Akash Systems brings diamond cooling to AI infrastructure

2026-06-10 · Source: AI – SiliconANGLE · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Emerging Technologies & Innovation · Depth: Intermediate, extended

Summary

Akash Systems Inc. offers a solution for GPU heat problems in AI infrastructure using lab-grown diamonds, which are the most thermally conductive material on Earth, five times faster than copper. Originating from space technology for satellite solar radiation management, their diamond-cooled servers reduce GPU temperatures by 10°C, enhancing efficiency and solving data center densification issues. The company targets neoclouds and enterprises, providing products for both air-cooled and liquid-cooled environments. Akash Systems claims a \$2 million cash impact over four years and a 50% token increase per server, enabling immediate capacity and avoiding significant data center construction costs. They deploy diamond-cooled servers with Nvidia and AMD GPUs, including H200, SMC 300, and Mitech AMD 350 models.

Key takeaway

For AI Architects and MLOps Engineers facing GPU thermal constraints and densification challenges, consider Akash Systems' diamond-cooled solutions. Implementing this technology can yield a 50% token increase per server and a \$2 million cash impact over four years, allowing you to maximize existing air-cooled data center capacity and defer costly new infrastructure builds. Evaluate their retrofit options for current GPU deployments to immediately enhance performance and efficiency.

Key insights

Lab-grown diamonds offer superior thermal conductivity, significantly improving AI GPU efficiency and data center capacity.

Principles

• Diamond is 5x more thermally conductive than copper.
• Efficient heat removal lowers power consumption.
• Solving heat issues enables higher server densification.

Method

Akash Systems integrates lab-grown diamonds into GPU cooling systems, either in new diamond-cooled servers or as retrofits for existing hardware, to rapidly transfer heat away from components.

In practice

• Upgrade existing air-cooled data centers for more power.
• Deploy diamond-cooled servers for 50% token increase.
• Run data centers at 95°F instead of 75°F.

Topics

• AI Infrastructure
• GPU Cooling
• Lab-Grown Diamonds
• Thermal Management
• Data Center Efficiency
• NVIDIA GPUs
• AMD GPUs

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

Related on AIssential

• Israeli AI cooling startup ZutaCore raises $100 million Series C at estimated $600 million valuation - CTech — ZutaCore's HyperCool offers a waterless, two-phase direct-to-chip liquid cooling solution for high-density AI data centers.
• Why Everyone Is Moving Away from NVIDIA — Hyperscalers are vertically integrating AI infrastructure, developing custom silicon and energy solutions to overcome GPU bottlenecks.
• Orbital Industries secures $50M to scale data centre infrastructure systems — AI accelerates industrial hardware development, enabling faster materials discovery and deployment for high-density computing.
• The NVIDIA Rubin Platform: Six New Chips, One AI Supercomputer — The Vera Rubin architecture integrates co-designed CPUs, GPUs, and DPUs for unprecedented AI supercomputing performance and scalability.
• Liquid-cooled AI systems expose the limits of traditional storage architecture — Hybrid air/liquid cooling for AI infrastructure is inefficient and creates thermal liabilities for storage.
• The $1 Trillion Data Center Crisis: Why America’s AI Ambitions Are Built on Crumbling… — Current data center infrastructure is unsustainable, demanding a radical redesign for AI and quantum computing's resource needs.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI – SiliconANGLE.