AI Data Centers: Superconductors Could Solve Power & Efficiency Challenges

The relentless demand for computing power, particularly from artificial intelligence workloads, is pushing data center infrastructure to its limits. Traditional power delivery systems are struggling to keep pace, facing constraints in both capacity and efficiency. Now, hyperscalers like Microsoft are turning to a potentially transformative technology: high-temperature superconductors (HTS). These materials promise to dramatically reduce energy loss during transmission, enabling more power to be delivered in a smaller footprint.

The problem is straightforward: existing power grids and the internal distribution networks within data centers lose a significant amount of energy as heat during transmission. According to the U.S. Energy Information Administration, average transmission and distribution losses are around 5 percent annually, a figure that can be considerably higher in some regions. For AI data centers, where power density is already a critical challenge, even small percentage losses translate into substantial wasted energy and increased operational costs. The sheer scale of new data center builds is also exacerbating the problem, as there simply isn’t enough existing grid capacity to meet the growing demand.

Microsoft is actively exploring HTS as a solution, viewing it as a way to improve energy efficiency, enhance grid resiliency, and minimize the physical space required for power infrastructure. “Because superconductors take up less space to move large amounts of power, they could help us build cleaner, more compact systems,” explained Alastair Speirs, general manager of global infrastructure at Microsoft, in a blog post. The core principle behind this improvement lies in the fundamental difference between how copper and superconductors conduct electricity.

Copper, while a good conductor, inherently resists the flow of current, generating heat as a byproduct. This resistance limits the amount of current that can be transmitted and reduces overall efficiency. Superconducting materials, when cooled to cryogenic temperatures, exhibit almost zero electrical resistance. This allows for significantly higher current densities in a much smaller space, minimizing energy loss and voltage drop. While termed “high-temperature,” these superconductors still require cooling, but to temperatures significantly warmer than those needed for traditional superconductors.

The benefits extend beyond efficiency. HTS cables are smaller and lighter than their copper counterparts, offering greater flexibility in data center design and potentially reducing the need for numerous substations. Microsoft estimates that next-generation superconducting transmission lines can deliver an order of magnitude more capacity than conventional lines at the same voltage level. This increased capacity is crucial for supporting the ever-growing power demands of AI workloads.

To accelerate the development and deployment of this technology, Microsoft has invested $75 million into Veir, a company specializing in superconducting power technology. Veir’s conductors utilize HTS tape, primarily based on rare-earth barium copper oxide (REBCO), a ceramic superconducting material deposited as a thin film on a metal substrate. “The key distinction from copper or aluminum is that, at operating temperature, the superconducting layer carries current with almost no electrical resistance, enabling very high current density in a much more compact form factor,” stated Tim Heidel, CEO and co-founder of Veir.

A significant engineering challenge lies in maintaining the cryogenic temperatures required for superconductivity. Veir addresses this with a closed-loop liquid nitrogen system, circulating the coolant through the length of the cable, re-cooling it, and recirculating it. According to Heidel, liquid nitrogen is a readily available, cost-effective, and safe material with established industrial applications. Veir favors external cooling systems, feeding liquid nitrogen into the data center to minimize the internal footprint and operational complexity.

However, HTS isn’t a universal replacement for copper. The costs associated with the rare earth materials, cooling infrastructure, and cryogenic temperatures are substantial. Heidel emphasizes that the economics are most compelling in scenarios where power delivery is constrained by space, weight, voltage drop, and heat. “In those cases, the value shows up at the system level: smaller footprints, reduced resistive losses, and more flexibility in how you route power,” he explained. As manufacturing processes improve and volumes increase, the cost of HTS is expected to decrease, making it more viable for a wider range of applications.

AI data centers are emerging as the ideal proving ground for HTS technology. Hyperscalers are willing to invest in innovative solutions to overcome the power and efficiency challenges posed by these demanding workloads. Microsoft’s Husam Alissa, director of systems technology, highlighted the company’s focus on validating and derisking the technology with partners, concentrating on systems design and integration. “HTS manufacturing has matured—particularly on the tape side—which improves cost and supply availability,” Alissa noted.

The transition to HTS-based power delivery systems represents a significant undertaking, requiring careful consideration of system design, integration, and operational procedures. However, the potential benefits – increased efficiency, reduced footprint, and enhanced grid resiliency – are substantial, positioning HTS as a critical component in the future of AI-powered computing infrastructure.