AI Co-Processor
Research Project
MultiSpin.AI is a EU-funded research project that aims to revolutionise neuromorphic computing by developing and fabricating a highly efficient n-ary spintronic based edge computing co-processor capable of performing AI algorithms like Deep Learning at unprecedented speeds while drastically reducing energy consumption, ultimately enabling transformative applications like autonomous vehicles, robots, and medical devices, and contributing to the EU chip industry and reducing CO2 emissions from AI inference.
01
Reduction of energy consumption
We develop an energy-efficient co-processor to tackle the rapidly growing energy demands linked to data processing, envisioning a future where sustainability challenges are seamlessly addressed.
02
Ethical considerations
At the core of our work lies an ethical framework encompassing environmental stewardship, equity, human rights, and social responsibility, , guiding our technology development for positive societal impact.
03
Unprecedented speeds
We are surpassing conventional speed thresholds with advancements exceeding tenfold, forging a path towards unprecedented data processing velocities.
Spintronics & AI Integration
Get to know MultiSpin.AI’s approach to the integration of spintronic hardware and artificial intelligence. Browse through the four main project steps to understand our innovative technological project.
Expertise Areas
Together with the leading research institutes, universities, and SMEs, MultiSpin.AI brings interdisciplinary competencies of research to commercialization.
Innovation Impact
The MultiSpin.AI initiative is poised to strengthen research in spintronics and artifical intelligence and be ready for commercialization. Go through our Innovation Impact section to understand how we plan to do it.
Publications
Read our latest research outputs as part of our commitment to innovating the field of neuromorphic artificial intelligence.
Need more information?
Timeline of the project
Check out our detailed timeline for the mapping out our initiative.
● Design and fabrication of the MultiSpin.AI post-von Neumann architecture
● Characterization and Testing of SLMMS and M2TJs
● AI algorithm design, validation and specs
● Proof of Concept Demonstration (PoC)
December 2024
Design and fabrication of the MultiSpin.AI Post-Van Neumann architecture
First Single-Layer Multistate Nanomagnetic device
June 2025
Characterization and Testing of SLMMS and M2TJs
Demonstration of 16+ discrete magnetic states in SLMMS and 8-level M2TJs.
December 2025
AI algorithm design, validation and specs
MAC operation benchmarking
June 2026
Design and fabrication of the MultiSpin.AI Post-Van Neumann architecture
Qualification of nanofabrication process for a 4x4 array of 4-level M2TJ devices
June 2026
AI algorithm design, validation and specs
Production of a simplified task database for the PoC
June 2026
Proof of Concept Demonstration (PoC)
Operational demonstration of PCB with spintronic crossbar
September 2026
Characterization and Testing of SLMMS and M2TJs
Showcasing high-speed read/write of a 16-level M2TJ demonstration
December 2026
Proof of Concept Demonstration (PoC)
Device benchmark
Upcoming events
Join us in our efforts to revolutionize neuromorphic computing. Browse through our upcoming events for more ways to engage.
Events
News
Keep up with the latest work and engagements of MultiSpin.AI.
- All Posts
- Updates
Want to learn more about us?