AI Hardware-maker Sustainable Metal Cloud Claims 50% Reduction In Data Center Energy Use Amid AI Boom

The surge in artificial intelligence (AI) is driving up demand for powerful processors. From OpenAI to Meta and Google – the top AI businesses are struggling to reduce the energy needed to keep their AI data centers operational.

Amid this challenge, Sustainable Metal Cloud, a company specializing in AI hardware solutions, is making headlines with a groundbreaking claim.

Sustainable Metal Cloud (SMC) has announced a significant leap in data center efficiency with its HyperCubes technology. Their new data centers reportedly slash energy consumption by up to 50% compared to traditional air-cooling systems. Moreover, the immersion cooling technology boasts a 28% lower installation cost than other liquid-based cooling systems. Sounds too good to be true, right?

Yet, that’s not all as Tim Rosenfield, co-founder and co-CEO of SMC, highlighted other benefits. “It enables high-density hosting for GPUs and supports platforms like Nvidia’s Grace Blackwell,” Rosenfield said, referring to Nvidia’s latest AI graphics processors.

So, what makes the HyperCubes technology tick and how can it help AI businesses?
 

What Did Sustainable Metal Cloud Announce?

SMC, which already operates “sustainable AI factories” in Singapore and Australia using Nvidia’s H100s, announced its new immersion cooling technology. HyperCubes are designed to be versatile and can integrate into existing data centers.

However, the key focus of this announcement was for its partnerships with major AI players like Nvidia, Google and Deloitte. In fact, SMC’s HyperCubes contain servers fitted with Nvidia H100 processors, submerged in a synthetic oil called polyalphaolefin to draw heat away more efficiently than air. The company said this is what helps reduce energy consumption by up to 50%, as compared to traditional air-cooling technology.
 

Why Does This Matter For AI Businesses?

Sustainable Metal Cloud’s advancements come at a crucial time as AI energy demands are under scrutiny. Tech giants are exploring various solutions, including nuclear power, to meet these growing power demands. Plus, concerns about the environmental impact and stability of new solutions are growing, with some suggesting that AI investments will ultimately balance out the higher electricity consumption.

Yet, as AI energy demands scale, data centers are increasingly turning to liquid cooling to address overheating risks. Giordano Albertazzi, CEO of Vertiv, anticipates a rise in liquid cooling adoption by 2024. Even Nvidia’s CEO Jensen Huang and Supermicro’s CEO Charles Liang have lauded the benefits of liquid cooling, citing energy savings and cost reductions in AI data centers.

However, challenges remain as SMC’s CEO Rosenfield acknowledged that many data centers are not yet equipped for liquid cooling. “The market is figuring out the best way to employ this technology, and there will be multiple approaches,” he noted.


Currently, SMC is raising $400 million in equity and $550 million in debt to support its expansion beyond Singapore. The company is planning to expand its footprint into Thailand and India, aiming to meet the growing demand for energy-efficient AI solutions. The funding will help bolster its efforts to deploy energy-efficient AI data centers globally.

Despite these challenges, the trend toward liquid cooling continues to gain momentum within the data center and AI industry.

Do you think Sustainable Metal Cloud’s HyperCubes approach will pave the way for more efficient AI infrastructure?

Let us know your thoughts in the comments below!