Snowcap Compute Secures $23M in Funding to Develop AI Chips Using Superconducting Technology
Snowcap Compute, a pioneering startup focused on developing artificial intelligence (AI) computing chips using superconducting technology, has raised $23 million in its latest funding round, as reported by Reuters.
The round was led by Playground Global, with participation from Cambium Capital, Vsquared Ventures, and additional investors. As part of the development, Pat Gelsinger, former CEO of Intel, will join Snowcap’s board to support the company’s ambitious vision.
Building AI Chips with Superconductors: The Next Frontier
Snowcap's innovative approach seeks to develop computers capable of outperforming the best AI systems of today, all while using significantly less energy. The company is focusing on superconductors, materials that allow electricity to flow without resistance, as the foundation for its AI chips.
While the concept of superconductor-based chips has been theorized for decades, practical applications were hindered by the need for cryogenic cooling systems that consume substantial energy to maintain the low temperatures required for superconductivity.
However, the rapid demand for computing power, spurred by the rise of AI-driven applications like chatbots, and the limitations of conventional chips, have prompted a shift in perspective. Snowcap believes that with the right performance gains, the energy costs of cryogenic cooling can be justified.
AI Chips: 25 Times More Efficient Than Conventional Solutions
Snowcap’s superconductor-based chips are projected to be 25 times more efficient than current best-in-class chips, in terms of performance per watt. This breakthrough could have profound implications for the AI industry, where energy consumption is a growing concern, especially in data centers.
According to Reuters, the upcoming "Rubin Ultra" AI data center server, set to launch in 2027, is expected to consume around 600 kilowatts of power per hour. Operating this server at full capacity for just one hour would use two-thirds of the energy consumed by an average US household in an entire month.
This has highlighted the need for more energy-efficient solutions, and Snowcap’s chips could be a game-changer.
The Road Ahead: High-Performance AI Solutions in 2026
Snowcap’s development timeline includes the release of its first basic chip by the end of 2026. Although full systems are not expected until later, the company’s approach is already attracting attention from industry leaders, including Pat Gelsinger, who previously emphasized the need for a drastic shift in the computing industry’s energy consumption.
Gelsinger noted, "A lot of data centers today are just being limited by power availability.”
Also read: AMD Unveils MI400 AI Chips, Teams Up with OpenAI
Collaborative Efforts with Industry Experts
Reuters reports Snowcap’s founding team includes Anna Herr and Quentin Herr, experts in superconducting chips who have worked with companies like Imed and Northrop Grumman. The startup also draws expertise from former executives from Nvidia and Google, enhancing its technical capabilities in chip design and AI infrastructure.
The company’s chips will be manufactured in standard factories, but their production will require specialized materials such as niobium titanium nitride, which are sourced primarily from Brazil and Canada.
