In 2026, the definition of Bitcoin mining companies is being fundamentally reimagined.
Traditionally, the narrative has centered on the competition for computing power: the higher the hash rate, the greater the mining output, and the higher the company’s valuation. But over the past two years, this logic has loosened significantly. As North American data center construction lags far behind the explosive demand for AI computing power, Bitcoin mining farms—once viewed as "commodity-style capacity"—have suddenly become the ideal landing sites for large-scale AI data centers. They possess power access that others can’t obtain, substations that would take four years to build are already connected, and land approvals that typically require five or six years have long been completed.
The market is now repricing a group of assets based on this shift. Leading Bitcoin miners like Iris Energy, Core Scientific, and Hut 8 are redirecting their power and data center resources from the Bitcoin network to AI computing clients. This transition is no longer just a conceptual play—it’s supported by billion-dollar contracts and is reshaping financial structures, capital expenditures, and valuation models.
Transformation Context: Power Becomes the Most Scarce Resource for AI Computing, Miners Hold Ready-to-Use "Powered Facilities"
To understand Iris Energy’s current narrative, you need to grasp the real supply situation in the North American data center market. Research from Bernstein shows that the median time to secure 1 GW of grid power access in the US is about 50 months. Even in Texas, where data projects are relatively welcome, utilities process applications in batches, resulting in actual wait times far exceeding user expectations. Meanwhile, there is a significant gap between the contracted capacity for hyperscale data center hosting and the volume actually delivered—much of the demand is locked on paper and cannot be converted into usable IT capacity in the short term.
In contrast, publicly listed Bitcoin mining companies collectively control over 27 GW of planned power capacity, with only about 3.7 GW currently tied to disclosed AI contracts, representing contract value exceeding $90 billion. In other words, most miners’ power capacity remains untapped by the AI market, precisely as AI data centers urgently need scalable expansion.
From a revenue perspective, the economic incentive for transformation is clear: 1 MW of power capacity generates annual revenue of about $57–$129 in Bitcoin mining, while the same 1 MW used for AI GPU hosting or cloud services can yield $200–$500. In this transition, power is no longer just fuel for computation—it becomes a core infrastructure asset that can be directly monetized.
Bernstein analysts recently dubbed these miners "AI’s power landlords" as they began covering multiple mining companies. The industry consensus behind this label is that the supply chain bottleneck in the AI computing race has shifted from chip manufacturing to power access. Miners, with their brownfield sites, grid connections, and substations, are positioned at this structural inflection point. Over the past two years, the industry has outsourced 6 GW of power capacity to hyperscale cloud providers and new cloud operators, involving 17 transactions totaling more than $110 billion—about 10% of all data centers currently under construction in the US. Bernstein projects that the Bitcoin miners it covers will see combined AI revenue grow from $1.2 billion in 2026 to $10.7 billion by 2030.
Iris Energy: From Bitcoin Hashrate Leader to AI Cloud Service Platform—A Practical Case Study
Among the most aggressive Bitcoin miners undergoing transformation, Iris Energy stands out for its fully locked contracts and transparent capital structure.
On the contract front, IREN signed a five-year AI cloud service agreement with Microsoft at the end of 2025, valued at about $9.7 billion. Microsoft will pay IREN 20% of the client prepayment, and the contract is expected to contribute roughly $1.9 billion in annual recurring revenue, with an estimated EBITDA margin of 85%. To support this contract, IREN entered a $5.8 billion procurement agreement with Dell Technologies for GPUs and related equipment, funding the purchase of NVIDIA GB300 GPUs. These will be deployed in four liquid-cooled data centers at IREN’s 750 MW campus in Childress, Texas, with major deployment expected to be completed in 2026.
Capital structure is the key metric validating IREN’s transformation feasibility. IREN secured about $3.65 billion in investment-grade debt financing for the above procurement, compressing borrowing costs to 6.00%. Combined with Microsoft’s $1.9 billion prepayment, these cover about 96% of GPU capital expenditures related to the contract. Additionally, IREN signed a $3.4 billion GPU cloud spend agreement with NVIDIA, which includes an arrangement granting NVIDIA the right to purchase up to 30 million shares of IREN stock at $70 per share over five years. By the end of Q1 FY2026, IREN had secured over $9.2 billion in financing sources, spanning client prepayments, convertible bonds, GPU leasing, and GPU financing, with quarter-end cash and equivalents totaling about $2.6 billion.
On the capacity side, IREN has locked in more than 4.5 GW of grid-connected power. Its newly acquired Oklahoma campus covers about 2,000 acres with 1.6 GW capacity; grid studies are complete, and phased power access is expected to begin in 2028. The company plans to deploy 140,000 GPUs by the end of 2026, which is expected to generate about $3.4 billion in annual recurring revenue.
Performance-wise, data for Q2 FY2026 (ending December 31, 2025) has been publicly disclosed. Total revenue for the quarter was $184.7 million, down 23.1% from the previous quarter’s $240.3 million and below analyst expectations. Bitcoin mining revenue was $167.4 million, while AI cloud service revenue reached $17.3 million—up 137% from the previous quarter’s $7.3 million. Net loss for the quarter was $155.4 million, mainly driven by non-cash and non-recurring items, including about $219.2 million in unrealized losses and debt conversion costs from convertible bonds, $31.8 million in ASIC miner impairment charges, and $58.2 million in equity incentive expenses, partially offset by $182.5 million in tax credits. Adjusted EBITDA was $75.3 million, with the adjusted EBITDA margin expanding from 38% in the previous quarter to 41%. At quarter-end, the company held about $3.26 billion in cash and equivalents.
It’s worth noting that the expansion of net loss this quarter was primarily due to accounting treatment, not operational deterioration. However, in terms of free cash flow, expansion-phase funding pressures are significant. In Q1 2026 alone, free cash flow was negative $1.38 billion, meaning IREN’s further expansion relies heavily on ongoing external financing and client prepayment support.
On analyst pricing, as of early June 2026, IREN’s stock traded at about $66.60. The 12-month average target price from 14 analysts was $80.49, with a high of $126 and a low of $41; the consensus rating was "Buy," with 10 recommending buy, 3 hold, and 2 sell. Bernstein set IREN’s target price at $96.00 on June 4, 2026, previously having set it at $100.
It should be noted that IREN’s transformation progress is highly dependent on the timely delivery of its 140,000 GPU deployment plan and sustained absorption of computing capacity under the Microsoft contract. Management must maintain prudent control over debt and cash flow during rapid expansion; any delays in chip delivery, power access, or weakening AI compute demand could impact market pricing of this narrative.
Comparative Analysis: Core Scientific and Hut 8’s Differentiated Paths
Iris Energy isn’t the only company following this path. Core Scientific and Hut 8 are also pivoting to the AI data center market, each with different asset structures and business models. All three share core resources—power, land, and existing data center facilities.
Core Scientific has chosen a relatively "asset-light" approach: it positions itself as an AI data center host, leasing computing capacity to hyperscale clients rather than operating its own GPU cloud services. In Q1 2026, its hosted AI business revenue reached about $77.5 million, up nearly 9x year-over-year and, for the first time, surpassed Bitcoin mining revenue. Bernstein raised Core Scientific’s target price from $24 to $32, maintained an "outperform" rating, and estimated that about 86% of the company’s target enterprise value comes from AI business, with only about 14% from crypto mining.
Hut 8 has gone even further. In May 2026, Hut 8 announced a 15-year lease agreement with a high investment-grade tenant for an AI factory with 352 MW IT capacity, valued at about $9.8 billion, with a 3.0% annual base rent escalation clause. The contract is structured as triple net lease and "pay-as-agreed," with the tenant being a confidential high investment-grade company. Including the prior River Bend campus’s 245 MW contract, Hut 8’s total contracted AI data center capacity reached 597 MW, with cumulative base contract value of about $16.8 billion. On the capital side, Hut 8 successfully issued $3.25 billion in 16.5-year senior secured notes—the first single-sponsor project financing completed in the data center sector via the investment-grade bond market. The company also disclosed a development pipeline exceeding 8,375 MW. Hut 8’s stock saw significant gains in 2026, with analysts’ latest target price around $124.
From a valuation perspective, Bernstein provides a quantifiable benchmark: Bitcoin miners with signed AI contracts are valued at about $6 million per planned MW, while pure mining companies without AI contracts are valued at about $3 million per MW. The roughly twofold valuation premium reflects the market’s response to contract certainty—power capacity locked in long-term contracts is viewed as a scarce infrastructure asset generating stable cash flows, whereas uncontracted resources remain exposed to Bitcoin price volatility and global hash rate pressure.
How to Gain Direct Exposure to AI Computing Assets via Gate Stock Services
For investors focused on the AI infrastructure theme, the aforementioned transformation targets—such as Iris Energy, Core Scientific, and Hut 8—are all tradable directly on the Gate platform, eliminating the need to open separate accounts with traditional brokers.
Gate recently announced a strategic partnership with Alpaca, a global leader in brokerage infrastructure, to provide eligible users with real stock trading services. This service covers over 10,000 US stocks and ETFs listed on the NYSE and Nasdaq. Users can trade stocks and ETFs directly using USDT, with fractional share trading starting at just $1. Alpaca handles core functions including order execution, clearing, settlement, and asset custody, as well as dividend distribution and corporate action processing.
The key advantage of this model is that users can use their USDT holdings directly for US stock trading—no need for repeated fiat conversions, new KYC, or separate brokerage accounts. The trading interface mirrors crypto asset trading, supports both market and limit orders, and offers a seamless experience with virtually no learning curve.
Additionally, Gate’s TradFi module offers tokenized stock products, also using USDT to trade assets linked to US stock prices. Regardless of the method, users can manage both crypto assets and traditional securities in a unified account.
For users holding stocks like IREN, CORZ, or HUT, Gate’s minimum investment threshold of 0.01 shares and $1 greatly lowers the cost of allocating to individual assets, enabling small-scale test positions without the upfront capital required for whole-share purchases.
Conclusion
By 2026, the transformation of Bitcoin mining companies into AI computing infrastructure providers is no longer industry speculation—it’s an empirical reality backed by billion-dollar contracts and quarterly financial data from public companies. In this structural shift, Iris Energy, with its $9.7 billion Microsoft AI cloud contract and 140,000 GPU deployment plan, offers one of the clearest and most robust frameworks for AI computing expansion. Core Scientific rapidly shifted its revenue structure through a hosting model, while Hut 8 established a foundation for long-term capacity expansion with a massive development pipeline and investment-grade project financing. The shared value among these companies lies in power and land—assets long considered heavy-resource investments—which are now becoming measurable sources of scarce computing supply in the AI era.
From a valuation standpoint, analysts’ pricing models are undergoing fundamental changes—valuations are no longer anchored to Bitcoin price or global hash rate difficulty, but focus on MW capacity, contract duration, and the ability to lock in investment-grade clients. For investors tracking this sector, quarterly GPU deployment progress, the pace of major AI lease agreements, and partnerships with hyperscale clients offer a more accurate reflection of intrinsic value than simply monitoring Bitcoin price trends. On the practical side, Gate’s stock services and tokenized asset products have greatly simplified the asset allocation process for participating in this wave of AI computing infrastructure growth.
