Cool new AI tech is coming to the rescue of old and strained power grids

The rapid growth of AI computing and data centers is creating unprecedented electricity demand, prompting innovation in power grid technology.

The core challenge: Power utilities face mounting pressure to deliver more electricity to data centers and other growing sources of demand, with distribution rather than generation emerging as the primary constraint.

• Data center operators and AI companies cite electricity access as their top operational concern
• Traditional grid infrastructure expansion is slow and expensive, creating opportunities for technological alternatives
• Growing adoption of electric vehicles and heat pumps adds to the strain on power grids

Innovative solution: Veir, Inc. is developing superconducting power cables that can transmit 10 times more electricity than conventional lines by using liquid nitrogen-cooled superconducting tapes.

• The technology replaces traditional metal cables with specialized pipes containing superconducting materials
• Veir’s system uses an evaporative cooling mechanism similar to human sweating, requiring less complex infrastructure than previous attempts
• The company has secured $75 million in new funding from investors including Microsoft, National Grid Partners, and Breakthrough Energy Ventures

Technical challenges and market considerations: While promising, superconducting cable technology faces several hurdles before widespread adoption.

• Previous experimental deployments proved expensive, with a 1-kilometer German project costing $14 million
• Maintaining extremely low temperatures (-321 Fahrenheit) requires specialized equipment and ongoing maintenance
• Veir hasn’t disclosed exact costs but suggests their solution will be competitive with premium cable projects exceeding $10 million per mile

Alternative approaches: Other companies are developing competing solutions to address grid capacity constraints.

• TS Conductor offers carbon fiber-core cables that can triple power capacity
• LineVision provides sensor-based “dynamic line rating” systems that optimize existing infrastructure
• LineVision’s UK project demonstrated 19% capacity increases and $17.5 million in savings for National Grid in 2022

Market uncertainties: Recent developments raise questions about future electricity demand from AI applications.

• Chinese company DeepSeek has demonstrated more energy-efficient AI model development
• Tech stock valuations have suffered following these efficiency breakthroughs
• Despite uncertainties, utilities continue investing in grid-enhancing technologies to adapt to changing energy landscapes

Looking ahead: Technology adoption outlook: While superconducting cables may remain a niche solution for specific high-demand applications, the broader trend toward grid-enhancing technologies appears robust as utilities adapt to evolving energy demands and distribution challenges. The success of these innovations will likely depend on their ability to deliver cost-effective solutions while meeting increasingly complex power delivery requirements.