Armenian external access — two borders open, two closed
A landlocked country whose shortest routes to the world have been shut for thirty years, and what that does to everything else.
Armenia's four land borders reduce to two usable ones: Georgia and Iran. Türkiye and Azerbaijan have been closed since the 1990s, which routes almost all trade through a single Georgian corridor and makes border reopening the largest economic variable in the module.
Overview
Armenia is landlocked and, in practice, half-enclosed. The border with Türkiye has been closed since 1993 and the border with Azerbaijan since the first Karabakh war, leaving Georgia and Iran as the only routes. The overwhelming majority of trade and effectively all of the northward connection to Russia and the Black Sea runs through Georgia, converging on the Upper Lars crossing on the Russian frontier — a single high-altitude route that closes for weather and has been closed for politics. The Iranian route provides a genuine second option but faces sanctions constraints and difficult mountain geography. The consequence is that Armenia pays more for everything, has limited redundancy against any single disruption, and holds an economic interest in normalisation that is structural rather than sentimental. The Lachin corridor, which connected Armenia to Karabakh between 2020 and 2023, is recorded here as a historical system: it does not exist now, and the events around its closure are covered in the humanitarian record.
Why it matters
This is the material basis of Armenian foreign policy. The reason border reopening appears in every negotiation, the reason the corridor question is existential rather than technical, and the reason Armenia's Russian and Georgian dependencies are so hard to reduce all resolve into this system. It also explains the asymmetry in the settlement negotiations: Azerbaijan has routes and alternatives, Armenia has two roads.
What depends on it
Near-total and structurally narrow. Georgia carries the large majority of Armenian trade and is the only route to Russia and the Black Sea; Iran is a real but constrained alternative. There is no maritime access, no route that avoids both, and no domestic substitute for imported energy and most goods. Redundancy exists only in the sense that two routes are better than one, and the northern route funnels through a single crossing.
What disruption does
The system's narrowness raises costs across the whole Armenian economy, concentrates risk on a single foreign transit state, and limits the country's ability to diversify away from Russia — because the route to anywhere else runs through Georgia too. It is also why border reopening is treated by Armenian governments as an economic and security question simultaneously: the same act would reduce cost, add redundancy and change the regional relationship in one move.
Nodes
Each node is a map marker this module already publishes, with its role in the system. Not a route: the list has no sequence and no geometry.
- Upper Lars crossing Border crossing · Border crossing
The only functioning crossing between Georgia and Russia, and therefore the bottleneck of Armenia's northern trade. Closes for weather and has closed for politics.
- Tbilisi Transit node · Capital
Georgia is the transit state for the large majority of Armenian trade.
- Poti Entry port · Port
Black Sea outlet used for Armenian containerised trade.
- Batumi Entry port · Port
- Yerevan Regional hub · Capital
- Gyumri Transit node · Major city
Near the closed Turkish border; the rail line through it has been out of service since 1993.
- Armenia–Türkiye border Chokepoint · Border zone · broad region
CLOSED since 1993. Recorded as the system's principal constraint rather than as a route.
- Syunik province Transit node · Border zone · broad region
The southern province carrying the Iranian route, and the territory any proposed Azerbaijani connection to Nakhchivan would cross.
- Lachin corridor (legacy) Alternative corridor · Corridor · supply & trade artery
HISTORICAL. Operated between 2020 and 2023 connecting Armenia to Karabakh under the trilateral arrangement; blockaded from December 2022 and superseded by the September 2023 events. It does not exist as a system now and is recorded for completeness, not as current capacity.
Alternatives and redundancy
Two, and both constrained. The Georgian route is the main artery and its northern extension depends on a single crossing at Upper Lars. The Iranian route is genuinely usable and is Armenia's hedge against Georgian disruption, but sanctions limit what can move along it commercially and the mountain road is difficult. Air freight covers high-value goods only. **The two closed borders are the alternatives that do not exist**: opening either — Türkiye in particular — would transform the system, which is precisely why it appears in the diplomacy. Proposed routes connecting Azerbaijan to Nakhchivan across Armenian territory are negotiation subjects, not infrastructure, and are not capacity.
Documented disruption history
Closure of the Turkish border, and of the Azerbaijani border earlier in the decade, removing Armenia's shortest western and eastern routes. Both remain closed and the rail line through Gyumri has been out of service throughout.
The Russia–Georgia war interrupted the northern route through Georgia, demonstrating that Armenia's principal artery depends on the stability of a country it does not control.
HISTORICAL. The Lachin corridor was blockaded and then closed, cutting Karabakh's connection to Armenia. Recorded here as a disruption to a system that no longer exists; the civilian consequences are covered in the humanitarian impact chain.
Seasonal and political closures of the Upper Lars crossing, which periodically sever the northern route entirely for days at a time.
Political implications
Gives Türkiye and Azerbaijan a standing instrument they do not have to use, and gives Armenia a structural incentive toward normalisation independent of any political preference.
Economic implications
Raises the delivered cost of imports and exports, constrains the scale Armenian producers can reach, and makes the closed borders a permanent growth constraint rather than a temporary one.
Humanitarian implications
Border communities in Syunik and the north bear the direct effects of closure and of any crossing disruption.
Actors with a stake
Campaign relevance
What remains unknown
- Armenian trade-route shares are reported on differing bases and informal cross-border trade is not measured.
- Crossing procedures, capacities and infrastructure detail are deliberately not published by this module.
- The commercial terms of Iranian trade arrangements are not public and include components that sanctions reporting does not capture.
- Cost estimates for the closed borders' economic effect vary widely between published studies and depend heavily on assumptions.
Readings the evidence also supports
- The persistence of the closed borders can be attributed to unresolved conflict, to Turkish domestic politics on the historical file, or to Azerbaijani sequencing that Ankara accepts. All three are argued and the record supports each in part.
- Armenia's continued Russian economic dependence can be read as a political choice or as the arithmetic consequence of having two open borders, one of which leads to Russia. The second reading explains more of the trade data than the first.
Listed because the record's own assessment is not the only one its sources permit — not as a hedge on the assessment above.
What to watch
Any movement on the Armenia–Türkiye border
Announced as under discussion repeatedly since 2021 and not done. Opening would be the largest single economic change available to Armenia.
Upper Lars closures and their duration
The northern route's single point of failure. Duration matters more than frequency.
Iranian route volumes
Armenia's hedge. Growth would indicate real diversification; stagnation indicates the sanctions constraint is binding.
Whether transport provisions of any settlement are implemented
Agreement and infrastructure are different things. Construction and operation are the observable measures.