An air-to-air missile is built on an assumption: that something else will carry it to the fight. It expects to be released at nine or ten kilometres, already doing Mach 0.9, from a wing rail that has spent twenty minutes climbing. It expects a fighter radar to tell it where to look. Take those two gifts away and you have a rocket with a very short memory.

Ukraine has now built an entire air defence system around taking them away on purpose. At the MSPO defence exhibition in Kielce, Poland, on 9 and 10 September 2026, Ukrainian industry displayed the working components of Shershen — Шершень, Ukrainian for hornet, and the spelling used by the developer and by every outlet that covered the reveal — a short-range surface-to-air system with its own command vehicle, its own surveillance radar, and launchers deliberately designed to fire whichever air-to-air missile happens to be in the warehouse that week.

The trucks are the least interesting part. What matters is the engineering underneath: what actually has to change when a missile designed to be carried, cued and released by a fighter is instead bolted to a rail at sea level and told to start from zero.

What Ukraine actually showed at Kielce

The display was put on by the National Association of Ukrainian Defense Industries (NAUDI), which describes one of its members as the programme’s lead developer and integrator. Ukrainian outlet Militarnyi, whose correspondent was on the floor, attributes the stand to an international consortium trading as Shershen Group and reports that it brought three things: a command post, a launcher, and a target-tracking radar. Separately, the OSINT account that first posted images from the hall credited the launcher to the Ukrainian radar house Radionix. Those attributions are not contradictory — this is a consortium product — but no single official release has yet pinned down who builds which box.

TWZ, which reported the system in detail on 10 September, describes a Shershen battery as one Perun command-and-control vehicle, one Athena surveillance radar, and four Borei launchers. Militarnyi’s account matches in shape if not in branding: a target-tracking station, a combat control post, and up to four launchers, plus loading and transport-reloading vehicles.

The launcher itself borrows a trick from the Israeli Barak family. The launch unit is lifted off the truck by a multilift system and left on the ground as a self-contained object, which means the vehicle can drive away and the launcher can sit camouflaged somewhere unhelpful for a Russian targeting cycle. The example at MSPO is reportedly the second operational prototype, and is described as close to its production configuration.

An earlier version of the story is worth keeping in view: this is not a first reveal. Shershen surfaced in February 2026 as a set of models at the World Defense Show in Riyadh. What changed in Kielce is that the models became hardware, with a manned command post showing commander and operator workstations.

The developer says the second prototype has completed initial testing including successful intercepts at a test range. No footage of those intercepts has been released, and no independent party has verified them.

The missile list, and what is actually confirmed

This is the part of the story most worth getting right, because a multi-calibre claim is easy to make and hard to check. Here is what the sources actually name.

TWZ reports that the company says Shershen can currently employ the R-27 (NATO: AA-10 Alamo), the R-73 (AA-11 Archer), the AIM-9 Sidewinder, und die AIM-132 ASRAAM, with compatibility for the R-60 (AA-8 Aphid) still being explored. Militarnyi lists the same four plus the R-60 without the caveat. That is the whole confirmed list: two Soviet infrared and semi-active radar types, one Soviet dogfight round, one American infrared round, one British one.

A sixth weapon is Ukrainian and does not exist yet in serial form. The R-120K — Militarnyi calls the displayed article a mock-up of the R-120 — is a new radar-seeker missile intended to work both as an air-to-air weapon and as a surface-to-air interceptor. TWZ reports that work is continuing on its rocket motor, which means its performance and production status are genuinely unsettled. It sat on a Borei launcher at Kielce as a statement of intent, not as a fielded round.

Two things are nicht on the list, and it matters. Nobody has claimed the AIM-120 AMRAAM for Shershen — the AMRAAM is NASAMS’ round, and Ukraine is short of it. And the RIM-7 Sea Sparrow, which does appear in Ukrainian air defence, belongs to a different programme entirely. Treat any coverage that folds those into the Shershen missile list as unsourced.

Honcharov also put a number on the testing that is more honest than most exhibition patter. Asked why the system had been fired with exactly five missile types, he told Militarnyi the figure described what had been done rather than what was possible.

Why launching from the ground is not a free lunch

Start with the energy. A Sidewinder leaving the wingtip rail of an F-16 at 10,000 metres and Mach 0.9 is handed roughly 270 metres per second of velocity and ten kilometres of altitude before its own motor contributes anything at all. Neither of those came out of the missile. They came out of the aircraft, which spent fuel and time acquiring them.

Bolt the same missile to a launcher in a field and it starts at zero, at sea level, pointing at the sky. The motor has not changed. The propellant load has not changed. Everything the aircraft used to provide now has to come out of the same finite burn — and it has to come out of it in the worst possible place.

Air at sea level is about three times denser than air at 10,000 metres — roughly 1.23 kilograms per cubic metre against 0.41 in the standard atmosphere. Drag scales directly with that density and with the square of speed. So a ground-launched round burns a large share of its boost accelerating from a standstill through the thickest air there is, while also climbing against gravity, before it reaches the thin layer where an air-launched missile would already have been cruising.

The result is not a small penalty. It is the reason a family of missiles credited with impressive reach from a fighter becomes a short-range weapon on a truck. Militarnyi offers a usefully concrete data point: the R-27ET1, the infrared-homing variant used in India’s ground-based Samar-2 system, is credited there with an engagement range against an aerial target of no more than about 20 kilometres. The same airframe, the same motor, a very different envelope.

That is the honest context for Shershen’s advertised numbers. TWZ reports a claimed capability against aircraft, cruise missiles and drones out to 25 kilometres and up to 10,000 metres, and notes plainly that the figures have not been independently verified and that it is unclear what target types and engagement conditions produced them. Since the system is explicitly designed to fire missiles with wildly different motors and seekers, a single headline range is close to meaningless. The number a given Shershen battery can actually achieve depends entirely on which rounds are on the rail that morning.

Seekers that expect an aeroplane

Energy is the obvious problem. Guidance is the harder one.

An infrared air-to-air missile does not find its own target. On a fighter it is cued — the radar, the infrared search and track set, or the pilot’s helmet-mounted sight tells the seeker roughly where to look, the seeker slews there, acquires, and growls. All of that happens over an interface built for that specific aircraft, with the aircraft’s power, cooling and mission computer behind it. R-73, R-60, AIM-9 and ASRAAM all arrive expecting that conversation.

On the ground, every part of it has to be rebuilt. Something must compute the target’s bearing and elevation, hand that to the seeker, cool it, confirm lock before launch, and do it fast enough to matter against a cruise missile. Britain’s containerised Gravehawk system, built around Ukraine’s R-73 stocks, solves this by packing an electro-optical and infrared targeting camera into the ISO container alongside the missiles and the generators — the container is not just a box, it is the aircraft the missile no longer has.

Radar-guided rounds bring a different bill. Semi-active types need a ground illuminator holding the target through intercept, which constrains how many engagements a battery can run at once and ties the launcher to the radar. TWZ reports that when Shershen employs radar-guided missiles, the engagement architecture can reportedly guide as many as six simultaneously — a claim that, if accurate, says more about the Perun command vehicle than about any of the missiles.

Shershen’s answer to the sensor problem is to refuse to have one. Honcharov has been blunt that the system is not married to a radar: there is, he told Militarnyi, no strict requirement that only Thales, HENSOLDT, Teledyne or another supplier’s set be used, and the control system can be adapted — a claim he pairs with the assertion that it costs at least three times less than comparable systems abroad. Militarnyi adds that the system can also take target data from the Oreanda and Virazh-Planshet command-and-control systems used by Ukraine’s air force and air defence troops, or from the Krechet control vehicle, which aggregates feeds from other radars.

There is a geometry penalty too, and it is the one that matters most against Shahed-type drones. A fighter looks down. A ground radar looks out, and the horizon gets in the way. A cruise missile or a one-way attack drone flying low and slow is detected late, which compresses the time available to cue a seeker, launch, and let a missile that started from zero climb into a firing solution. An 85-kilometre detection range against a high-flying aircraft is not an 85-kilometre detection range against something hugging the terrain.

From FrankenSAM to Franken-architecture

None of this is new in principle. What is new is the intent.

The FrankenSAM effort, run with United States support from roughly 2023, took the existing Soviet hardware Ukraine already operated and taught it to fire Western rounds. The best-documented variant mates a Buk transporter-erector-launcher-and-radar with RIM-7 Sea Sparrow missiles — a ship’s point-defence weapon fired from a tracked vehicle in a Ukrainian forest. A separate strand paired Soviet-era radars with AIM-9M Sidewinders. The conversions were tested in the United States before fielding.

The same instinct produced a small zoo of conversions. Ukraine has fielded Osa-AKM vehicles carrying R-73s instead of their original 9M33 interceptors. The United Kingdom supplied Raven, a pair of ASRAAMs on a Supacat HMT 600 six-by-six. The UK and Denmark funded Gravehawk, the containerised R-73 launcher. R-73-based launchers have even turned up on Ukrainian uncrewed surface vessels.

Every one of those is a point solution: one launcher, one missile, one integration job, done once. The moment the stock of that particular missile runs out, the launcher is furniture. That is precisely the failure mode Honcharov described, and it is not hypothetical — TWZ has reported Ukrainian F-16s running low on Sidewinders, AMRAAM shortages hitting the F-16 force, and Patriot interceptor stocks so thin that ballistic missiles reaching Kyiv have gone unengaged.

Shershen inverts the order of operations. Instead of adapting a launcher to a missile after the fact, it starts from an open architecture and treats the interceptor as a variable. The launcher rail, the umbilical logic, the pre-launch initialisation and the fire-control interface are designed from the outset to speak several dialects. That, not the truck, is the actual product — and it is also why the software is the hard part rather than the metalwork.

Where Shershen fits — and what it does not replace

Shershen is a short-range system and its developers do not pretend otherwise. Militarnyi places it, in missile terms, in the same band as the German IRIS-T SLM or the Norwegian-American NASAMS rather than anywhere near Patriot. TWZ is blunter: Shershen is focused on the short-range layer.

That leaves the ceiling to other programmes. Ukraine’s Koral, a medium-range interceptor traced back to a 2021 Luch design bureau concept, made an apparent first public appearance earlier in 2026 on a Tatra chassis. Patriot and SAMP/T remain the only answer to ballistic threats. Shershen does nothing about any of that.

What it does address is the arithmetic problem underneath the whole Ukrainian air defence effort: a shooter is worthless without ammunition, and every Western system Ukraine operates ties its launchers to one manufacturer’s round. A battery that can fire a Soviet R-73 on Monday, a donated ASRAAM on Tuesday and a Ukrainian R-120K in 2028 is not a better interceptor. It is a hedge against running out.

Whether it works is a separate question, and one nobody outside Ukraine can answer yet. The intercepts are claimed, not shown. The envelope figures are the developer’s. The indigenous missile still needs a rocket motor. Treat the Kielce display for what it was: a serious piece of systems engineering, presented at a trade fair, by people who need it to sell.

Sources: TWZ (Thomas Newdick, 10 September 2026); Militarnyi (Mykhailo Liuksikov, 9 September 2026 and 12 February 2026); UNITED24 Media; Ukrainska Pravda; UK Defence Journal; The Armourer’s Bench; Defense Express.