# [Video] Why Europe Still Throws Away Its Rockets (The Physics of Reuse)

> Fuel is 0.3% of a rocket's cost and the first stage around 60%. That is why SpaceX flies its boosters up to 37 times, while Europe's reusable stages are still waiting for their first hop.

- Published: 18 Oct 2026 (2026-10-18)
- Author: Stanislav Milchev, Editor
- Section: Markets
- Tags: reusability, Falcon 9, Themis, European Launcher Challenge, Prometheus
- Canonical URL: https://www.spacecapex.com/en/analysis/why-europe-is-losing-reusable-rocket-race

Imagine flying from Sofia to London and throwing the plane away after landing. A new Boeing 737 lists at more than $100m, so the ticket would be a little steep. Almost the entire history of spaceflight has looked exactly like that. Europe still flies that way today.

### A rocket made of fuel

Blame the Tsiolkovsky equation. Low orbit requires a speed of around 7.8 km/s. Add the losses to gravity and air resistance, and the bill comes to around 9.4 km/s. But speed grows only with the logarithm of the ratio between the mass with fuel and the mass without it. That is why a rocket is mostly propellant wrapped in a little metal. Falcon 9 weighs around 549 tonnes on the pad and at best puts 22.8 tonnes into low orbit. In other words, about 4% of everything that lifts off reaches orbit.

<https://youtu.be/hH5ZaWWbUPs>

### Fuel is cheap, metal is expensive

Here's the surprise. Back in 2013 Elon Musk said propellant accounts for only 0.3% of a rocket's cost, [Space.com](http://Space.com) reports. In 2016 he specified that fuelling a Falcon 9 costs around $200,000, while building one costs around $60m, [Manufacturing.net](http://Manufacturing.net) writes. The money is in the hardware. When unveiling the Block 5 version in 2018, Musk broke the cost down like this: the first stage is close to 60%, the upper stage around 20%, the fairing around 10% and launch operations another 10%. That account comes from [NASASpaceflight](https://www.nasaspaceflight.com/2018/05/block-5-spacex-increase-launch-cadence-lower-prices/). Drop the first stage in the ocean and most of the money sinks with it.

![SpaceCAPEX](https://omlclbaklhdpibyhxubz.supabase.co/storage/v1/object/public/media/articles/1792355810429-s04l1v.png "SpaceCAPEX")

### How to land a 40-metre stage

At separation the first stage is at around 70 km altitude and travelling at more than 2 km/s. It flips over and fires its engines to slow down. It then passes through the dense layers of the atmosphere, steering with grid fins, and fires once more just before the ground.

That last burn is the hardest. Even at minimum throttle, a single Merlin engine pushes harder than the almost empty stage weighs, so it cannot hover. The engine has to bring the speed down to zero at exactly zero altitude. Engineers call this a hoverslam. Light it too early and the stage starts climbing again. Too late and it slams into the deck.

### Landing has a price

The propellant held back for these burns, plus the legs and fins, cuts into payload. When the stage lands on a ship, Falcon 9 carries around 17.5 tonnes instead of 22.8, roughly a quarter less. Then every booster has to be inspected and readied for the next flight.

So reusability only pays if the rocket flies often. A booster good for 20 flights saves almost nothing if you need four launches a year. That was precisely Europe's argument a decade ago. In 2017 then Arianespace chief Stéphane Israël argued that reusability "may not bring the same benefits in Europe as in the US", because the Americans have a huge government market with guaranteed flights. The quote comes from [Europe in Space](https://europeanspaceflight.substack.com/p/low-launch-rates-dont-kill-reusability). SpaceX solved the problem in the simplest way possible: it became its own customer and launched thousands of Starlink satellites. According to the company, in the second quarter of 2026 alone, 28 of its 38 launches were for its own needs.

### The result

On 25 August 2026, booster B1067 flew for the 37th time, and it was also Falcon 9's 100th launch of the year, [Space.com](http://Space.com) reports. In 2018 SpaceX designed Block 5 for ten flights with minimal refurbishment, so reality has now beaten the plan almost fourfold. Ariane 6 flew four times in the first nine months of 2026.

SpaceX is no longer alone. Blue Origin's New Glenn landed its first stage in November 2025. On 18 August 2026, private Chinese company LandSpace landed a Zhuque-3 stage on a pad around 390 km from the launch site, [NASASpaceflight](https://www.nasaspaceflight.com/2026/08/zhuque-3-second-flight-successful-landing/) writes. The booster toppled over shortly afterwards, but the landing counts. A month earlier, state giant CASC had caught a Long March 10B booster at sea.

### A continent without a rocket

For Europe the issue isn't just cost. In 2023 Ariane 5 retired, Vega C was grounded after its December 2022 failure, and Russia's Soyuz had already left Kourou. Europe was left without its own rocket for medium and heavy payloads, while the small Vega was flying its last missions. The Euclid space telescope, Galileo navigation satellites and the Hera probe, sent to a binary asteroid, all flew on Falcon 9. One of the missions of today's record-holder B1067 was in fact Galileo L13.

### The answer starts with an engine

Prometheus burns liquid methane instead of kerosene. Methane burns cleanly, the engine doesn't clog with soot and is easier to fly again. The target is around €1m per engine, roughly a tenth of the cost of the Vulcain 2, Ariane 5's main engine.

Prometheus powers Themis, a 28-metre first-stage demonstrator standing on a pad at Esrange, in the Swedish Arctic. Its first flight will be a low hop. In January ESA was aiming for the first quarter, but the pad was buried in snow. "So we are waiting for spring," ESA's Toni Tolker-Nielsen told [European Spaceflight](https://europeanspaceflight.com/esa-waiting-for-spring-to-launch-themis-reusable-rocket-demonstrator/) at the time. Spring came and went, and so did summer. According to [SSC](https://sscspace.com/news/european-prototype-themis-completes-second-wet-dress-rehearsal-at-esrange/), on 29 September Themis completed its second full wet dress rehearsal with fuelled tanks, and the teams are still "defining a window" for the next test. The next Kiruna winter is weeks away.

The smaller Callisto is a joint project of France's CNES, Germany's DLR and Japan's JAXA. It stands 14 metres tall and is due to fly from Kourou. Its first flight has slipped from 2026 to 2027, [European Spaceflight](https://europeanspaceflight.com/cnes-call-reveals-inaugural-callisto-flight-test-pushed-to-2027/) reports. The project was conceived in 2015, the year Falcon 9 first landed.

### The start-ups

MaiaSpace, an ArianeGroup subsidiary, is designing a small rocket whose first stage lands on a barge and can fly up to five times. It will launch from the former Soyuz pad in Kourou, [ASME](https://asme.org/Topics-Resources/Content/European-Rocket-Makers-Aim-for-Reusability) writes. Spain's PLD Space wants to recover Miura 5's first stage using a combination of engines and parachutes.

On 27 August ESA handed out €543.6m under the European Launcher Challenge: €197.8m to Isar Aerospace, €186.9m to RFA and €158.9m to PLD Space. That is according to [European Spaceflight](https://europeanspaceflight.com/esa-awards-e544-million-in-european-launcher-challenge-contracts/). The condition is that the companies reach orbit by the end of 2027, and the terms contain no reusability requirement. Isar has already done it: on 5 September its Spectrum rocket reached orbit from Andøya in Norway. It is the first rocket to put a payload into orbit from Western European territory, according to [Space.com](http://Space.com). Spectrum is expendable, of course. The contract with MaiaSpace, whose rocket is designed around a reusable stage, is still being negotiated.

### Who pays for Europe's rockets of the future

| Who pays | To whom | For what | How much | Source |
| --- | --- | --- | --- | --- |
| ESA | ArianeGroup | Themis T1E and Prometheus development (2024 contract) | €230m | [European Spaceflight](https://europeanspaceflight.com/esa-waiting-for-spring-to-launch-themis-reusable-rocket-demonstrator/) |
| EU (Horizon Europe) | SALTO consortium led by ArianeGroup, 25 partners from 12 countries | Themis testing at Esrange | Not stated | [SSC](https://sscspace.com/news/european-prototype-themis-completes-second-wet-dress-rehearsal-at-esrange/) |
| ESA (European Launcher Challenge) | Isar Aerospace | Launches to 2030 and rocket upgrades | €197.8m | [European Spaceflight](https://europeanspaceflight.com/esa-awards-e544-million-in-european-launcher-challenge-contracts/) |
| ESA (European Launcher Challenge) | RFA | Launches to 2030 and rocket upgrades | €186.9m | [European Spaceflight](https://europeanspaceflight.com/esa-awards-e544-million-in-european-launcher-challenge-contracts/) |
| ESA (European Launcher Challenge) | PLD Space | Launches to 2030 and rocket upgrades | €158.9m | [European Spaceflight](https://europeanspaceflight.com/esa-awards-e544-million-in-european-launcher-challenge-contracts/) |
| CNES, DLR and JAXA | The Callisto programme | Landing demonstrator, flights from Kourou | Not disclosed | [European Spaceflight](https://europeanspaceflight.com/cnes-call-reveals-inaugural-callisto-flight-test-pushed-to-2027/) |

European Launcher Challenge money pays for rockets that reach orbit, not rockets that come back. Only the first two rows and the last pay directly for reusability.

### Four numbers that explain reusability

| Number | What it means |
| --- | --- |
| \~4% | of Falcon 9's launch mass reaches orbit |
| \~0.3% | of the rocket's cost is propellant |
| \~60% | of the cost is the first stage |
| 37 | flights by a single booster (B1067, August 2026) |

### What I think

Landing a rocket is a solved problem. Paying for it isn't. Europe is paying for two jobs at once: prototypes learning to land, and rockets that first have to reach orbit at all. That is sensible, but not enough. A reusable stage makes sense if it flies every week, and Ariane 6's biggest customer today is Amazon, an American company building a rival to Starlink. IRIS², the EU's own constellation, will have around 290 satellites. That means a handful of launches, not a weekly schedule. My bet is that Europe will have a landing stage around 2030, and that it is more likely to come from a start-up than from ArianeGroup. Whether there will be anything to use it for depends on whether Europe becomes its own customer, as SpaceX did. Without such a customer, Themis risks becoming the most expensive exhibit in Kiruna.

---
Source: SpaceCAPEX — https://www.spacecapex.com/en/analysis/why-europe-is-losing-reusable-rocket-race