HardwareJun 22, 2026, 5:00 AM

NVIDIA Unveils 45°C Liquid Cooling for AI Servers

TickrWire Editorial Desk·Jun 22, 2026, 5:00 AM·1 min read AI-assisted, human-reviewed

Reported by NVIDIA AI Blog: Hotter Than a Hot Tub: The 45°C Breakthrough to Cool AI’s Biggest Machines. Analysis and context written by TickrWire.

30-second summary

NVIDIA announces AI servers capable of operating with liquid cooling at 45°C, improving energy efficiency by allowing higher operating temperatures.

TickrWire
NVIDIA Unveils 45°C Liquid Cooling for AI Servers
Key takeaways
  • NVIDIA’s new AI servers support liquid cooling at 45°C, a 5°C increase over conventional limits.
  • Higher operating temperatures reduce cooling energy demands, potentially cutting data center power usage by up to 50%.
  • The technology targets 'AI factories'—large-scale AI infrastructure where cooling is a major cost driver.
  • NVIDIA’s announcement comes from its official blog, indicating a direct product update rather than third-party speculation.
  • The innovation aligns with broader industry trends toward energy-efficient AI hardware.
Full story

NVIDIA has unveiled a breakthrough in liquid cooling technology for its AI servers, enabling them to run at up to 45°C (113°F) without performance degradation. This higher temperature threshold reduces the need for aggressive cooling, leading to significant energy savings in data centers. The innovation targets 'AI factories'—large-scale AI training and inference facilities—where cooling costs are a major operational expense. By tolerating warmer liquid temperatures, NVIDIA claims its systems can cut cooling power consumption by up to 50% while maintaining reliability.

Why this matters
Developers

Enables more efficient deployment of AI workloads in data centers, reducing operational costs and environmental impact.

Businesses

Lowers total cost of ownership for AI infrastructure, particularly for companies running large-scale models.

Investors

Signals progress in energy-efficient AI hardware, a key differentiator for data center operators and cloud providers.

Students

Demonstrates how thermal management innovations can improve AI system sustainability and performance.

Everyone

Highlights advancements in making AI infrastructure more sustainable and cost-effective for end-users.

Glossary
AI factories
Large-scale data centers dedicated to training and running AI models.
liquid cooling
A method of removing heat from computer hardware using circulating liquid, more efficient than air cooling.
data center
Facility housing servers and computing infrastructure for processing and storing data.
operational cost
Recurring expenses associated with running a business or infrastructure, including energy use.

AI bias estimate: Source is NVIDIA’s official blog; likely promotional but factually specific and credible. (Automated estimate, not a definitive judgement.)

Sources · 1
Read next
More stories