Großwald Curated No. 42 — The eyes got cheap, the shots didn't

Europe bought back overhead reconnaissance at startup speed

This week ICEYE, the Finnish builder of synthetic-aperture radar (SAR) satellites, announced a Series F round whose structure matters more than its headline total: €450 million of primary capital led by General Atlantic at a valuation above €10 billion, with a secondary placement of more than €550 million taking the headline total past €1 billion (dispatch). The primary €450 million is the part that funds satellites; the secondary is existing holders selling down, so none of it reaches orbit — and the round is announced rather than closed, pending regulatory approvals expected in the third quarter. What the fresh capital is buying is the point. Sovereign and strategic balance sheets anchor it — the Qatar Investment Authority, Finnish state vehicles, Nokia — beside ordinary growth funds. Read that as more than venture money chasing launch stories: a continent is funding a sovereign alternative to the space capability the United States signalled on Friday it means to keep — the ability to see from orbit, in any weather, day or night.

The synthetic-aperture distinction matters here: unlike optical satellites, SAR sees through cloud and darkness — the conditions an eastern-flank crisis would actually unfold in. ICEYE is selling revisit speed: a constellation large enough that any point on earth is re-imaged in hours, then minutes, which is what converts a satellite from a mapping tool into a tracking one. The firm is doubling production from roughly fifty satellites a year toward a hundred by 2028 — and that manufacturing cadence, more than the imagery, is what matters.

The cadence is the whole argument: it is what moved overhead reconnaissance across the seam — from a capability Europe could in principle own to one it can now actually build at speed. Consider Poland. Under a €200 million contract signed in May 2025, ICEYE built and delivered a sovereign four-satellite SAR constellation — designated POLSARIS, the Polish SAR Intelligence System, operated by ARGUS, the geospatial-intelligence agency Warsaw's defence ministry stood up only in 2024 — and took it from signature to delivery in under twelve months. A sovereign overhead-reconnaissance capability, contracted and orbiting inside a year, is a timeline that has no equivalent anywhere else in this briefing. It is faster than a tank order, an ammunition line, or a single Patriot battery's delivery. The layer that was supposed to be the hardest to replace turns out, in its imaging form, to be the fastest.

ILA was where the German version organised itself. Rheinmetall and ICEYE used the show to formalise Rheinmetall ICEYE Space Solutions — a joint venture, Rheinmetall holding sixty per cent, built in Neuss to anchor a sovereign German SAR constellation and pull the country's NewSpace tier in behind it: Reflex Aerospace, OroraTech, ConstellR and LiveEO as launch partners, on a deliberately open architecture. The anchor contract already exists — SPOCK, awarded to the Rheinmetall–ICEYE venture in December 2025 at €1.7 billion, for roughly forty SAR reconnaissance satellites by the end of the decade — and it sits inside the €35 billion Berlin committed in February to low-earth-orbit resilience. In a parallel stand, Airbus assembled its own all-German space-intelligence alliance — Rohde & Schwarz, ConstellR, OrbInt and HPS, the thermal-imaging firm ConstellR turning up in the Rheinmetall lineup too — under the same banner of sovereign reconnaissance, and a third effort surfaced the same week as OHB tied up with Helsing for space-based observation and tactical targeting, having been shut out of the no-tender SPOCK award. At least three sovereign-constellation efforts, stood up in one country in one week, cut two ways. They confirm that the capital and the political demand are both real — but they also carry a familiar risk, one the overlapping rosters already hint at: a finite German demand split across three sub-scale constellations is the way to end up with none of them reaching a tracking-grade revisit cadence. Consolidation, or a clean division of orbits, is the thing to watch next. And one dependency sits a tier up and outside the SAR line entirely — responsive, sovereign launch, being built by the same new tier on the same compressed clock (indicators).

The honest objection is that a picture is not a target. SAR compresses the front of the kill chain: it is a find-and-fix sensor, and it is targeting-relevant — the same fusion of radar imagery with signals intelligence is what China has built to accelerate its own find-fix-finish cycle, and the law-of-war literature now treats commercial SAR as targeting infrastructure precisely because it works. But the back of the chain is a different enterprise. Boost-phase missile warning — the infrared tier that catches a launch in its first seconds — is not what ICEYE builds. Assured positioning and timing is the back-of-chain layer where Europe holds a sovereign card, in Galileo's encrypted Public Regulated Service, but the card is not yet fully played: military receiver integration across the force lags, and anti-jam parity against high-end electronic warfare is not demonstrated. And the fused engagement layer that turns a detection into a clearance to fire is the connective authority that remains unbuilt. Europe is buying eyes, and buying them fast; it has not yet bought — or, for PNT, not yet fielded at scale — the nervous system the eyes report into. The space contribution the United States is keeping is mostly that nervous system, which is why the SAR build-out and the American reassurance sit alongside each other rather than cancelling out.

The jammer Europe still does not have — and the route it chose to it

The second information layer on display is the one Europe is most behind on; the route it has chosen to close the gap is the story. Hensoldt brought the broadest spread: Kalætron Attack, an airborne jamming family built to suppress enemy air defences; Kalætron Integral, an integrated signals-intelligence suite; the PEGASUS strategic-reconnaissance system; and TAERVUS, a full electronic-warfare system combining direction-finders, receivers and jammers across the COMINT and ELINT bands with the firm's own Spectrum Battle Management software. It also unveiled, for the first time publicly, a "Battle Lab" — a live demonstration of sensors, effectors and command levels meshed into one software-defined network, which is the connective layer this briefing keeps naming as the thing Europe does not build, shown as a working demonstrator rather than a slide.

The gap underneath all of it is specific and long-standing: NATO outside the US Navy has no equivalent of the EA-18G Growler — no escort jammer that flies with a strike package and jams for it. The SEAD platforms Europe does field — German Tornado ECRs at Schleswig, Italian ones at Ghedi, US F-16CMs at Spangdahlem — can find and shoot at radars but cannot jam for other aircraft. Italy has ordered two EA-37B jamming aircraft for late this decade; that is the extent of NATO's organic stand-in outside American squadrons. Germany's own fix, the Eurofighter EK — fifteen jets NATO-certified by 2030, Saab's Arexis suite and the AGM-88E — is stand-off SEAD; its airborne support-jamming "Step 2," the part that actually replaces the Growler, is not planned before the early-to-mid 2030s. On the manned, prime-led path, the escort-jamming gap stays open the better part of a decade.

Which is why the most revealing electronic-warfare exhibit at ILA was not a conventional aircraft. Helsing unveiled the CA-1EA, an autonomous electronic-attack drone — a common-airframe variant of its CA-1 Europa, running the Centaur autonomy stack the company flew in beyond-visual-range trials on a Saab Gripen E last year, with a Hensoldt broadband jammer and a Grob airframe — pitched as a European Growler alternative for service around 2031. But look past the timeline to the design — that is where the real bet sits. Escort jamming against a peer air-defence network is a power-and-aperture problem: a Growler concentrates a large electrical supply and a big antenna in one airframe. A four-tonne attritable drone has a fraction of both — and Helsing concedes the point by adding a second generator dedicated solely to the jamming suite — the clearest sign that power, not the airframe, is the binding constraint — for a stated reach of around 100 kilometres. So the CA-1EA is more than a cheaper Growler arriving sooner. It is a wager on a different doctrine: that distributed low-power jamming from many cheap nodes can substitute for the concentrated high-power jamming of one expensive crewed platform. That is a contestable claim, and a more interesting one than the timeline race — but it means electronic attack is itself a split capability: the airframe and the autonomy around the jammer move quickly, while the watts inside it do not.

The find outruns the finish, and the reach has no answer at all

Turn from the layers made of information to the layers made of mass, and the clock stops. The week's clearest exhibit of the problem also looked, at first glance, like its solution. Diehl and Polaris unveiled the Cobra 600, a jet-powered drone whose entire reason to exist is to carry a single IRIS-T air-to-air missile aloft and extend its reach from the roughly 40 kilometres of a ground launcher to a few hundred kilometres airborne, returning to be reused or expended one-way. It is an elegant idea — a cheap, runway-or-highway-launched truck that multiplies the footprint of a proven missile. And the gain is real but narrow. A forward, airborne launch improves an engagement's geometry — better aspect, more energy, an earlier shot — and so can raise the hit probability per round, which, exactly as better sensing does, deepens the effective magazine. What it cannot do is change the rate at which the magazine refills: the Cobra 600 still expends one IRIS-T per engagement, and the IRIS-T, like every effective interceptor, comes off a finite production line. Against saturation it is the refill rate that binds, and a cheap truck riding on a rate-limited round does not raise it. The drone is attritable; the missile it carries is not, and the missile is what was scarce.

This is where the "effector" story has to be split in two, because the two halves run on different clocks and conflating them is the error of the week's optimistic coverage. On the offensive side, cheap mass is arriving. MBDA used its ILA presence to push what it calls "smart mass" — effectors designed for cost and volume through civil-military co-production. The One Way Effector is a turbojet loitering munition built to saturate air defences — MBDA's public figures give it a three-metre wingspan and a payload around 50 kilograms, with France's procurement agency having signed the first development contract in January and a first production batch due in 2027; the heavier Crossbow variant runs past 800 kilometres with a 300-kilogram modular payload, and it is that one MBDA says can be in series production as early as this quarter. Cheap one-way strike, at saturation volumes, on a timeline measured in quarters — that is real, and it is the appropriable answer working as advertised.

On the defensive side, it is not — and that asymmetry is where the seam actually runs. A hit-to-kill round has several pacing items, most of them mass that will not get cheap. The solid rocket motor is the most widely documented chokepoint in Western munitions; and on the modern kinetic rounds, the divert and attitude-control thrusters that steer the kill vehicle in its terminal second are a precision-mechanical subsystem with their own narrow, export-controlled supply base and no commercial volume to ride. The seeker alone points anywhere else, because part of it is sensors and compute. But only part. The digital back-end that rode the commercial curve is already the cheap half of a seeker; the costly half — a cryo-cooled infrared focal-plane array, or the export-controlled g-hardened MMICs of a Ka-band active seeker — did not. And export control is a security regime — volume cannot dissolve it the way it dissolves an ordinary price. The escapable residue, the one-shot qualification burden amortised over rounds that are then expended, is smaller than it looks. It is why the PAC-3 line in Camden and the Aster ramp (which MBDA publishes in ratios, not units) stay dear — and why a war-economy entrant like Ukraine's Fire Point can cheapen every part of its round and still stall on the seeker it has to import. So you cannot make defensive mass cheap by bolting a still-expensive seeker onto a cheap airframe, which is exactly what Cobra 600 does. The one structural escape on display was directed energy — MBDA's DefendAir trains a high-energy laser, with a theoretically unlimited number of shots, on close-in targets — but a laser's bottomless magazine holds only at short range, in clear air, against the cheap drones at the bottom of the threat set, not the ballistic and cruise rounds at the top. The deep magazine stays as shallow as it was, and putting more launch platforms in front of it draws it down faster.

Then there is the gap ILA produced no answer to: the tanker. The United States is withdrawing all eight aerial refuellers it committed to NATO's Force Model — the committed allocation in full, within a wider US theatre tanker fleet that a separate leaked count put at a trim from 71 to 63 (No. 41) — and aerial refuelling is what makes every other aircraft usable at range, the fighters that survive the cut included. Europe's answer is incremental and entirely conventional, and it long predates the show: Italy had formalised six Airbus A330 MRTT tankers for about €1.39 billion back in April, the NATO pooled fleet at Eindhoven continues its slow build, the A400M can buddy-refuel as a part-time stopgap — and ILA itself produced no dramatic tanker news at all. No. 41 filed the tanker in the appropriable column — Europe owns the A330 conversion line, so the gap read as unbought rather than unbuildable — and that classification still holds. What this edition adds is why "unbought" is so sticky: there is no NewSpace tanker, no startup disrupting air-to-air refuelling, no cost collapse, because a wide-body aircraft costs what it costs and takes the years it takes. The owned line is real and still offers no fast clock. Of all the layers the American withdrawal named, the tanker has the slowest clock and the fewest contenders — the purest case of mass that cannot be made cheap, and so cannot be made fast.

The eyes got cheap, the shots didn't — and the bill comes due before Ankara

One mechanism runs under all three sections. A capability gets cheap when commercial volume pulls its price down, and stays dear when there is none — which is why the eyes, riding the same cost curves as the rest of the space economy, fell fast, and the shots, riding nothing, did not. The two in-between cases prove the rule rather than breaking it: the jammer is quick in its airframe and slow in its watts, and the seeker is cheap only where it was always cheap — its digital back-end — while export control, a security rule no volume can buy down, keeps the rest beside the rocket motor and the kill vehicle. In each, the part made of information moved and the part made of mass did not. The one honest complication is that the eyes are not wholly cheap either: the boost-phase warning, the fused firing authority and the assured timing they report into are bespoke, not mass-produced, and so sit on the slow side with the shots. Europe is buying the picture fast; it has not bought the nervous system that turns a picture into a clearance to fire.

The week put a price on the gap. Two UK ministers resigned over an underfunded Defence Investment Plan, and Al Carns, on his way out, named it in money: platforms that cost billions, he said, can be defeated by systems that cost thousands — the same seam read at the muzzle rather than the production line, since mass that is dear to build is also dear to expend. The mismatch does not bite evenly, and the edition's own war line is the proof. At the bottom of the threat set — reconnaissance drones, the refineries and depots Ukraine spent the month burning — cheap finding paired with cheap effect is genuinely potent, which is exactly why that campaign reads as success. It bites at the top: against ballistic and cruise saturation and contested air, where the finish depends on the scarce interceptor and the missing tanker, a force can see and decide faster than it can shoot or reach. Better sensing does deepen the magazine — fewer rounds wasted, the rest better allocated — so the danger is precise rather than total: at the high end the eyes validate more targets than the magazine can service. FCAS is the same fact at its sharpest. The airframe broke on the who-leads dispute this franchise named before, but the metal was never the problem (No. 41; Signal No. 78); what the collapse exposed is that the engine — the most sovereignty-critical pillar, the one we argued worked on the merits — is now orphaned, because Germany has no national fighter-engine integrator to house the core it cannot build.

What Washington has asked Europe to backfill comes down to one question: whether the pledges are denominated in mass — tankers, maritime patrol, interceptor lines, in which case the slow clock governs and the cost lands in the 2030s — or in the bet that information can substitute for it, a thicker sensor-and-autonomy layer letting fewer shooters do the work of more, the faster wager the week was implicitly placing, and unproven against a peer that can contest the spectrum it runs on. Ankara, on 7–8 July, is the deadline that forces the answer; the answer itself is settled earlier. The harder signals come earlier: the Brussels ministerial on 18 June, four days out, where the backfill ask either becomes assigned forces with dates or stays a communiqué of intent; the German Patriot decision; whether serious RFI responses land by 21 August. By the time the summit laminates the result into a press release, those will have shown which clock Europe is actually on.