197 Shell Hops, 425 MMSI Changes, 383 Renames: The 1,005 Identity Mutation Events in 33 Days

The Setup

The vessel identity graph holds 23,586 temporal edges across the active fleet. Most are routine: ownership assignments, ship manager links, flag registrations. Of the 23,586 edges recorded between April 2 and May 5, 2026, exactly 1,005 — 4.3% — are identity mutation events. A vessel changed its MMSI, changed its name, or passed through a shell structure.

Of those 1,005, 197 are shell_hop designations. The most recent was recorded today.

The Chain

Temporal edges record how vessels relate to corporate entities and to their own prior identities over time. The full distribution across the 33-day tracking window:

Stable identity edges:

• SHIP_MANAGER: 7,651 (32.5% of all edges)
• OWNED_BY: 7,649 (32.4%)
• flagged_as: 7,276 (30.8%)

Identity mutation edges:

• mmsi_changed: 425 (1.8%)
• renamed_to: 383 (1.6%)
• shell_hop: 197 (0.8%)

MMSI changes and renames are administrative events. Sales, reflaggings, and operator transitions all generate them — not every MMSI change signals evasion. Shell hops are different in kind, not just degree. A shell_hop edge is assigned when a vessel's beneficial ownership traces through an intermediate entity displaying structural obfuscation characteristics: short corporate lifespan, single-vessel operation, jurisdiction mismatch between registration and operator, or a combination of those markers.

For every shell_hop in the 33-day window, there are 2.16 MMSI changes and 1.94 renames. The ratio matters: a fleet evading at the structural level tends to also cycle identifiers, but the correlation is not one-to-one. Some MMSI changes are clean transactions; some shell hops exist without any corresponding identifier change in the same window.

The Implication

197 shell_hop edges in 33 days is not a negligible count. Against a fleet of roughly 45,889 tracked vessels, it represents a meaningful fraction of ownership structures flagged as obfuscation-pattern-consistent.

Shell companies are not inherently sanctions violations — they are also the standard instrument in conventional ship financing through single-purpose vehicles. But the OFAC playbook for continued trading after designation follows a specific sequence: rename, reflag, transfer beneficial ownership to a newly incorporated entity, and repeat. The temporal edge graph is designed to track exactly that sequence.

The 425 MMSI changes deserve separate attention. MMSI is the vessel's primary radio identifier in the AIS system. A change means the vessel appears, for some window, as an unknown new entity to screening systems that don't cross-reference identity history. Combined with the motif detection system showing 1,925 AIS manipulation alerts in the overlapping period, MMSI cycling represents a compounding risk: structural obfuscation plus identifier rotation plus active AIS manipulation in the same fleet population.

What to Watch

Monitor the shell_hop-to-MMSI ratio over rolling 30-day windows. The current ratio sits at 1:2.16 (shell_hops to MMSI changes). If shell hops begin closing on the MMSI change count — ratio approaching 1:1 — it suggests obfuscation is escalating at the structural layer, not just the identifier layer. That is a materially different risk posture.

Renames at 383 in 33 days are a baseline, not an alarm. The alert condition is renames that co-occur with MMSI changes on the same entity within a tight window — compound indicator, not single signal.

Limitations

Shell_hop classification is a heuristic designation derived from entity graph characteristics, not a legal determination. A single-purpose vehicle used in standard shipping finance can satisfy the same structural criteria as one used for evasion. The edge label reflects pattern match, not confirmed intent.

These counts cover the 33-day window for which temporal edges have been recorded. Vessels with no recent identity changes don't appear in mutation counts — the denominator for "what fraction of the total fleet is mutating identities" is not directly readable from this table alone, since stable vessels produce only their background relationship edges.

Sample: 23,586 temporal edges from April 2 – May 5, 2026. Underlying vessel table: 45,889 vessels.

Data as of 2026-05-05. Source: Axiom Overwatch temporal_edges table.