From mining to computing infrastructure: TeraWulf's financing expansion and HIVE's convertible bonds reveal the path of AI transformation for mining companies

In mid-April 2026, the global mining sector saw two landmark capital actions. The U.S.-listed miner TeraWulf issued 47.4 million shares of common stock at $19 per share, raising $900 million; Canadian miner HIVE Digital Technologies followed suit, announcing plans to privately issue $75 million of zero-coupon convertible senior notes due 2031.

The two financings point to the same goal: AI data center construction and GPU procurement, not the expansion of traditional Bitcoin mining rigs. This is not an isolated case. From Core Scientific selling off its Bitcoin holdings to secure a $10 billion-plus AI hosting contract, to Hut 8 and AI unicorn Anthropic signing a $7 billion infrastructure leasing agreement with Google-supported Fluidstack, miners’ “main-business turnaround” has evolved from individual companies’ strategic adjustments into a systemic, industry-wide migration.

The Capital Logic Behind the Two Financings

TeraWulf priced this share issuance on April 14. The issuance size was raised from the initially planned $800 million to $900 million. Underwriters also hold a 30-day over-allotment option, which allows for an additional subscription of up to 7.11 million shares. The funds will mainly be used for the construction of the AI data center park in Housville, Kentucky, and for fully repaying the outstanding amounts under the bridge credit facilities. Morgan Stanley served as joint bookrunner for this offering, and Cantor Fitzgerald served as an equity capital markets advisor.

Meanwhile, TeraWulf released preliminary first-quarter results for the period ended March 31, with revenue expected to be between $30 million and $35 million and adjusted EBITDA up to $3 million. As of the end of the first quarter, the company held $3.1 billion in cash and equivalents, and its total debt was $5.8 billion.

HIVE’s $75 million convertible note financing took a different route. The notes are structured as zero-coupon, with a maturity of 2031, issued by HIVE Bermuda 2026 Ltd. via a private placement to qualified institutional investors. The initial buyers were also given the option to additionally subscribe for $15 million in notes. The company also designed capped subscription transactions to reduce dilution for existing shareholders.

Regarding the use of funds, HIVE clearly stated that the proceeds will be used for GPU procurement, data center expansion, and general corporate purposes. Worth noting is that HIVE also disclosed that it is gradually cutting back its ASIC mining activities at the Boden facility in Sweden, converting the site into a Tier III high-performance computing data center, and deploying its first GPU cluster in Paraguay for early-training workload of large language models.

Why Are Miners’ AI Transformations Accelerating in 2026?

Miners’ AI transformation is not a phenomenon that started today, but data from the first quarter of 2026 highlights the urgency of this trend.

After the Bitcoin halving in 2024, miner profit margins fell by about 50%, while costs did not decline in parallel. Entering the second half of 2025, the continued weakness in Bitcoin prices further squeezed miners’ survival space. Industry estimates show that in Q4 2025, the weighted average cash cost for listed miners to produce one Bitcoin had risen to about $79,995; after factoring in depreciation, operations, and maintenance costs, total costs were even higher. As of April 17, 2026, Bitcoin’s price was about $74,729.9, creating a severe cost-price inversion, with some miners experiencing net losses exceeding $19,000 per coin mined.

In Q1 2026, the price of computing power fell to about $29 per PH per day, the five-year low. Meanwhile, Bitcoin’s total network hash rate fell from a peak of about 1.16 ZH/s to about 853 EH/s, a decline of roughly 22%, directly reflecting some miners shutting down and computing power flowing out.

In March 2026, Core Scientific confirmed it had sold about 1,900 Bitcoins, raising $175 million in cash, and planned to liquidate nearly all of its Bitcoin holdings to accelerate AI infrastructure expansion. In the same period, MARA Holdings sold more than 15,000 Bitcoins to reduce leverage, and Bitdeer’s Bitcoin holdings fell to zero.

From April 14 to April 16, 2026, TeraWulf and HIVE completed financing announcements one after another, pushing miners’ AI transformation to a new scale.

The drivers of miners’ AI shift come from two directions. First is the structural deterioration of Bitcoin mining economics—difficulty continues to rise, coin prices remain weak, and cash costs stay high; this triple squeeze makes traditional mining models difficult to sustain. Second is the gravitational pull from explosive growth in the AI computing power market—retrofitting and rebuilding mining sites typically takes about 18 to 24 months, while constructing new data centers from grid-connection application to commissioning often takes more than five years. Miners’ advantage in power infrastructure thus forms a scarce barrier.

The Scale and Pattern of Miners’ AI Transformation

From the perspective of industry data, miners’ AI transformation has moved from scattered trials to a systemic trend. The following is a quantified analysis from three dimensions: contract scale, financing structure, and revenue changes.

AI order scale: As of early 2026, multiple miners have accumulated total AI and HPC orders of about $38.5 billion. Key projects include the $12.8 billion AI data center order signed by TeraWulf and Fluidstack, the five-year $9.7 billion agreement signed by IREN and Microsoft, the $7 billion infrastructure leasing contract signed by Hut 8 and Google/Anthropic, and the more than $10 billion 12-year managed hosting agreement signed by Core Scientific and CoreWeave.

Changes in revenue structure: Core Scientific’s AI hosting revenue surged 268% year over year in Q4 2025, far outpacing the revenue performance of its mining business. HIVE reported revenue of $93.1 million in its latest quarter, up 219% year over year. Although it recorded a net loss of $91.3 million due to depreciation related to expansion projects, management explicitly prioritized long-term infrastructure growth.

Diversification of financing methods: Miners’ AI transformation financing sources show a diversified character. TeraWulf chose a public equity issuance, obtaining capital at the cost of diluting equity; HIVE chose zero-coupon convertible notes, exchanging future conversion rights for current liquidity; Core Scientific also expanded its credit line to $1 billion and systematically liquidated its Bitcoin holdings. Different financing paths reflect each company’s differentiated choices regarding capital structure and risk appetite.

Divergent signals from market feedback: Compared with the short-term decline in stock prices after financing announcements, there is a contrast with the long-term reassessment of valuations. TeraWulf’s stock is up 68.06% year-to-date; HIVE is up about 37% year-to-date; and Core Scientific’s return over the past year reached 173%. Some institutional analysts believe the long-term contract value of the AI hosting business has not yet been fully reflected in current valuations.

The pricing logic in the capital market is shifting. In the past, miners’ valuations were highly correlated with Bitcoin prices. After the AI transformation, the market has begun to re-evaluate these companies using the “data center REITs framework with power-access capability.” This switch in valuation logic is precisely the fundamental reason miners’ AI transformation is receiving positive feedback in capital markets.

Dissecting Public Sentiment: Support, Doubts, and a Calm Review

Miners’ shift to AI is an inevitable choice in the “computing power-centric” era. A scholar from the Digital Economy Institute at the Shanghai Academy of Social Sciences noted that Bitcoin is merely an early Token generated by a specific historical stage of computing power. When miners clear out Bitcoin and pivot to AI infrastructure, it represents a “decoupling” from the old Token system and an investment in new infrastructure aligned with computing power. At its core, this is a historic shift of tracks. Yu Jianing, Chairperson of the Academic Committee of the Hong Kong Registered Digital Asset Analysts Association, believes that power entry is becoming a more scarce strategic resource than the chips themselves; miners have accumulated a decade of first-mover advantages in power access, which is precisely the capability that AI giants need most.

The practical difficulty of the transition has been underestimated. On the one hand, Bitcoin mining hardware (ASICs) and AI servers (GPUs) cannot be directly repurposed, and mining farms’ power systems need to be upgraded to support higher-density computing loads. The total cost of a modern 100 MW AI data center exceeds $4 billion, with about 70% of spending going to servers and GPUs, far higher than the capital intensity of traditional mining sites. On the other hand, miners generally carry high levels of debt. With TeraWulf’s total debt at $5.8 billion, debt pressure may become a limiting factor before AI projects generate positive cash flow.

Some industry observers believe the scale and speed of miners’ AI transformation should be treated separately. Not all miners have the capability to build high-density data centers needed for AI hosting, nor do they have sufficient power redundancy space. The divergence between leading miners and small-to-medium miners will further intensify; ultimately, the companies that can complete the transition may be limited to the leading tier with high-quality power resources, strong capital-operations capabilities, and technological reserves.

Industry Impact Analysis: Mining Ecosystem, Network Security, and Capital Flows

Miners’ large-scale shift to AI data centers will bring multiple impacts to the Bitcoin mining industry and even the broader crypto ecosystem.

Changes in the Bitcoin network security landscape: The total network hash rate fell from the peak of about 1.16 ZH/s to about 853 EH/s, a drop of about 22%. A decline in hash rate means the cost of network attacks decreases. Although the economic cost of 51% attacks remains extremely high, continued hash-rate outflows may pose a structural challenge to long-term network security. At the same time, hash rate concentration may further increase—companies able to maintain mining operations will be more concentrated among leading miners with low-cost power.

Changes in Bitcoin market supply pressure: Miners’ large-scale liquidation of Bitcoin holdings created notable sell-side pressure in Q1 2026. MARA sold more than 15,000 BTC, and Riot Platforms cleared more than 3,700 BTC. Combined with sell-downs by companies such as Core Scientific and Cango, the collective selling behavior of listed miners exerted an undeniable downward impact on Bitcoin market prices.

Reshaping capital market fund flows: The investment logic for the miners sector is being rebuilt. Traditionally, investors viewed miners’ stocks as leveraged substitutes for Bitcoin. After the AI transformation, some institutions have begun to price these companies using valuation frameworks based on data center REITs. This change in valuation paradigm means the correlation between miners’ stock prices and Bitcoin prices may gradually decrease, while the linkage to the health of the AI infrastructure market may increase.

Upstream and downstream transmission effects: Reduced demand for mining rigs from miners will put pressure on ASIC manufacturers; the reallocation of power resources will affect regional energy market patterns; mining operations and maintenance service providers face a business model transition. The ripple effects of miners’ AI transformation are spreading across the upstream and downstream of the crypto industry chain.

Multi-Scenario Evolution Projections

Based on current information and industry trends, the following projections lay out three different evolution paths.

Scenario 1: Accelerated transition, leading miners establish a dual-core business structure

Projection logic: AI computing power demand continues to grow at a high rate. According to multiple research institutions, the global AI data center market will have a CAGR between 25% and 36% starting in 2026, and by 2034 the market size could reach the trillion-dollar (hundreds of billions to one trillion) level. Against the backdrop of power resources becoming a bottleneck for AI expansion, the value of miners’ existing power infrastructure will continue to rise. After leading miners complete the first round of AI infrastructure deployment, they can use stable cash flows generated by long-term contracts to support their Bitcoin mining business, forming a dual-core structure of “AI hosting plus Bitcoin mining.” In this scenario, the sector’s overall valuation midpoint moves upward, and the correlation with Bitcoin prices further weakens.

Scenario 2: Differentiation intensifies; only a few companies complete the transition

Projection logic: The capital intensity of AI data center construction far exceeds that of the mining business. The total cost of owning a 1 GW data center is about $55 billion, with the GPU chip share being the highest; power and cooling-related infrastructure follows. Many small and medium-sized miners cannot withstand the enormous upfront capital expenditures, or may struggle to obtain AI hosting orders due to non-compliant power infrastructure. After the transformation window closes, miners that successfully embed themselves in the AI supply chain gain a capital premium, while companies that fail the transformation face the fate of being acquired or exiting the market. Industry concentration rises significantly, and Bitcoin hash rate further concentrates among the small number of miners that either have not transformed or failed to transform; however, these companies will continue to bear the operating pressure of low profit margins.

Scenario 3: Bitcoin mining economics recover, and some miners slow their transition

Projection logic: If the Bitcoin price rebounds significantly to the $90,000 to $100,000 range within the next one to two years, mining economics will improve substantially. Some miners that have already completed part of their AI deployment may slow down further transformation, reallocating more resources back to the mining business. However, this scenario does not change the long-term structural trend. After the 2028 fifth halving, the block reward will be further reduced to 1.5625 BTC; even if the coin price is high at that time, unit output from mining will still be cut by half. Therefore, a rebound in the Bitcoin price may delay the pace of transition, but cannot reverse the industry-wide direction of structural migration.

These three scenarios are not mutually exclusive and are more likely to overlap over time. In the short term, Bitcoin price volatility will affect the speed of miners’ transition. In the medium term, whether AI data center contracts can be delivered smoothly and generate positive cash flow is the key to validating the feasibility of the transition. In the long term, the 2028 halving will further reinforce the inevitability of miners shifting to AI infrastructure.

Conclusion

TeraWulf’s $900 million equity issuance and HIVE’s $75 million convertible bond are landmark milestones in the 2026 miners’ AI transformation wave—not the endpoint. From Core Scientific to Hut 8, from MARA to IREN, the industry-wide trend of miners turning toward AI data centers is fundamentally a re-pricing and reallocation of the core asset value of “computing power.”

For the Bitcoin mining industry, this transition means a reshaping of network security, changes in market supply structure, and a recalibration of capital flows. For participants in the crypto market, miners’ AI transformation is rewriting the traditional investment logic of “mining stock = Bitcoin leverage.” And for the broader technology industry, miners—leveraging a decade of power infrastructure accumulation—are becoming key providers of AI computing capacity.

No matter how the evolution unfolds, one trend is already clear: miners’ identity is changing from “workers” of the Bitcoin network into “power landlords” of AI infrastructure. This industrial relocation spanning both the crypto and AI sectors has only just entered the fast lane.