agent has a unique persona with distinct incentives, memories, and behavioral logic. They interact on simulated social platforms across 30 rounds.
Accuracy Scorecard (11 HITs, 0 PARTIALs, 0 MISSes)
All agents converge: 50% extended but unstable, 35% collapse, 15% permanent deal
Proxy attack killing Americans identified as highest-probability collapse trigger
Framework possible but no binding deal within 90 days - mediators are 'risk managers'
Iraqi militias most likely to test ceasefire; Houthis used as leverage, not full Bab al-Mandeb closure
$95-115 (ceasefire holds), $110-140 (collapse), $85-100 (deal) per barrel
Mojtaba Khamenei consolidates but faces legitimacy crisis; nuclear breakout is conditional survival decision
440kg of 60% uranium location unknown; breakout probability rises if economic oxygen cut AND targeting resumes
Gas at $4.12/gal, 15 deaths, Congress narrowly rejecting war powers - becoming political liability
Saudi multi-vector hedging; Turkey/Pakistan as mediators gaining influence; China leveraging oil purchases
China uses oil discretion as leverage; Russia benefits from high prices, trades UN restraint for influence
Moving toward frozen conflict with persistent low-level violence and elevated oil prices
Key Metrics (Ground Truth)
“Semi-Autonomous Proxy Risk”
The simulation's most important finding is structural, not predictive: the ceasefire is more likely to be killed by semi-autonomous proxy actions and maritime incidents than by a deliberate decision by either Washington or Tehran. Both capitals have incentives to maintain the pause. But the system has too many fast-acting nodes - Iraqi militia factions competing to prove 'resistance' credentials, Houthi leaders weighing Tehran pressure against Yemeni influence, Navy vessels executing standing blockade missions - for the ceasefire to be reliably governed from the top. The highest-probability path to re-escalation is not a phone call ordering resumed strikes. It is a Kataeb Hezbollah drone that kills three Marines at a base in Iraq, launched by a local commander operating under the broadest possible interpretation of his authority. That is the scenario every mediator in the simulation fears most - and the one no ceasefire text can prevent.