JCPOA — Iran Nuclear Deal
Snapshot · March 2026 · not live22 promises, 11 agents, 6 domains — Promise network analysis of the most complex arms control agreement in history
Start here: one promise, four steps
The 2015 Iran nuclear deal (the JCPOA), broken into 22 promises: who committed to what, to whom, and whether it was kept. This is a snapshot from March 2026 and is not updated live.
Step 1: One promise
Everything on this page is built from promises like this one.
- Who promised
- United States
- Promised to
- Iran
- The promise
- “Lift nuclear-related secondary sanctions and allow Iran access to global financial system”
- How it's checked
- Official filing, by U.S. Treasury OFAC sanctions list
- Status
- BrokenU.S. withdrew from JCPOA on May 8, 2018 and reimposed all nuclear-related sanctions plus new maximum pressure sanctions by November 2018.
Step 2: Promises lean on each other
6 other promises depend on this one:
- EU/E3 (UK, France, Germany): Lift EU nuclear-related sanctions including oil embargo and financial restrictions (slipping)
- United States: Refrain from imposing new nuclear-related sanctions or re-introducing lifted sanctions (broken)
- EU/E3 (UK, France, Germany): Support normalization of trade and economic relations with Iran (slipping)
- and 3 more.
Step 3: It did break
This one broke. Of the 6 promises that depend on it, all are now slipping or broken. The map cannot prove that one caused the others, but it shows where to look.
In the What-if you can pick any promise and see what else a change would reach.
Step 4: What we found
The deal came apart on both sides. Of its 22 promises, 14 are broken and 6 slipping; 1 is confirmed kept. Iran broke 10 of its 12 promises, and the United States all 3 of its own.
The rest of this page is the detail on all 22 promises.
Where the 22 promises stand
Overall
: 14 of 100
The average of every promise's status (confirmed kept 100, no verdict yet 60, slipping 30, no way to check 20, broken 0).
78 of 100
How settled these statuses are. It is low where nobody has checked.
Health by area
Promise score and certainty for each area, out of 100.
What to check first
Where a fresh check would tell us the most, in order.
Joint Commission: Provide dispute resolution mechanism through Joint Commission with escalation to UNSC
No one is assigned to check it.
No way to checkJCPOA-018
United States: Lift nuclear-related secondary sanctions and allow Iran access to global financial system
6 other promises depend on this one.
BrokenJCPOA-011
UN Security Council: Replace prior UN Security Council resolutions with UNSCR 2231 framework — phase out arms/missile restrictions
Checking it now adds the most certainty.
BrokenJCPOA-013
Iran: Resolve all outstanding issues on Possible Military Dimensions (PMD) of nuclear program
3 other promises depend on this one.
SlippingJCPOA-010
United States: Refrain from imposing new nuclear-related sanctions or re-introducing lifted sanctions
No one is assigned to check it.
BrokenJCPOA-014
Technical detailsModel diagnostics, k values, and cycle projections behind the summary above.
Network Health (same number as the promise score)
14 out of 100
Verification Dynamics
Bayesian certainty — how confident we are in the health assessment.
Dynamical regime distribution
Verification Priorities
?Promises where verifying NOW has the highest marginal impact on network confidence.
- JCPOA-01874%Composting risk
Joint Commission: Provide dispute resolution mechanism through Joint Commission with escalation to UNSC
Composting risk (k=0.34) — approaching verification window closure. 3 review periods without assessment.
Resolution expected by cycle 3Optimal review: every 3 cycles (reviewed more often than it can change)Long-term: resolution trending · P(met) at cycle 10: 76% · P(not met): 11% - JCPOA-01159%Bottleneck
United States: Lift nuclear-related secondary sanctions and allow Iran access to global financial system
Structural bottleneck: 6 downstream promises depend on this. Under-verified for its network importance.
- JCPOA-01352%Composting risk
UN Security Council: Replace prior UN Security Council resolutions with UNSCR 2231 framework — phase out arms/missile restrictions
Composting risk (k=0.43) — approaching verification window closure.
- JCPOA-01050%Bottleneck
Iran: Resolve all outstanding issues on Possible Military Dimensions (PMD) of nuclear program
Structural bottleneck: 3 downstream promises depend on this. Currently well-verified — confirm it's holding.
Resolution expected by cycle 5Optimal review: every 3 cyclesLong-term: resolution trending · P(met) at cycle 10: 72% · P(not met): 2% - JCPOA-01449%Composting risk
United States: Refrain from imposing new nuclear-related sanctions or re-introducing lifted sanctions
Composting risk (k=0.34) — approaching verification window closure.
High-Leverage Promises
Ranked by how much else would slip if each promise broke, in the model's what-if engine. Ranks only.
Reliabilities are default estimates, not measured.
Structural Diagnostic
Five-field analysis: epidemiology · FMEA · information theory · incentive alignment. Each applies a published model's form to this network's structure; no coefficient was fitted to promise data.
Cascade Risk
A single violation is unlikely to cascade beyond its direct dependents.
R₀ = 0.68 (network) · R₀ hubs = 2.50
Watch list: slipping or weakly verified
Answers a different question from High-Leverage Promises: which promises are currently at risk and least checked, not which hold the plan up.
Ranked by Risk Priority Number (RPN = Severity × Occurrence × Detection). Network reliability: 11.3%.
Severity 10 · Occurrence 10 · Detection 5
High severity: failure cascades to 6 downstream promises across 3 domains. High likelihood of failure based on current status (violated). Moderately detectable: verification is filing.
Severity 10 · Occurrence 7 · Detection 3
High severity: failure cascades to 5 downstream promises across 2 domains. High likelihood of failure based on current status (degraded). Easily detected: verification is audit.
Severity 6 · Occurrence 10 · Detection 3
Moderate severity: failure cascades to 2 downstream promises across 2 domains. High likelihood of failure based on current status (violated). Easily detected: verification is audit.
Severity 5 · Occurrence 7 · Detection 5
Moderate severity: failure cascades to 2 downstream promises across 1 domain. High likelihood of failure based on current status (degraded). Moderately detectable: verification is filing.
Severity 1 · Occurrence 10 · Detection 10
Low severity: failure cascades to 0 downstream promises across 0 domains. High likelihood of failure based on current status (violated). Hard to detect: verification is none.
Verification Infrastructure
34% of this network's state is unobservable.
Incentive Alignment
Highest Agency Cost:
United States: Lift nuclear-related secondary sanctions and ...
United States: Refrain from imposing new nuclear-related san...
Empirical Diagnostics
Each figure carries its evidence class: fitted on data, an estimate, or a rule we wrote.
Network Fragility
Which promises the rest depend on most: see High-Leverage Promises above, ranked by what would slip if each one broke.
Stranded Promises (1)
1 promise has no pathway to resolution:
Recommendation: add verification infrastructure or structural dependencies
Cascade Calibration
Outcome projection
Outcome projection: Using a three-state transition model — fitted on 11,994 MONA conditions for the composting and transitional regimes; the computing and pressure parameters are estimates not yet validated on a cohort — 3 promises in this network are in composting dynamics (a promise left unchecked gets harder to resolve) and 1 is in computing dynamics (observation-driven resolution) . 3 promises are at risk of over-review — they are being reviewed more often than they can change .
1 promise has no structural pathway to resolution
No dependencies, no verification. Adding a dependency connection or verification mechanism would move it out of stasis.
JCPOA-018
Want your commitments mapped like this?
We build interactive promise graphs for organizations, advocates, and policy teams.