← Back
MIT News AI

Startup’s nuclear-inspired cooling system could make data centers more sustainable

6 min read
#compute#inference
Startup’s nuclear-inspired cooling system could make data centers more sustainable
Level:Intermediate
For:AI Engineers
TL;DR

Ferveret, a startup founded by Reza Azizian and Matteo Bucci, is developing a nuclear-inspired cooling system for data centers that uses a specialized liquid to absorb heat, reducing electricity usage and water consumption. The company's Adaptive Phase Cooling (APC) solution has shown a 15% improvement in computational power efficiency compared to state-of-the-art liquid cooling solutions. By combining APC with a power control system, Ferveret claims to enable data centers to generate 35% more tokens from their AI models with the same amount of power. This innovation has the potential to make data centers more sustainable and efficient. The practical implication for engineers building AI systems is that they can potentially reduce their energy consumption and increase their computational power efficiency by adopting Ferveret's cooling system.

⚡ Key Takeaways

  • Ferveret's APC solution leads to a 15% improvement in computational power efficiency compared to state-of-the-art liquid cooling solutions.
  • The company's cooling system uses a specialized liquid that absorbs heat more efficiently than air from a fan, with no water consumption.
  • Ferveret's system enables data centers to get 35% more tokens from their AI models with the same amount of power when combined with a power control system.
  • The startup is already testing its solutions with companies including CleanSpark, FuriosaAI, and Switch.
  • Ferveret's goal is to make data centers as sustainable as possible and help them use every single watt of power to generate tokens.
💡 Why It Matters

The development of Ferveret's cooling system has significant implications for the sustainability and efficiency of data centers, which are projected to account for 9-17% of total electricity usage in the U.S. by the end of the decade. By reducing energy consumption and increasing computational power efficiency, Ferveret's technology can help data centers generate more tokens from their AI models w

✅ Practical Steps

  1. Consider adopting Ferveret's Adaptive Phase Cooling (APC) solution to improve computational power efficiency and reduce energy consumption in data centers.
  2. Evaluate the potential benefits of combining APC with a power control system to optimize operating conditions and generate more tokens from AI models.
  3. Explore partnerships with Ferveret to test and implement their cooling system in data centers.

Want the full story? Read the original article.

Read on MIT News AI

More like this

Surprise upset: GPT-5.5 beats Claude Fable 5 on brutal new Agents’ Last Exam benchmark

VentureBeat AI#agents

Graviton5’s improved design increases speed and energy efficiency — beyond Moore’s law

Amazon Science#compute

The Practitioner’s Guide to AgentOps

Machine Learning Mastery#agents

Stop hand-tuning kernels: How Neuron Agentic Development accelerates AWS Trainium optimizations

AWS ML Blog#mcp

EXPLORE AI NEWS

Daily hand-picked stories on LLMs, RAG, agents and production AI — curated for engineers who ship.

BROWSE NEWS

GET THE WEEKLY DIGEST

Join engineers getting the Monday signal-over-noise AI breakdown. No spam, unsubscribe anytime.

LEARN AI ENGINEERING

Curated courses, research papers, repos and tutorials built for engineers leveling up in AI.

START LEARNING