Hut 8 Corp. is constructing dedicated AI data-center sites instead of converting its existing bitcoin-mining facilities, CEO Asher Genoot said.

This strategic pivot highlights the technical divide between cryptocurrency mining and artificial intelligence workloads. While both require massive power, AI training demands specific infrastructure and cooling that repurposed mining sites may not provide efficiently.

The company has focused its expansion on several high-capacity projects. In Nueces County, Texas, Hut 8 signed a 15-year lease [2] for the Beacon Point AI data-center campus valued at $9.8 billion [1]. This move contributed to a contracted capacity of 597 MW [3].

Hut 8 is further expanding its footprint with a second Texas campus. The company has planned a total capacity of one gigawatt [5] for this second site, utilizing an EPCM contract awarded to Jacobs to deliver the facility.

Infrastructure needs extend beyond power to water management. In West Feliciana Parish, Louisiana, the company has invested $16 million [4] into water infrastructure to support its AI campus operations.

Genoot said the decision to build new sites is driven by the need to meet the rising demand for AI model training. By avoiding reliance on legacy bitcoin-mining hardware, the company aims to optimize the performance, and reliability of its AI infrastructure.

“Hut 8 is building dedicated AI data-center sites rather than converting its existing bitcoin-mining facilities.”

Hut 8's approach signals a shift in how digital infrastructure companies are diversifying. Rather than treating AI as a software layer that can sit atop existing mining hardware, the company is treating it as a distinct real estate and engineering challenge. This suggests that the specialized cooling and power density requirements of modern AI models are too stringent to be met by the retrofitting of older bitcoin mines.