Crypto news

Public mining companies are actively transforming their energy assets and data centers into infrastructure for artificial intelligence and high-performance computing. This trend is gaining momentum amid a sharp increase in capital expenditures in the AI sector and an acute shortage of sites with access to cheap electricity.

Nvidia's Record Bond Issuance as a Market Signal

In mid-June, Nvidia issued $25 billion in bonds, with demand exceeding $85 billion. Although the deal was not directly related to data center financing, it clearly demonstrates the enormous investor interest in AI-related infrastructure, where Nvidia's graphics processors remain key equipment.

This is Nvidia's first corporate debt issuance since 2021. Initially, the company planned to raise about $20 billion but increased the amount due to overwhelming demand. The issuance was split into seven tranches with maturities ranging from 2028 to 2056 and coupon rates from 4.25% to 5.625%. The proceeds will be used for general corporate purposes, including refinancing old debts. In my assessment, Nvidia needs this deal more to create a credit benchmark and increase liquidity rather than to directly finance data center construction — unlike Meta or Alphabet, Nvidia does not build them but supplies critical equipment.

Miners Are Selling Not Hashrate, But Energy and Infrastructure

The demand for AI infrastructure is radically changing the economics of mining companies. For computing clients, not only GPUs have become scarce, but also land plots, grid connections, cooling systems, and ready-made data centers. Major players already have all of this.

In May, Hut 8 signed a 15-year lease agreement for 352 MW of IT capacity in Texas. The base contract value is $9.8 billion, and with extension options, it could reach $25.1 billion. The campus is designed for 1 GW of capacity and will use the Nvidia DSX architecture.

In August, TeraWulf signed two 10-year agreements with the AI cloud platform Fluidstack for over 200 MW of load. The contracts imply approximately $3.7 billion in revenue over the base term and up to $8.7 billion including extensions. In May, the company also purchased a site in Eastern Kentucky for HPC infrastructure with a potential of over 1 GW.

In February, CleanSpark announced the development of a multi-gigawatt AI infrastructure platform and gained access to up to 890 MW of capacity in the Houston area. Company CEO Matt Schultz stated: "We are advancing negotiations with data center tenants in parallel with efforts to secure sites and electricity, which support sustained demand from AI and HPC."

Why Miners Are Moving from Bitcoin to AI

The shift to AI is driven not only by growing demand for computing. After the halving and increased mining difficulty, the profitability of Bitcoin mining has significantly decreased. Companies are seeking more stable and predictable sources of cash flow.

For AI clients, mining companies are attractive as owners of ready-made energy and data center infrastructure. However, transitioning to HPC requires significant additional investments: data centers for GPUs differ from mining sites in terms of reliability, cooling, networks, and customer service requirements. Not every site can be quickly retrofitted, but companies with large energy capacities and access to capital gain a unique opportunity to diversify their business beyond Bitcoin mining.

Let me remind you that by November 2025, seven out of the ten largest public miners by hashrate reported generating revenue from AI or HPC. And Nvidia's report in May 2026 pushed mining company stocks higher, confirming sustained demand for AI infrastructure.

My analysis: The current trend is not a temporary fad but a fundamental shift in the business model of the mining industry. Companies that manage to retrofit their capacities for HPC and secure long-term contracts with AI giants will gain a significant competitive advantage. Those who remain solely focused on Bitcoin mining risk facing further margin compression.