DMG Blockchain pivots Bitcoin miner to AI data centre amid revenue slide

DMG Blockchain Solutions, the Vancouver-listed digital-asset and data-centre operator, reported a sharp deterioration in its second-quarter 2026 results — revenue fell 42% year-on-year to CA$7.3 million — while simultaneously doubling down on a strategic pivot that encapsulates a trend reshaping the entire bitcoin-mining sector: the migration from proof-of-work compute to AI infrastructure.

The company mined 69 bitcoin during the quarter, flat against Q1 2026 but down 25% on the same period a year ago, as network-wide difficulty and lower average token prices compressed economics across the industry. Net loss widened marginally to CA$3.5 million, and total assets contracted to CA$109.9 million, partly driven by a CA$16.7 million markdown on the fair value of the company's digital-currency holdings — a reminder of how balance-sheet volatility remains structurally embedded in mining-company accounts.

From proof-of-work to AI compute

The strategic pivot is centred on DMG's Christina Lake facility in British Columbia, which management says it is converting into an AI data centre capable of providing at least 50 megawatts of critical IT load. CEO Sheldon Bennett said the company is "encouraged by what we see in the market," pointing to dual strategic pillars: core data-centre infrastructure and a second strand branded Core+, which houses Systemic Trust — a digital-asset financial services platform DMG describes as being built out to offer a "broad range of services."

The numbers behind the pivot are telling. Operating and maintenance expenses fell to CA$5.2 million from CA$7.6 million a year ago, in part because the company retired inefficient miners and benefited from favourable non-firm electricity rates — a cost structure that improves the economics of repurposing the same power envelope for AI workloads, which command materially higher rack-revenue per megawatt than bitcoin mining at current prices.

Convergence context: mining hardware meets AI infrastructure

DMG is not alone. Across North America, bitcoin miners that accumulated large power contracts and purpose-built facilities during the 2021–2022 bull cycle are now competing for AI hyperscaler tenancy and high-performance-computing (HPC) contracts. The structural logic is compelling: both workloads are power-hungry, both require dense cooling, and the land and grid-connection assets are already in place. What differs is the contractual profile — AI colocation deals typically carry multi-year, fixed-fee revenue that is structurally decoupled from token-price volatility.

For cross-sector investors, the DMG story sits at the intersection of two capital narratives that have been running in parallel but are now colliding. On one side, sovereign and institutional capital is flowing into AI infrastructure — data-centre REITs, hyperscaler capex, and edge-compute buildouts are all expanding. On the other, the bitcoin-mining sector is under structural margin pressure as the April 2024 halving continues to work through industry economics. The companies that survive the compression are those that can credibly reposition their power and facility assets toward the higher-margin AI compute market before their balance sheets erode.

DMG's CA$47.4 million in cash, short-term investments and digital assets gives it a reasonable runway to execute the Christina Lake conversion, though the CA$18.9 million current portion of loans payable — up sharply from CA$10.9 million at the prior year-end — means refinancing risk deserves attention. The company notes it expects its balance sheet to weather a crypto downturn, though that claim should be read in the context of a CA$24.6 million total current liabilities position versus CA$52.3 million in current assets.

The broader question for the sector is whether AI hyperscalers and enterprise compute buyers will treat converted mining facilities as credible Tier III-equivalent infrastructure, or whether the heritage of the assets creates a reputational discount when competing against purpose-built data-centre operators. That answer will define which miners successfully complete the transition — and which remain stranded in a proof-of-work economics trap.