Snowcap Compute Secures $23 Million to Develop Superconducting Chips for AI

Snowcap Compute, a U.S.-based startup working on superconducting technology for artificial intelligence, has raised $23 million in a funding round led by Playground Global. Other investors include Cambium Capital and Vsquared Ventures. The company also announced that former Intel CEO Pat Gelsinger will join its board, bringing decades of industry leadership to the ambitious project.

Snowcap is developing next-generation AI chips built using superconductors—materials that conduct electricity with zero resistance. Unlike conventional silicon-based chips, these superconducting processors promise significantly higher energy efficiency and computing power. According to the company, their technology could deliver up to 25 times more performance per watt compared to leading current chips, even after accounting for the energy needed for cryogenic cooling systems.

While cryogenic cooling has traditionally been seen as a barrier due to its high energy requirements, Snowcap’s CEO Michael Lafferty argues that the trade-off now makes sense. As AI workloads place increasing demands on global power infrastructure, using some of that energy to cool superconducting chips becomes a viable option if it leads to massive performance improvements.

Founded by engineers with experience at Northrop Grumman, Nvidia, Google, and imec, the startup is building its chips using a superconducting alloy called niobium titanium nitride, sourced from Brazil and Canada. The company expects to produce its first chip prototype by the end of 2026, with full systems rolling out shortly after.

Snowcap’s approach could offer a much-needed solution to the growing energy challenges in data centers, especially as AI hardware like Nvidia’s future “Rubin Ultra” servers are expected to consume massive amounts of electricity. Gelsinger emphasized that the semiconductor industry must pivot toward radically more efficient architectures to keep pace with AI’s rapid growth.