Bernstein: TeraWulf, Cipher to see 9x AI revenue by 2030

Bernstein began coverage of bitcoin miners TeraWulf and Cipher Digital in a new analyst note, projecting AI-related revenue for the two firms could grow ninefold by 2030.

The firm cited rising demand for AI training and inference and the miners' access to low-cost electricity and existing data-center infrastructure as drivers of the opportunity.

Bernstein's projection assumes miners can market spare capacity-space, cooling and power-to support GPU-based AI workloads. The note lists potential revenue sources as colocation for AI hardware, managed training clusters and selling rack and power capacity to cloud and software companies.

TeraWulf and Cipher Digital are publicly traded operators that build and run large-scale bitcoin mining facilities. Both companies have sought preferred power contracts and expanded data-center operations that can be repurposed or augmented to host AI servers.

The note identified risks that could affect the size and timing of AI revenue: converting or equipping facilities for GPUs requires capital investment, a steady supply of specialized hardware and different cooling and networking arrangements than ASIC-based bitcoin rigs. The note also lists competition from established data-center operators and cloud providers, regulatory approvals and permitting, and fluctuations in cryptocurrency markets as further uncertainties.

The note says miners would need to install GPU clusters or partner with firms that supply and manage those systems. It names availability and cost of GPUs, the pace of enterprise demand for colocated AI compute, and each miner's ability to fund and execute the transition while maintaining mining operations as key determinants.

Bernstein estimates AI services could form a substantially larger share of total revenue for both companies over the coming decade.