Interspectral has been selected as a key partner in a multi-million SEK research initiative awarded by Sweden’s national innovation agency.
Vinnova has elected Interspectral, along with Saab, AMEXCI and Scaleout Systems, to develop the next generation of AI-driven quality assurance for additive manufacturing in aerospace, defence and other safety-critical industries.
They will carry out this work via the Trusted Federated Intelligence for Additive Manufacturing (TRUSTAM) consortium.
Vinnova believes that for additive manufacturing to fulfil its potential in sectors like aerospace and defence, quality must be ‘embedded, traceable and provable in real-time’ across facilities, machines and highly sensitive environments. Aligning with this belief, Interspectral deemed the TRUSTAM partnership a natural opportunity to pursue.
“We entered this collaboration because the challenge it addresses is one we encounter with our customers every day,” said Isabelle Hachette, CEO of Interspectral. “How do you scale AI-driven quality assurance across multiple production sites and different machine environments without ever compromising data security or IP ownership? That question demands a collaborative answer, and this consortium is uniquely positioned to deliver it.”
TRUSTAM will apply federated learning to quality assurance in AM, a framework where AI models improve collectively across production environments without raw data ever leaving the site where it was generated. Only model updates are exchanged to preserve full data confidentiality.
Interspectral will take a central technical role whereby it will lead the development of the local AI model and build it directly into its AM Explorer platform. With this integration, the hope is that TRUSTAM will enable on-premise AI models that are fine-tuned to specific machines and production conditions. The consortium also aims to develop a validated framework for secure cross-site AI collaboration, while real-world demonstrators in live aerospace and defence environments will help to prove out the technology’s capabilities.
The project will run through early 2028 – culminating in a demonstrator phase and full dissemination of results to the broader additive manufacturing community – and accelerate planned capability enhancements in the AM Explorer Platform.
