Snowcap Compute, a startup dedicated to developing artificial intelligence computing chips through superconducting technology, announced that it has successfully secured $23 million in funding, along with the news that the former CEO of Intel will be joining its board.

Snowcap's objective is to create computers that have the potential to outperform the best artificial intelligence systems available today while consuming significantly less electricity. To achieve this, Snowcap intends to utilize a novel type of chip constructed from superconductors, which are materials that enable the flow of current without any electrical resistance.

Researchers have a thorough understanding of superconductors and have speculated about their use in computer chips since at least the 1990s; however, they have encountered a significant obstacle: for these chips to function, they must be maintained at extremely low temperatures within cryogenic coolers, which in turn require a large amount of electricity to operate.

Also Read: The Quantum Computing Age is Upon US Sooner than Expected

For many years, this made the concept of superconductor chips impractical, until the rise of AI chatbots created a substantial demand for computing power, coinciding with the limitations of conventional chips in terms of maximizing performance from electrical power and straining electrical grids.

For instance, Nvidia's upcoming "Rubin Ultra" AI data center server, expected in 2027, is anticipated to use around 600 kilowatts of power. This implies that running that single server at its maximum capacity for one hour would consume approximately two-thirds of the average energy that a typical US household uses in an entire month.

The founding team of Snowcap comprises two scientists—Anna Herr and Quentin Herr—who have substantial experience working with superconducting chips at the chip research company Imed and the defense contractor Northrop Grumman, along with former chip executives from Nvidia and Google, part of Alphabet.

Also Read: Disney, NVIDIA & Google DeepMind Usher Expressive Robotics Revolution

Although these chips can be produced in a conventional factory, they will require a specialized metal called niobium titanium nitride, which Lafferty mentioned relies on Brazil and Canada for essential components.

Snowcap aims to develop its first basic chip by the end of 2026, though complete systems will not be available until a later date.