Baltic air policing became air defence. The layers beneath it are still arriving

The finding

For twenty-two years the most visible piece of NATO military presence in the Baltic states was a pair of fighters on alert. Estonia, Latvia and Lithuania joined the Alliance in 2004 without fighter aircraft of their own, and Allies have taken turns, four months at a time, to keep jets ready at Šiauliai in Lithuania — later also at Ämari in Estonia, and since 2024 with Lielvārde in Latvia able to host them. The mission was called Baltic Air Policing, and what it did is in the name: meet an aircraft that approaches or enters allied airspace, identify it, and escort it out.

On 1 August 2026 it stopped being called that. Agreed at NATO’s summit in Ankara on 8 July, an air defence mission replaced Baltic Air Policing. The first fighters to fly it were the same kind of rotation that had flown the old one — Turkish F-16s at Ämari, Italian Eurofighters at Šiauliai — from the same bases, on the same four-month cycle.

So what actually changed? NATO’s answer is that the mission’s purpose widened: from threats posed by aircraft to a full spectrum of crewed and uncrewed aircraft and cruise missiles, with fighters joined by ground- and sea-based air defence and counter-drone systems. This Note tests how much of that wider system exists on the flank, and where the public evidence of it stops. The short answer is that the mandate moved faster than the layers beneath it. The mission has been redefined; the systems that would make it more than fighters are mostly national, arriving on their own schedules, and NATO does not say which of them serve it.

Air policing was one layer of a larger system

Air policing was never “nothing but jets”. NATO has run it since 1961 as part of its integrated air and missile defence, and the jets were the visible end of a chain. National radars and control centres in each Baltic state, NATO’s airborne early-warning aircraft and allied sensors feed one recognised air picture. NATO describes its integrated air and missile defence system as “a network of interconnected national and NATO systems”, operating under the authority of the Supreme Allied Commander Europe. Allied Air Command at Ramstein provides command and control around the clock from three combined air operations centres, and when an interception is needed the relevant centre decides which base launches.

NATO’s own policy, adopted in February 2025, defined the task narrowly: “the continuous presence of fighter aircraft and crews which are ready to react quickly”, to preserve “the integrity of the NATO Airspace in peacetime”. That is what the mission was built for — an aircraft that strays, probes or flies without a transponder or flight plan. It is identified, followed and, where necessary, escorted. It was not built to stop a drone swarm or a cruise missile, and NATO now says so: on its rewritten page, Air Policing “was primarily oriented to protect against threats from aircraft through interception by Allied fighter jets”.

The figure shows the structure that makes air policing an allied rather than a national activity, and the boundary that matters for everything that follows. The fighters fly under a NATO chain that Vigil’s records draw. The ground-based systems that would add depth sit in national armed forces, and the records hold no published command relationship between them and that chain. That is not evidence that none exists — NATO’s policy says national authority to defend Allies is assigned to NATO on a permanent basis — but it is the point at which the public evidence thins.

Why drones and cruise missiles change the requirement

The threat the Baltic mission was designed for is few, fast, large and generally identifiable. The threat that arrived in September 2025 was different. On the night of 9–10 September, Poland’s prime minister told parliament, Polish airspace was violated nineteen times by drones over about seven hours; three were confirmed shot down. NATO listed the response: Polish F-16s, Dutch F-35s, an Italian airborne early-warning aircraft, a NATO tanker and German Patriot systems. It worked, in the sense that some drones were downed. It also exposed the problem NATO’s commanders have described ever since.

NATO’s policy defines the threat spectrum as running “from small, low- and slow-flying UASs to all types of cruise and ballistic missiles”. Each end stresses a different part of the system. Small, slow, low-flying objects are hard to detect against the ground and harder to identify: Estonia and Latvia said the drones shot down over their territory in May and June 2026 had entered under Russian electronic warfare, and both were reported to be Ukrainian drones that had gone astray — which makes “is it hostile, and whose is it?” a real question rather than a formality. Cruise missiles are fast and low, and they leave little time. And cheap drones turn every engagement into a cost problem: a fighter firing a missile worth far more than its target cannot do so night after night. SACEUR said on the day Eastern Sentry was launched that NATO needed “lower cost weapons”. Allied Air Command’s commander said in September 2026 that using air power against drones and cruise missiles is “not putting us on the right side of the cost curve”.

None of this makes fighters obsolete. The interceptions over the Baltic states published since — a Romanian F-16 over Estonia in May 2026, a French Rafale over Latvia in June, an Italian Eurofighter over Lithuania in September — were all flown by the rotating fighters. What it shows is that a system sized around fighters needs other layers to deal with the newer threats at volume, and that the binding constraints shift from detecting to identifying, deciding and sustaining.

What the August change added — and what it did not

What NATO published about the change is short. Its topic page, updated on 29 September, says that throughout 2026 NATO upgraded its Air Policing mission and integrated it with other activities — Eastern Sentry and the rotational model for air and missile defence agreed in 2023 — into “a comprehensive Air Defence posture”. It names the components: quick-reaction fighters, sea- and surface-based air and missile defence, and counter-drone systems. It keeps the command chain — SACEUR with overall responsibility, Allied Air Command overseeing — and the four-month rotation. The Ankara summit declaration does not mention the change at all.

The detail comes from the host states. Estonia’s defence minister said on 8 July that SACEUR would receive broader authority, that pilots would receive certain pre-authorisations, and that it would no longer be necessary to consult at the political level before shooting down a drone that had strayed into NATO airspace; he said Ämari would become a designated NATO operational air base rather than an auxiliary one, and that rotations for three years were agreed. Lithuania’s president described the change as a broader mandate and faster decisions. On 31 July the Estonian Defence Forces wrote that the new mission would make it possible to counter aerial attacks with NATO’s integrated air defence system.

The pattern is consistent. The change NATO and its members describe is above all a change in deciding: who may order an engagement, and how quickly. It is matched by national changes on the same step. Romania legislated in 2025 to let its forces shoot down drones in peacetime; Lithuania’s parliament amended its laws in September 2025; Estonia’s government sent a drone act to parliament in April 2026; and Poland says that from 1 October 2026 radar-based identification replaces the visual check before an engagement. None of these rules is published in full, and NATO publishes no rules of engagement at all, so the size of the change cannot be measured — only that it was made at the step the newer threats stress most.

What the change does not establish is the ground layer. NATO names the classes of system that make up the mission, not the systems. Its air commander said in September that small drones are “typically a national responsibility”. The next question is therefore what the national layers actually consist of.

Delivered is not deployed, and deployed is not integrated

Air-defence procurement on the flank is announced loudly and delivered quietly, and the two are easily confused. “Country X bought Patriot” and “Patriot is defending this sector under NATO command” are different claims separated by several steps, each with its own evidence. The figure below takes seven cases through those steps and stops each one where its public evidence stops.

Three things stand out. First, the Baltic medium-range layer is real but young. Estonia’s Air Defence Wing received the first of its three IRIS-T SLM fire units at Ämari on 22 June 2026, nearly three years after the contract; the Air Force commander said it would take months to train the operators and field the system. Latvia ordered the same system in November 2023; its Air Force commander said on 1 October that the first package is due by the end of the year and combat duty in the first half of 2027. Lithuania’s Air Defence Battalion began receiving its second NASAMS battery, ordered in 2023, in April 2026; a third, contracted in October 2024, is due in 2028. These are three different stages, and a single headline — “the Baltic states are buying air defence” — would erase all of them.

Second, the most mature national layer is outside the Baltic states. Poland’s first Patriot squadron reached full operational capability in December 2025, seven years after the contract. Its counter-drone system, contracted on 30 January 2026 for about fifteen billion złoty, shows no delivery yet.

Third, integration is the stage the public record reaches least. Latvia’s Air Force commander put the distinction plainly: learning the procedures is one thing, and integrating the system into NATO’s overall surveillance and response system is another. Only one allied ground-based unit in the Baltic states is publicly described by its owner as integrated — the Spanish NASAMS unit in Latvia, there on rotation since June 2022, which Spain’s defence staff calls “fully integrated into NATO’s air defence network”. Rotations elsewhere show the same pattern of temporary depth: Germany’s Patriot deployment in Poland, protecting the logistics hub for support to Ukraine as part of NATO air and missile defence, ran from January to December 2025 and ended; the Netherlands took over and has extended its Patriot deployment to December 2026 at the latest.

And underneath all of it sits the stage nobody publishes. A multinational contract of January 2024 covers up to a thousand Patriot missiles for four Allies; no delivery under it is established, and no flank state publishes an interceptor stock or a reload rate. A defence that cannot be measured at its last step cannot be measured as a whole.

Eastern Sentry: a bridge, not a structure

Between the September 2025 incursions and the August 2026 change, the additional assets came through Eastern Sentry. SACEUR ordered it on 12 September 2025, two days after the drones crossed into Poland, under his existing authorities; it created no new headquarters and no standing force. Allies sent what they chose and for as long as they chose: Danish F-16s and a frigate, French Rafales in Poland, German Eurofighters — from Poland between December 2025 and March 2026 — British Typhoons, Czech helicopters. NATO never published a consolidated list. In 2026 Allied Air Command describes itself as leading the activity, and NATO ran counter-drone trials under it in Romania and over the Baltic states.

NATO’s own description of Eastern Sentry is now in the past tense. Its air defence page says the activity “enhanced NATO Air Policing and served as a bridging mechanism while NATO worked on more permanent reinforcement”, and that the assets first deployed under it have been integrated into NATO Air Defence along the eastern flank. NATO has not announced an end to it, and its eastern-flank page of the same date still describes it in the present tense.

What Eastern Sentry added was flexibility and mass on demand, and an experiment in stitching national systems together. What it did not add was permanence: its contributions came and went, and its name never established which ground-based systems were under NATO command. The air defence mission inherits both its assets and that limit.

What deeper air defence would look like

If the posture is deepening, it will be visible in a small number of observable facts, and none of them is an announcement:

• a delivery statement naming the receiving force for a ground-based or counter-drone system — Latvia’s IRIS-T, the first elements of Poland’s counter-drone system, Lithuania’s third NASAMS battery;
• an operational declaration for a delivered system, as Poland made for its Patriot squadron and as Estonia has not yet made for its IRIS-T;
• a statement, by NATO or by the owning state, that a named system serves the air defence mission or is integrated into NATO’s air defence network, as Spain has made for its NASAMS in Latvia;
• a rotation that becomes persistent, or a published change to the mission’s rotation or bases;
• a published delivery of interceptors, or a published stock;
• a published rule or delegation for engagement under the new mission.

Vigil tracks these on its standing question on air and missile defence depth, which this Note has updated.

Where the public evidence ends

The mission’s change is documented by NATO and by the host states. The delivery stages in the figure are documented by the receiving ministries and armed forces, with one exception noted. Beyond that the record thins quickly. NATO does not say which national systems serve its air defence mission, or under whose command they engage. It publishes no rules of engagement, so the change in authority rests on a minister’s account. No flank state publishes interceptor stocks. And no public source states where any system stands or what it protects — which is why this Note draws no map of coverage and infers none from the range a manufacturer quotes.

What can be said is narrower and firmer. The Baltic mission now has an air defence mandate and, by its members’ account, faster authority to use it. The fighters that fly it are the same rotation that has flown since 2004. The layers that would make it air defence in depth are arriving unevenly, mostly under national command, and mostly not yet at the stage where the public evidence says they are part of the mission. That is not a verdict on whether NATO can defend the Baltic states. It is a statement of what has been built, what is being built, and where the record stops.