Bitfarms vs Hut 8: The energy and compute divide

Bitfarms faces a Canadian energy squeeze

The sudden 23% surge in Canadian hydroelectric hosting costs forces Bitfarms to move its operations toward Manitoba because the 340 MW Regie de l’energie cap in Quebec strictly limits their future growth. This regulatory cap in Quebec prevents the company from expanding its current footprint within the province. Bitfarms must now prioritize its migration to Manitoba to secure more stable power costs. You already know that energy costs dictate mining margins. While the company seeks relief in Manitoba, the loss of capacity in Quebec creates a significant gap in its Canadian strategy.

Hut 8 and the Quebec bottleneck

Hut 8 remains caught in the same regional energy constraints that affect Bitfarms. The 340 MW Regie de l’energie cap in Quebec limits the availability of power for all large-scale miners in the region. This capacity limit forces a competitive struggle for remaining hydroelectric resources. Because the 23% increase in hosting costs applies to the broader Canadian hydro market, Hut 8 faces similar margin pressure. The migration of players like Bitfarms to Manitoba changes the competitive landscape for all Canadian miners. Will Bitfarms find enough cheap power in Manitoba to offset these Canadian losses?

Why Core Scientific’s pivot matters

Core Scientific captures value through a Texas AI colocation revenue pivot that matches the movement of organizations from AI experimentation into operational use. This transition requires developers with skills in Python and Azure services to manage agentic workflows and multimodal scenarios. The industry needs production-ready agents, which requires expertise in generative AI, visual understanding, and LLM-first text analysis. Microsoft provides a way to validate these skills through the Microsoft Certified: Azure AI Apps and Agents Developer Associate Certification. To earn this, candidates pass Exam AI-103: Developing AI Apps and Agents on Azure. This beta exam focuses on agent development, multimodal generation including images, video, and speech, and information-extraction pipelines for retrieval-augmented generation (RAG). The exam tests a developer’s ability to plan and manage Azure AI solutions and implement computer vision solutions. Microsoft Foundry provides the unified platform for model management, agent configuration, and evaluation pipelines. Candidates must work with business stakeholders, solution architects, data scientists, DevOps engineers, and cloud security engineers. Microsoft offers a discount of 80% for the first 300 people who take the beta exam on or before May 7, 2026. Registration requires the code AI103Claxton. This discount is unavailable in Turkey, Pakistan, India, or China. General availability for the certification arrives in June 2026. After the exam goes live, beta rescoring begins, and final results arrive approximately 10 days later.

The winner in a compute-driven market

Core Scientific comes out ahead because its Texas AI colocation pivot targets a different demand profile than traditional mining. Bitfarms and Hut 8 struggle with the 23% hydroelectric cost increase and the 340 MW capacity cap in Quebec. These companies remain tethered to the physical limitations of regional power grids. Core Scientific instead targets the growth of AI applications that require high-performance compute for tasks like information extraction and multimodal workflows. As organizations move toward the operational use of AI agents, the demand for specialized colocation facilities increases. Core Scientific’s revenue relies on the compute needs of the AI industry rather than the volatile energy costs of Canadian hydro.