This article is based on an official press release from the European Space Agency (ESA).

ESA and MT Aerospace Deploy AI to Slash Rocket Inspection Times by 95%

The European Space Agency (ESA) has announced a significant leap forward in the manufacturing of launch vehicles, revealing that the integration of artificial intelligence (AI) into its production lines has drastically reduced quality assurance timelines. In a statement released on January 21, 2026, ESA detailed how its collaboration with German manufacturing partner MT Aerospace has successfully applied machine learning to the production of the Ariane 6 rocket.

The initiative, conducted under ESA’s Future Launchers Preparatory Programme (FLPP), focuses on automating the complex analysis of metal forming and welding. According to the agency, the most immediate impact has been observed in the inspection of friction stir welds, where the introduction of AI has cut analysis time by 95% compared to traditional manual methods.

By shifting from labor-intensive human inspection to data-driven algorithmic monitoring, ESA aims to increase production rates and reduce costs, critical factors in an increasingly competitive global launch market.

Revolutionizing Friction Stir Welding

The core of this manufacturing update centers on Friction Stir Welding (FSW), a solid-state joining technique used to construct the massive fuel tanks for the Ariane 6. Unlike traditional welding, which melts materials to fuse them, FSW uses a rotating pin to generate friction and heat, joining metals without reaching their melting point. While this produces exceptionally strong joints, verifying their integrity has historically required time-consuming analysis.

Under the new system, machine learning algorithms monitor digital telemetry directly from the welding equipment. This includes data points such as weld force, torque, and temperature. The system processes this data to automatically verify the shape and quality of the final weld seam.

Daniel Chipping, ESA Project Manager for Software-Centred and Digitalisation Activities, highlighted the operational impact of this technology:

“Artificial intelligence, such as machine learning, in combination with new digital technologies is transforming launcher manufacturing… from automating complex analysis tasks to reducing tedious machine stop-starts, we are starting to see the benefits across all materials and shaping processes.”

, Daniel Chipping, ESA Project Manager (FLPP)

Precision in Shot Peen Forming

Beyond welding, the initiative has applied AI to “shot peen forming,” a chaotic process used to shape the dome heads of the Ariane 6 fuel tanks. This technique involves blasting metal sheets with small spherical shots to bend them into specific curves without applying heat, which preserves the material’s structural integrity.

Predicting the Unpredictable

Historically, shot peening has been difficult to model precisely because the impact of thousands of individual shots is physically unpredictable. This often necessitated a trial-and-error approach to achieve the correct geometry. ESA reports that MT Aerospace has now trained machine learning models to predict exactly how the metal will deform under specific bombardment patterns.

This predictive capability allows manufacturers to achieve the desired dome shape with a tolerance of just 2 millimeters, significantly reducing the time required to set up and calibrate the machinery.

Advancing Carbon-Fibre Composites

The FLPP initiative also extends to the “Phoebus” project, a collaboration aimed at replacing heavy metallic upper-stage tanks with lightweight carbon-fibre reinforced plastic (CFRP). Reducing the mass of the upper stage is a priority for ESA, as every kilogram saved on the structure translates to additional payload capacity.

In this application, laser sensors combined with machine learning models are used to detect and classify manufacturing defects “on the fly” during the automated fibre placement process. By identifying issues immediately as layers are applied, the system prevents long production stoppages associated with manual checks, streamlining the fabrication of these complex composite parts.

AirPro News Analysis

The integration of AI into the Ariane 6 supply chain represents a necessary evolution for the European space sector. While new entrants like Relativity Space have garnered headlines for 3D-printing entire rockets, ESA’s approach demonstrates how legacy manufacturers can modernize established industrial processes to achieve similar efficiency gains.

The 95% reduction in weld analysis time is more than a technical statistic; it addresses a primary bottleneck in rocket production. In an era where launch cadence is dictated by how quickly vehicles can roll off the assembly line, removing manual “stop-starts” is essential for Ariane 6 to meet its commercial and institutional targets. By validating these technologies through the FLPP, ESA is effectively de-risking the transition to a more automated, data-centric future for European aerospace.

Sources

Sources: ESA (Primary Source)