Snowcap Compute launched with $23M seed funding led by Playground Global to develop superconducting...
The AMW Read
The article introduces a new player in the AI infrastructure segment targeting the power bottleneck through post-CMOS superconducting architectures, representing a significant structural shift in silicon design (cross.§H).
NoveltySignificance
AI Infra · Player MapSilicon Substrate
Snowcap Compute launched with $23M seed funding led by Playground Global to develop superconducting AI chips that promise 25x better performance-per-watt than current systems. Former Intel CEO Pat Gelsinger joins as board chair, marking his first public investment since joining Playground. With Nvidia's 2027 Rubin Ultra projected to consume 600kW, Snowcap's approach using Josephson junctions instead of transistors addresses the critical power bottleneck facing AI data centers. The company plans to release its first chip by end of 2026, manufactured in conventional fabs using niobium titanium nitride. This signals a significant shift toward post-CMOS computing architectures for sustainable AI infrastructure.