Space Technology · 2026-10-07
Themis Rocket: Europe Is Finally Building a Rocket That Lands Itself — and the Test Just Happened Inside the Arctic Circle
A rocket is sitting in a Swedish forest, minutes from the Arctic Circle. The ground team just pumped in supercooled liquid oxygen and liquid methane — cold enough to frost the steel hull in October air. They ran through every pre-launch system. Twice. Then they didn't launch. That was the plan. And somehow, that quiet non-event matters more than most actual launches.
This is Themis — ArianeGroup's reusable rocket demonstrator. On October 7, 2026, ArianeGroup and the Swedish Space Corporation completed the second wet dress rehearsal of the Themis vehicle at Esrange Space Center in Kiruna, Sweden. ESA confirmed the test successful. After 15 years of watching SpaceX stick landings while European programs wrote the cheques, something is finally moving in Lapland.
What is a wet dress rehearsal — and why does two of them matter?
A wet dress rehearsal is the closest you get to launch without launching. Engineers fully fuel the rocket with real cryogenic propellant, cycle through every system, run the countdown to T-zero, and then safely drain the tanks. It stresses the vehicle exactly as a real launch would — thermal loads, pressurisation cycles, sensor responses — without the risk of ignition.
SpaceX ran multiple WDRs before every major vehicle. Artemis I completed three before SLS finally lit. Completing a second WDR means Themis has now survived full cryogenic fuelling twice in sequence. The hardware is behaving predictably. The vehicle is learning to be a rocket.
What is the Themis rocket, exactly?
Themis is not a satellite launcher. It is a technology demonstrator — a prototype first stage designed to prove one thing: that Europe can do what SpaceX has done 250-plus times. Launch. Separate. Land vertically. Fly again.
It is powered by the Prometheus engine, a methane/liquid-oxygen engine developed entirely by ArianeGroup. Methane burns cleaner than kerosene, produces far less coking residue between flights, and — this is the part that sounds like science fiction — can theoretically be synthesised from atmospheric carbon dioxide and subsurface ice. It is the same reason SpaceX chose methane for Raptor. The same reason Rocket Lab chose it for Neutron. The engineers who build rockets for the next century are all betting on this fuel.
Esrange Space Center, where Themis is being tested, has been operational since 1966. It is one of Europe's oldest and most experienced launch ranges — 5,200 square kilometres of Swedish Lapland wilderness, one of the largest overland launch corridors on the continent. This is not an improvised test site. It is decades of institutional expertise pointed at a fundamentally new problem.
Why did Europe fall so far behind on reusability?
This is the uncomfortable part. Europe was not late to space. Ariane 5 was, for nearly two decades, the most reliable heavy launcher on Earth. But a decision was made — repeatedly — to keep every rocket expendable. Every vehicle that flew was thrown away. Meanwhile, SpaceX began iterating on landing technology from 2013 onward and achieved it in 2015.
When Ariane 5 retired in 2023 and its successor, Ariane 6, accumulated years of delays, Europe found itself in a deeply awkward position: unable to access orbit on its own schedule at competitive cost, renting launches on vehicles built by the entity it was theoretically competing with.
The gap is not only price. Reusable rockets can fly on much shorter turnaround cycles — the same booster can launch multiple times in a month. Expendable rockets are manufactured from scratch each time, like building a new car for every taxi ride and scrapping it at the destination. Europe's current position is exactly that metaphor, competing against a fleet that refuels and reruns the same vehicle daily.
Is Themis actually enough to close the gap?
Themis is a demonstrator, not a production rocket. Think of it the way SpaceX thought about Grasshopper — the small test vehicle they flew repeatedly in 2012 and 2013 to prove landing algorithms before applying them to Falcon 9. Themis is Europe's Grasshopper moment. The technology, if it works, feeds directly into what comes next.
The current candidate for a production successor is sometimes referred to as Ariane NEXT — a partially or fully reusable launcher designed to be competitive in the 2030s market. No final design has been locked. The outcome of Themis tests will almost certainly shape what that rocket looks like.
The competition is not standing still. China's Zhuque-3 came within metres of a successful vertical landing before toppling over — that story is in the SkyLens archive. Rocket Lab is developing Neutron with reusability built in from the start. Blue Origin's New Glenn already recovers its first stage. The 2030s launch market will likely have four or five operators doing what only SpaceX does today. Europe's question is not whether reusability is the future. It is whether Themis buys enough time to build a competitive vehicle before the market hardens around existing players.
Reusability milestones — where Themis fits:
What happens after the static fire?
After a successful engine ignition test, the next step is a hop test — Themis lifts off a short distance, hovers, and lands. SpaceX used hop tests to validate both Grasshopper and early Starship prototypes. It is the moment you find out whether the landing software actually functions under real aerodynamic loads, not just simulation.
If the hops go well, ArianeGroup moves to progressively higher-altitude test flights over Esrange's 5,200-square-kilometre range. The wilderness of Swedish Lapland provides something rare: room for failure without consequence. Room to iterate. Room to learn.
It will not be fast. Aerospace never is. But when Themis eventually lifts off and — if the engineers have it right — settles itself back onto its landing legs, it will be the most significant moment in European rocketry since Ariane 5 first flew in 1996. Learn how orbital mechanics and rocket staging actually work if you want to understand exactly what the landing manoeuvre demands.
SkyLens editorial — live CelesTrak + NASA/JPL data (15968 objects)
Related stories

