September 17, 2025

Why We Invested in Mueon: Building Data Center Scale AI Infrastructure

Written by
Neb Mela & Rayfe Gaspar-Asaoka

At Geodesic Capital, we look for exceptional teams tackling fundamental bottlenecks in technology. Our investment in Mueon, which emerged from stealth with a $15.5 million seed round, reflects this focus. Mueon is pioneering Data Center Scale infrastructure designed to solve the most pressing constraints in scaling artificial intelligence: power delivery, thermal management, and memory-compute integration.

The rapid growth of AI models has exposed hard physical limits in existing server and interconnect architectures. Bandwidth between compute and memory is increasingly inadequate, power delivery efficiency cannot keep pace with rising density, and thermal management has become a gating factor for reliability. These are not incremental engineering issues but structural constraints that cap performance, cost-effectiveness, and scalability for next-generation AI workloads.

Mueon is taking a fundamentally different approach. Rather than making piecemeal improvements at the chip, board, or rack level, the team is rethinking system design from the ground up. Their architecture introduces modular, cube-based wafer and panel scale systems that tightly integrate compute and memory. This reduces interconnect complexity, increases bandwidth, and improves both thermal and power density.

Our conviction to invest came from the depth and credibility of the team. Wilfred Gomes and Jack Hwang and the Mueon team of industry experts have spent their careers building high-performance systems where success depended on solving the hardest integration problems of Memory, power, cooling, and IO. Their track record spans large-scale computing environments and advanced packaging programs, giving them the technical foundation to deliver on Mueon’s ambitious vision.

Equally important is the alignment with where the industry is heading. Advanced packaging, 3D stacking, and Memory architecture are rapidly becoming central to next-generation compute. Mueon’s architecture is designed with these trends in mind, positioning the company to play a key role as adoption accelerates. In addition, Mueon’s early collaborations with partners across semiconductors and AI infrastructure show that demand for system-level solutions is real and already taking shape in the market.

Looking ahead, AI models will continue to grow in size and complexity, pushing existing server architectures beyond their limits. Meeting this demand requires more than incremental chip improvements; it calls for a rethinking of how compute, memory, power, and cooling are integrated. Mueon’s system-level approach is designed for this shift, creating infrastructure capable of sustaining wafer-scale performance at scale. The opportunity is significant, and we believe Mueon is well positioned to play a central role in building the foundation for the next generation of AI.

At Geodesic, we are excited to partner with Mueon as they begin this next chapter. Their vision, technical depth, and system-first approach embody the type of innovation we believe will define the future of computing.