Phononic, known for its solid-state cooling technology, is expanding its range of advanced cooling solutions for networking, GPUs, and AI data centres. The company, with deployments among hyperscalers, has introduced the qualified and deployed 1.6T HVM solution, strengthening its presence in the evolving data centre thermal ecosystem.
The latest innovations include a GPU HBM cooling solution that seeks to improve energy efficiency, hardware lifespan, and computing performance while supporting strong return on investment. As demand grows for higher bandwidth, faster performance, and greater energy efficiency, Phononic’s solutions aim to address key challenges in high-performance computing environments.
Key features include:
- Next-Gen HBM4 Aligned GPU HBM Cooling Solutions: Designed for up to 75% higher heat dissipation, supporting sustained GPU performance and reducing thermal-induced throttling.
- High-Performance Pluggable Optics for 1.6T and Beyond: Expanded capability to handle approximately 50% greater heat loads without increasing power consumption.
- CPO-Ready Thermal Kit: Provides targeted cooling and packaging optimised for co-packaged optical engines in scalable deployments.
Phononic’s development in CPO cooling aims to address the thermal challenges posed by external laser sources and aligns with the industry shift toward co-packaged optics. This approach seeks to reduce signal-path distances and associated energy consumption while supporting the scale and bandwidth density required for next-generation networks.
The Gen 2 GPU HBM Cooling Solution integrates Phononic's Thermal Kit as a central subsystem. With up to 75% improved cooling, the system helps maintain GPU performance, reduces thermal throttling, and enhances cluster stability. Designed for seamless integration into existing data centre hardware, it allows operators to adjust energy efficiency, computing output, and component longevity across the installation.
Phononic’s solutions aim to provide precise, node-level thermal management for GPU and AI workloads, supporting efficiency, performance, and infrastructure reliability in high-density environments.
