Baltic Energy Transit Risk

How exposed the new Baltic gas import terminals are

Tobyn Smith

The question

Before 2022, most of the Baltic region’s gas came by pipeline from Russia. After the invasion of Ukraine and the Nord Stream sabotage, that route was shut. What replaced it is five working coastal terminals, plus a sixth planned at Gdańsk.

How concentrated is that infrastructure, and how exposed is it?

The size of the hole

Russian pipeline gas into these eight countries:

  • 2021: about 70 bcm
  • 2023: about 0.3 bcm, essentially nothing

Operating LNG capacity built in the region: around 23 bcm, roughly a third of what stopped arriving. Lower demand closed some of the rest. Pipeline gas from the west and north closed the remainder.

The terminals

Hover for a name, click for the figures. Each point is sized by capacity.

What I am measuring

An exposure score for each terminal, from three ingredients:

  • how close it sits to Russian territory,
  • how much of the region’s import capacity it holds,
  • how long its supply route past the Danish Straits runs.

I picked the weights myself. Anyone can change them on the site.

Finding 1: concentration

Import capacity is lopsided.

  • Two southern terminals, Świnoujście (Poland) and Mukran (Germany), hold most of the region’s capacity.
  • The northeast runs on one or two terminals each: Inkoo for the Gulf of Finland, Klaipėda for Lithuania.

Finding 2: proximity

The terminals closest to Russia are the small ones:

  • Hamina, at the eastern end of the Gulf of Finland,
  • Klaipėda and Gdańsk, beside the Kaliningrad exclave.

The two largest terminals sit farthest from Russia. I did not rank on size alone for that reason.

Finding 3: one way in

Every LNG cargo passes the Danish Straits. The Kiel Canal is the only other entrance, and it is far too small for LNG carriers. The Straits are the door that actually matters.

Terminals deep in the Gulf of Finland have the longest run once they are inside the sea.

Disruptions since 2022

Date Asset What happened
Sep 2022 Nord Stream 1 & 2 Pipelines sabotaged off Bornholm
Oct 2023 Balticconnector Pipeline and cable damaged by an anchor; repaired 2024
Dec 2024 Estlink 2 Power and telecoms cables cut; back in service 2025

Baltic energy and data links have been hit repeatedly, mostly by ships dragging anchors across the seabed. I left those events as context.

Who leans on whom

Not every country has its own terminal.

  • Estonia, Latvia, Sweden: no terminal; they depend on neighbours through the Balticconnector and regional pipelines.
  • Lithuania, Germany, Poland: a single terminal each, so one outage takes out their seaborne supply.
  • Denmark: low dependency, on the back of its own North Sea gas.

Is the capacity enough?

Capacity only means something next to demand.

  • Finland and Lithuania hold several times their own annual consumption. They built a regional buffer that neighbours can draw on.
  • Germany and Poland cover only part of much larger demand by sea, and still lean on pipelines from the west and north.
  • Estonia, Latvia, Sweden have no terminal at all.

The terminals sitting closest to Russia are also the ones carrying the region’s spare capacity. I would keep those two facts together.

Lose one terminal

If you apply the EU’s N-1 idea to seaborne capacity alone:

  • Finland falls from over three times its demand to about a fifth, on Hamina alone.
  • Lithuania, Poland and Germany fall to zero seaborne capacity.

They are not failing the real N-1 test, which also counts pipelines and storage. The fallback is the same pipelines and cables that have been cut.

What I would take from this

  • Finland and Lithuania each lean on one terminal; the fallback is the pipelines that have already been hit.
  • Estonia, Latvia and Sweden import through neighbours, not terminals of their own.
  • One entrance through the Straits serves all eight countries, so patrols there cover everyone at once.
  • FSRUs were quick to add and can be moved, but they are few and chartered.

Limitations

  • This measures geographic exposure, not the probability of disruption.
  • The score uses proxies (distance, capacity share), not a causal model.
  • It is a single snapshot: capacity, routes and terminals change.

I would not cite this as a forecast.

The full project

The maps, the data, and the exposure ranking, all built from public sources:

tobynsmith.me/baltic

Code and open issues on GitHub.