The increase in demand for GPUs is driving cryptocurrency miners to transform data centers. In this way, they can meet the needs of the artificial intelligence (AI) industry.
According to a recent Bloomberg report, with AI’s unprecedented demand for data center and GPU resources, cryptocurrency mining companies are finding new opportunities in the AI sector by converting their existing infrastructures.
Computational power demand driven by AI industry
The AI boom has led to an unprecedented shortage of equipment and energy supplies for industry, along with the electricity needed to power these resources.
The report indicated that new facilities under construction are quickly being leased. In fact, around 83% of the capacity is already claimed by AI companies and cloud service providers. That’s according to a March report from commercial real estate firm CBRE Group.
To satisfy the growing demand for data center space, many companies are turning to cryptocurrency mining companies. Since many of them already have the necessary infrastructure and are looking for more profitable uses for their properties.
“Just getting cold calls or texts from industry peers who are looking for capacity because they’re finding it just doesn’t exist,” Matt Brown, chief operating officer at Core Scientific, described his experience at a conference hosted by Nvidia.
“Everyone is scrambling to figure out over the next 12 months where they can find capacity, and that is getting exponentially more difficult,” Brown said.
Core Scientific Partnership with CoreWeave
This collaboration trend was highlighted when Core Scientific announced a partnership with AI startup CoreWeave, which is expected to generate approximately 3.5 billion USD in revenue over 12 years.
“We see the opportunity in AI today where we can convert the existing infrastructure we have to host customers who want to install very large sets of GPUs for their customers who are ultimately AI customers,” said Adam Sullivan, CEO of Core Scientific.
The company has identified about 500 megawatts of capacity available for conversion into artificial intelligence facilities , potentially the largest share of GPU dedicated to AI worldwide. This capacity is enough to supply hundreds of thousands of homes for a year.
