In a major step toward redefining the future of artificial intelligence hardware, Snowcap Compute, a startup pioneering superconducting AI chips, has secured $23 million in funding. The company aims to build ultra-efficient computers that can outperform today’s top AI systems while using only a fraction of the electricity.
Snowcap plans to achieve this with chips made from superconductors—materials that allow electrical current to flow without resistance. These chips promise massive leaps in energy efficiency, addressing one of the most pressing issues in modern computing: power consumption.
The company also announced a high-profile addition to its board—Pat Gelsinger, former Intel CEO, who led the investment via Playground Global. His involvement signals strong confidence in Snowcap’s disruptive potential.
Why Superconductors Matter in AI
The concept of superconducting chips has been around since the 1990s, but they’ve long been sidelined due to the requirement of cryogenic cooling, which uses large amounts of power. However, the explosive demand for AI processing—especially from large language models and generative AI—has changed the game.
Modern AI servers are power-hungry. For instance, Nvidia’s upcoming Rubin Ultra server, slated for release in 2027, could consume around 600 kilowatts—enough energy to power a U.S. home for two-thirds of a month in just one hour. With energy limits now a real bottleneck, Snowcap believes its superconducting chips—despite cooling requirements—could be 25 times more efficient per watt than current best-in-class solutions.
“Power efficiency is nice, but performance is what sells,” said CEO Mike Lafferty, who previously led chip innovation at Cadence Design Systems. “We’re pushing performance through the roof and pulling power consumption down.”
A Team of Elite Engineers
Snowcap’s leadership features Anna Herr and Quentin Herr, two leading scientists with decades of experience in superconducting technologies from firms like Imec and Northrop Grumman. They’re joined by former executives from Nvidia and Google, bringing together deep technical expertise and market knowledge.
Their first prototype chip is expected by the end of 2026, although complete systems will likely roll out later. The chips will use an exotic material called niobium titanium nitride, with critical components sourced from Brazil and Canada.
Industry Powerhouse Support
Beyond Playground Global, the funding round attracted support from Cambium Capital and Vsquared Ventures. Gelsinger emphasized that without a radical change, data centers will soon be paralyzed by power limitations.
“Today, many data centers are already hitting power ceilings,” Gelsinger noted. “This kind of tech is essential if we want to keep scaling.”
Snowcap Compute’s mission is clear: create the next generation of AI chips that marry performance and sustainability, helping the tech industry break free from
