Satellite internet at sea: how to choose between LEO, VSAT and L-band

Choosing satellite connectivity for a ship has quietly got harder over the last three years. The question used to be single — “how much bandwidth, for how much money” — and the answer was a single VSAT contract. Low Earth orbit (LEO) services broke that equation: a step change on speed and price that barely admits comparison, but two new questions came with it. Coverage now depends on the country; safety communications is an entirely separate matter.

This guide treats connectivity not as one choice but as three separate layers — because most of the mistakes we see in the field come from trying to solve all three with one contract.

Scope note. This is an independent guide. It is not the publication of any service provider and contains no product recommendation. Coverage, licensing and tariff information changes quickly; confirm current status with your provider and flag state before deciding.

Layer 1 — Primary broadband: this is where the job changed

A conventional VSAT-equipped ship budgeted connectivity like fresh water: a shared, metered resource with operations first in the queue. A LEO-equipped ship effectively has the connection of a small shore office — enough capacity for simultaneous video calls, cloud applications, telemetry streams and software updates at the same time.

The market answered by turning connectivity from a fixed installation into a managed service. In March 2026 Marlink launched a multi-LEO service combining Starlink and Eutelsat OneWeb capacity in a single data package, with tiers running from 500 GB to 10 TB and dynamic switching between networks based on performance and coverage.

The real decision here is not which brand to pick: are you tying yourself to a single network, or delegating network selection to a provider? A ship tied to one network has also inherited that network’s coverage and licensing map.

Layer 2 — Failover: the question to ask before signing

Hybrid architectures — blending LEO, GEO VSAT, LTE and L-band under a single policy layer — are becoming standard equipment among serious operators. The reasoning is simple: no single network is assumed reliable enough to carry safety-related traffic on its own.

The common arrangement works like this:

Layer	Role	Typical technology
Primary	High-bandwidth traffic	LEO (broadband)
Secondary	Automatic handover when primary degrades	GEO VSAT or LTE / shore network
Safety	Distress and safety communications	Recognised GMDSS system (L-band)

The critical detail is that the handover is driven by policy, not by hand. “We have a backup” is not enough; which traffic uses which link, what keeps moving when the primary fails and what waits, all have to be defined in advance. Without that definition the backup link exists on paper and congests at the moment of need.

So the pre-contract question is exactly this: is the handover automatic, how many seconds does it take, and which applications stay up during it?

Layer 3 — Safety communications: no choice here

This is the clearest section of the guide and the one most often skipped.

Starlink is not recognised by the IMO as a GMDSS provider. In low Earth orbit, the network formally recognised for GMDSS distress and safety communications is Iridium — the LEO provider holding that certification since 2020. Its constellation of 66 active satellites operating on L-band frequencies provides global coverage including both poles.

So whatever you do on the broadband side, the safety layer is built on a separate, recognised system. LEO broadband does not replace that layer; it runs alongside it. This should be read not as a cost line but as a scope condition.

Why coverage varies by country

The common assumption is that a satellite service works everywhere by its nature. For LEO broadband that is not the case: the service requires separate telecommunications landing rights in each country’s territorial waters and exclusive economic zone. Coverage can therefore differ between the high seas and territorial waters, and from country to country.

The practical consequence: a ship trading predominantly on the high seas and a ship calling frequently and spending much of its time in territorial waters should not make the same connectivity decision. For the second group the failover layer is not a comfort, it is an operating condition.

Before signing, get coverage confirmation for your actual trading area from the provider — a general marketing map is not adequate for this job.

The connectivity decision is also a cyber decision

As the pipe grows, so does the surface. An always-connected ship becomes a network in the same class as a shore office: crew traffic, operational traffic and supplier remote access start flowing through the same link.

That is why the bandwidth decision and the network segmentation decision belong on the same table. That crew-welfare traffic should not sit in the same network segment as navigation and machinery systems is no longer a contested view. On the newbuild side that expectation has entered class rules — for the regulatory side of this: Ship cyber security: which rules actually apply to your vessel?

Decision order

1. Map the trading pattern. High seas versus territorial waters, which countries are called. The coverage question follows from this.
2. Budget the three layers separately. Primary broadband, secondary failover, safety layer. A setup that believes one contract covers all three usually leaves the safety layer short.
3. Ask for the handover policy in writing. Automatic or not, in how long, which traffic survives.
4. Decide segmentation at the same time. Segmentation added after the connectivity contract is signed is always more expensive and less complete.
5. Tie coverage confirmation to the contract. Not a marketing map — current written confirmation for your trading area.

How connectivity is changing ship operations — crew welfare, remote support, operations moving ashore — is a separate subject: LEO connectivity is quietly rewiring ship operations.