SnowCap Compute is an AI startup that works on next-generation computing chips built with superconductors, raising $23 million in funding and earning a high-profile addition to the board: former Intel CEO Pat Gelsinger. Cambium Capital and Vsquared Ventures also participated in the funding round.
The California-based startup is building AI chips that run on superconducting technology. It could make them far more energy efficient than today’s chips if they can make it work on a large scale.
Snowcap goal? Building computers that can surpass top-class AI systems while using only a small portion of electricity. That’s a big claim, but it’s rooted in real physics. Superconductors are nothing new. Scientists have been studying them for decades, and making practical chips from them has always hit a wall. And that usually means a cooling system that is hungered for a lot of power.
However, the AI boom changed the equation. Chatbots and large language models are pushing chip makers to limits, and energy demand is beginning to surpass what the power grid can offer. When efficiency is more important than ever, revisiting superconductor technology suddenly makes more sense.
Nvidia’s upcoming “Rubin Ultra” AI Server is expected to draw around 600 kilowatts of electricity, expected to be in 2027. If you drive in just an hour with full tilt, two-thirds of electricity will burn out the average US household in a month, Reuters reported.
Such mathematics opens the door to alternatives, especially when performance is improved, when it outweighs the cost of keeping the chip extremely low and cool. Snowcap CEO Michael Lafferty says he’s on track to do just that. He claims the company’s chips offer 25 times better performance than today’s best processors, even after factoring them out with cooling overhead.
“The power (efficiency) is great, but the performance sells,” Rafferty said. “So we push up our performance levels and at the same time we’re reducing our power.”
Lafferty is nothing new for futuristic hardware. Before SnowCap, he led Advanced Chip R&D at Cadence Design Systems. The founding team also includes Anna Herr and Quentin Herr, who previously worked on the superconducting chips of IMED and defense contractor Northrop Grumman, as well as former chip executives from Nvidia and Google.
The chip itself does not require exotic manufacturing. They can be produced in standard factories. However, the materials are a different story. To build these chips, Snowcap uses the Niobium Titanium Titanium nitride. This is a special metal that relies heavily on Brazilian and Canadian suppliers.
The basic version of the SnowCap chip is expected by the end of 2026. The full system will appear later.
Gelsinger, who led investments through Playground Global and currently on Snowcap’s board of directors, said the industry has reached a turning point. AI is chewing on electricity, and electricity is quickly becoming a bottleneck in data centers.
“Many data centers today are limited by power availability,” he said.
SnowCap is still early, but the 2025 bet feels very 2025. A smarter chip, less power, and shots that overtake the Giants.
🚀Want to share the story?
Submit your stories to TechStartUps.com in front of thousands of founders, investors, PE companies, tech executives, decision makers and tech leaders.
Please attract attention
Source link
