On December 4–5, 2025, the MultiSpin.AI consortium held its two-year meeting at INESC-ID in Lisbon, Portugal, gathering partners from across Europe to review achievements, align on technical directions, and define the priorities for the next phase of the project.
Over the past two years, MultiSpin.AI has been working toward a new generation of energy-efficient AI hardware based on n-ary spintronic technologies. The Lisbon meeting provided an important opportunity for partners to present results, exchange insights, and strengthen collaboration across the full technical stack — from materials and device fabrication to algorithms, benchmarking, and use-case integration.
Key Highlights of the Meeting
During the two-day , each work package shared its main achievements, demonstrating significant progress toward the project’s objectives. Discussions focused on:
- Hardware and device design: advancements in spintronic memory elements and early integration steps for multi-level devices.
- Benchmarking and modeling: development of methodologies to evaluate the performance, reliability, and energy efficiency of the emerging technology.
- Algorithmic and software integration: progress in adapting AI workflows to the characteristics of n-ary devices, enabling efficient on-chip computation.
- Use-case alignment: coordination with application partners to ensure that system requirements and expected performance match real-world needs.
- Communication and dissemination: the consortium jointly refined the project’s outreach strategy to increase visibility, strengthen stakeholder engagement, and showcase results as they mature.
Looking Ahead
The meeting concluded with a clear roadmap for the months ahead. Partners will continue working in close collaboration to:
- finalise and optimise the design of key device and circuit components,
- advance benchmarking activities to validate performance gains,
- strengthen integration with selected use cases,
- and expand dissemination and communication efforts to share progress with the research community, industry stakeholders, and the public.
Acknowledgements
The consortium extends its gratitude to INESC-ID for hosting the meeting and to all partners for their active participation and strong commitment to driving innovation in next-generation AI hardware.
As MultiSpin.AI enters the second half of the project, the team remains focused on delivering impactful results and demonstrating the potential of spintronic technologies to reshape the future of sustainable, low-power artificial intelligence.
