Cryptocurrency miners are reportedly scrambling to boost their revenues by forging deals with AI developers.
As the Financial Times (FT) reported Wednesday (July 17), these miners operate vast, powerful computing sites and have struggled to turn a profit due to high energy costs and reduced rewards for mining.
Now, the report says, the miners seek to capitalize on rising demand for chips known as graphics processing units (GPUs), used in both their field and in artificial intelligence (AI) processing.
Core Scientific, one of the world’s largest bitcoin miners, is “aggressively pursuing” AI deals, CEO Adam Sullivan told the FT.
“It’s an incredibly important part of the business,” he added.
The company, which has data centers in Georgia, North Carolina and Texas, recently reached an agreement with AI cloud provider CoreWeave last month that the companies project will be worth $4.7 billion in revenue over 12 years.
CoreWeave itself is a former crypto miner that made the move to AI and saw its valuation climb to $19 billion in May, when it raised $7.5 billion in debt financing.
The FT report notes that AI firms need a lot of energy and computing infrastructure, both of which crypto miners can offer, a better proposition than AI firms building their own high-performance computing (HPC) data centers.
(Big tech companies such as Google and Microsoft are nonetheless spending billions on AI data centers. And as PYMNTS argued in April, they will likely also make their own AI chips.)
“It [normally] takes 3-5 years to build an HPC-grade data center from scratch,” J.P. Morgan analysts wrote in a recent note, per the FT. They added that this timeline has grown even longer due to rising AI demand.
An analysis earlier this year by the Energy Information Agency (EIA) estimated that large-scale cryptocurrency operations in the U.S. eat up more than 2% of America’s electricity, roughly the equivalent of adding another state to the nation’s power grid.
Meanwhile, the International Energy Agency (IEA) issued in January published projections for global energy use over the next two years, including estimates for electricity consumption from data centers, crypto and AI.
The IEA said that, put together, this usage was the equivalent of nearly 2% of the world’s energy demand in 2022, and that this demand could double by 2026. That would be almost the same amount of electricity used by Japan.