AI Data Centers Study High-Temperature Superconductors

Data Centers for AI Transforming Power Generation:

As the demand for energy among AI data centers surges, traditional power grids struggle to keep up, resulting in significant transmission and distribution losses, especially in some parts of the world. Hyperscalers like Amazon Web Services, Google Cloud, and Microsoft Azure seek more efficient ways to harness and distribute energy. Microsoft is promoting high-temperature superconductors (HTS) as a superior alternative to copper wiring, promising enhanced energy efficiency, reduced transmission losses, and decreased environmental impact. Utilizing materials like rare-earth barium copper oxide (REBCO), HTS cables operate with negligible electrical resistance, providing much higher current density in a more compact design. Though employing liquid nitrogen for cooling adds costs, the high efficiency and reduced footprint of HTS technology make it particularly valuable for space-constrained or highly demanding AI applications. Microsoft has invested $75 million in Veir, a company developing superconducting cables, aiming to enhance system-level value through improved scalability and reduced resistive losses.

Key Points:

Significant power inefficiencies in existing grids can’t support the growing energy needs of AI data centers.
Hyperscalers like Microsoft are investigating high-temperature superconductors (HTS) as an alternative to conventional copper wiring.
HTS technology promises major efficiencies in space, power density, and heat reduction, though at a higher initial cost.
Use of HTS is viable where space and power delivery constraints are major issues.
Microsoft, in collaboration with Veir, is working to validate and deploy HTS technology in AI data centers.