Rocket One Inc. has entered into a research agreement with Virginia Commonwealth University (VCU) aimed at developing SKYMRAM, a novel skyrmion-based magnetic memory technology intended for next-generation artificial intelligence (AI) chips and computing systems. The collaboration was announced on October 13, 2026, and is set to commence on November 1, 2026.
The agreement outlines a one-year research initiative focused on nanoscale fabrication, material optimization, and device testing. A key goal is to demonstrate the switching capabilities of SKYMRAM’s soft magnetic layer, which is essential for its intended applications.
Robb Knie, CEO of Rocket One, emphasized the importance of hardware advancements for next-gen AI computing. “This agreement establishes a concrete development path for SKYMRAM,” he stated, highlighting the need for rigorous testing and validation at the device level. He noted that the initiative aims to build a technical foundation for future chip architectures applicable to AI, space, and defense sectors.
Research Program Details
The SKYMRAM program will investigate skyrmions—nanoscale magnetic structures—as the basis for a new magnetic memory architecture. Key research activities will include:
- Nanoscale device fabrication: This involves patterning and etching magnetic material stacks intended for skyrmion hosting.
- Materials engineering: The project will refine thin-film structures to facilitate device testing.
- Switching studies: Researchers will explore the control of the soft magnetic layer, a central focus of the proof-of-concept objective.
- Device optimization: This will include efforts to create reliable switching mechanisms for skyrmions sized approximately 100 nanometers.
Jayasimha Atulasimha from VCU will lead the project. Additionally, a subaward will be made to Northwestern University to support thin-film growth and materials characterization, promoting a collaborative approach between material development and device testing.
The program includes quarterly goals that will guide the progression from initial material fabrication to device optimization. The outcomes are expected to provide evidence regarding the technical feasibility of SKYMRAM, laying the groundwork for potential future AI chip architectures.
Rocket One regards advanced magnetic memory as a key component of its long-term computing strategy, particularly for applications in AI, space, and defense. The agreement also allows Rocket One to negotiate licenses for any intellectual property developed during the research.
As the research progresses, the company will continue to evaluate the technology’s potential for integration into practical AI applications.
