Superconducting logic for AI

Superconducting (SC) electronics are emerging as a compelling pathway toward energy-efficient, high-performance computing (HPC), artificial intelligence (AI), and quantum information systems. Their unique electromagnetic properties enable ultra-fast, low-dissipation logic and interconnects, offering a route to overcome the thermal and power constraints of conventional semiconductor technologies.

We present a NbTiN-based platform that integrates Josephson junctions, SC interconnects, capacitors, and engineered failure mode mitigation structures within a unified device stack. Fabricated using 193i lithography on 300 mm wafers, the platform enables compact, dense and scalable SC circuit fully compatible with existing CMOS infrastructure and provides a clear path towards large-scale manufacturing.  Building on this platform, we aim to demonstrate key functional circuits as a first step toward high-speed energy-efficient system-level validation.

The platform’s flexibility also supports technology co-integration and enables a broad range of applications with highly sensitive photon detectors, integrated photonics, and cryo-CMOS. These capabilities position NbTiN superconducting technology as a viable foundation for next-generation HPC and AI systems.