A Bit of Data Center Heat Can Be Turned Back Into Electricity

· Source: IEEE Spectrum · Field: Technology & Digital — Cloud Computing & IT Infrastructure, Emerging Technologies & Innovation, Robotics & Autonomous Systems · Depth: Intermediate, short

Summary

PyroDelta Energy, a subsidiary of First Tellurium, is developing innovative thermoelectric generators to convert waste heat from AI data centers, engines, and drones into usable electricity. Current thermoelectric materials, primarily bismuth telluride, are limited by high manufacturing waste, brittleness, and restricted shapes. PyroDelta's founder, Michael Abdelmaseh, has devised a capillary casting method that grows bismuth telluride crystals directly into desired shapes, reducing material waste by 60-80% and increasing durability by approximately 10 times. This method allows for curved designs, such as rings that fit around liquid cooling pipes. Prototypes include an energy harvester for data centers, capable of powering sensors and cameras, and a thermoelectric car radiator that could improve internal combustion engine efficiency by 5%. The company is also applying this technology to drones for the DARPA Lift Challenge.

Key takeaway

For CTOs and VPs of Engineering exploring sustainable data center operations, PyroDelta's thermoelectric technology offers a novel approach to energy recapture. Your teams should investigate integrating these advanced thermoelectric generators into existing liquid cooling infrastructures to power auxiliary systems like sensors and cameras, potentially reducing operational costs and enhancing energy efficiency without replacing primary cooling methods.

Key insights

PyroDelta's capillary casting method significantly improves thermoelectric generator efficiency, durability, and form factor versatility.

Principles

Method

PyroDelta uses a capillary casting method to grow bismuth telluride crystals directly into specific shapes within molds, eliminating sawing waste and improving material properties for thermoelectric generators.

In practice

Topics

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Hardware Engineer, AI Architect, Research Scientist

Related on AIssential

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by IEEE Spectrum.