One number stopped. Another started, eighty-four metres away.

The GLOBAL SEAWAY series has spent seven parts on registry records — who owned a hull, who managed it, when a name changed. This piece starts somewhere else, with a vessel that stopped transmitting.

On 28 July 2026 at 11:00 UTC, a tanker at anchor in the Gulf of Suez broadcast its last position under MMSI 632001125, at 29.709854 N, 32.609753 E. On 31 July at 06:42 UTC, MMSI 273123320 began broadcasting from 29.709580 N, 32.608940 E.

Those two points are eighty-four metres apart — a fraction of the vessel's own length. Two days and twenty hours separate them.

The numbers say what the names don't

A ship's MMSI is not arbitrary. Its first three digits are a Maritime Identification Digit assigned by country, and they change when a vessel changes flag.

632 is Guinea. 273 is Russia.

So the hull that went quiet at that anchorage was broadcasting a Guinean identifier, and the hull that resumed from the same berth three days later was broadcasting a Russian one. The reflag is legible in the numbers themselves, before any registry is consulted.

What the registry adds, and where it is weaker

The vessel is IMO 9409467, and its Equasis name history runs:

Five names in eighteen years, the most recent arriving in the month after the broadcast identity changed.

Note what the registry cannot tell you here. These rows come from IHS Maritime, which records identity history to the month — every date renders as the first of a month, as all 614 IHS name rows held for this investigation do. "AMARANT since 01/08/2026" means August 2026, not 1 August.

The position record is the more precise source. It places the change between 28 and 31 July, at a named anchorage, to the hour. A registry search would have told you this vessel was renamed sometime in August. The transmissions tell you when it happened, and where the ship was standing.

The designation is filed against a name the hull no longer carries

The European Union designated this vessel on 25 February 2025, and the listing names it "Pravasi" — which was correct. Pravasi is the name it carried from February 2024 to August 2026. The United States has designated it as well.

Then, eighteen months after the EU listing, the name went.

This is the pattern Part 4 documented on CETUS, which became MERIDIAN while its EU designation stayed filed against IMO 9418482. The mechanism is identical and the consequence is the same: a screening list that reaches this hull by number does not reach it by name.

What is different here is the evidence. On CETUS the rename was visible only through the registry, and the registry gave it a month. On this hull the change was observed while it happened, from position data, at a specific berth.

Evidence and confidence

What this doesn't establish

Two MMSIs at one anchorage is not, by itself, one ship. Nothing in the position record proves the hull that stopped and the hull that started are the same vessel. What the record shows is that the second identity began transmitting eighty-four metres from where the first one ended, less than three days later, and that the registry places a name change on that hull in the same window. Those facts sit together; this piece does not claim they are one observation.

No fleet-mate, cargo, counterparty or voyage claim is made. This is a single-hull identity reconstruction.

The six-authority claim this project's own working record carries is not repeated here. That record states designations by six governments. Checked by exact IMO-field match against the datasets held locally, only the European Union and the United States are confirmed. The others may well exist; they are not evidenced in this piece and are not asserted.

Nor is the screening result asserted. An internal note records that a name-only sanctions search for "Amarant" returns nothing while an IMO-anchored search returns the full record. That test has not been re-run for this piece, and a claim about what a live sanctions database returns today should not rest on a check made weeks ago. It is stated as an open question below, not as a finding.

What would prove this wrong

• The 84-metre separation is computed from two single position reports, one on each MMSI. A bad fix on either would change the distance materially, and neither has been cross-checked against a second source.
• A three-day gap is not an absence of broadcast. The old MMSI may have transmitted after 28 July outside this project's reception; the new one may have transmitted before 31 July. Both dates are first- and last-observed, not first- and last-transmitted.
• MMSI-to-country mapping identifies the flag administration that issued the number, not necessarily where the vessel is registered today. A vessel can carry a stale MMSI after a flag change.
• The name history is Equasis's current record. A later correction to that record would not be visible here.
• This piece does not establish why. A rename and reflag are ordinary commercial events, and nothing here shows intent.

Method

Position data is this project's own AIS record, read for both MMSIs across the transition window and reported as first- and last-observed fixes with their timestamps and coordinates. Separation is computed by haversine.

MMSI-to-country resolution uses this project's canonical Maritime Identification Digit table, not an inferred mapping — a standing rule here, after an earlier incident in which guessed MID values produced wrong flags across a whole cohort.

Name history is the vessel's own Equasis record, treated as month-precision throughout. Designation status was checked against locally held European Union and United States datasets by exact IMO-field match rather than by name search — which, given that the listing names a hull this ship is no longer called, is the only method that would have found it.

Part 9

One thread carries forward, and it is the one this piece deliberately declines to assert: what a name-only search of a live sanctions database returns for "Amarant" today, against what an IMO-anchored search returns. If the first is empty and the second is not, that is this series' central claim demonstrated on a hull whose identity change was observed rather than inferred. It needs the search re-run, dated, and captured — not recalled.

Use

This reconstruction is free to use, but you must credit AIS Dark.

Want the same reconstruction run on a fleet or counterparty you are assessing? Get in touch.

As of 28 August 2026.