The valuation framework for publicly listed Bitcoin mining companies is undergoing a fundamental reconstruction. The core metrics have shifted entirely from traditional hash rate scale and BTC holdings to energized capacity, contracted IT load, and project delivery capability. Land, substations, and grid connection rights once built specifically for mining operations are now being redefined as scarce infrastructure assets that AI data center developers are competing to acquire.

This paradigm shift is driven by a macro reality where electricity has become the scarce link connecting Bitcoin mining and artificial intelligence. The International Energy Agency's April 2026 forecast update projects global data center electricity consumption will nearly double from approximately 485TWh in 2025 to roughly 950TWh by 2030. More notably, the IEA previously estimated that U.S. data centers would account for nearly half of all new electricity demand added between 2024 and 2030. Meanwhile, revenue per unit of Bitcoin mining hash rate continues to face persistent pressure, forcing miners to reassess their asset portfolios and reconfigure power infrastructure originally built for ASIC miners into critical resources for AI high-performance computing.

The breakdown of the traditional mining formula stems from sharply compressed profit margins and the commoditization of the business. Following the 2024 Bitcoin halving, block subsidies dropped to 3.125 BTC, and with network hash rate and difficulty rising over the long term, the revenue share per mining machine has been continuously diluted. VanEck data shows that as of mid-July 2026, Bitcoin's hashprice had fallen to approximately $30.6 per PH/s per day, with daily average miner revenue across the network around $28.5 million over the past 30 days. Inefficient mining machines are now approaching or breaching their breakeven points, meaning even without a significant drop in coin prices, intensifying competition continues to erode profits.

By contrast, once an AI data center secures a long-term lease with a creditworthy counterparty, it converts grid access and construction and operation capabilities into longer-duration, more predictable cash flows, thereby escaping real-time dependence on coin price volatility and network difficulty. CoreScientific (CORZ.US) financial performance serves as the most direct sample of this transformation. In the second quarter of 2026, the company reported total revenue of $164.2 million, surging 109% year over year. High-density hosting revenue reached $136.7 million, accounting for approximately 83% of total revenue, compared to just $10.6 million in the same period last year. In stark contrast, self-mining revenue plummeted 66% from $62.42 million in the prior-year period to $21.54 million, recording a gross loss of approximately $12.17 million with a gross margin of negative 56%. The high-density hosting business, however, achieved roughly $79.98 million in gross profit with a gross margin of 59%. The same set of assets produces dramatically different financial outcomes under two business models, marking a complete shift in the company's revenue focus from mining to AI infrastructure.

Billion-dollar contracts among industry giants further validate the new pricing logic for power capacity. In July 2026, TeraWulf (WULF.US) signed a 20-year data center lease with Anthropic covering its Justified Data campus in Hawesville, Kentucky, planning to deliver approximately 401MW of critical IT capacity with delivery expected to begin in the second half of 2027 and full operations by early 2028. TeraWulf stated the contract is expected to generate approximately $19 billion in revenue over the initial term and has secured investment-grade credit support. In the same month, Hut8 (HUT.US) announced a second 15-year lease valued at $9.8 billion at its Beacon Point campus in Texas, adding 352MW of IT capacity and bringing the same client's contracted capacity at that campus to 704MW. The base-term contract value at Beacon Point now reaches $19.6 billion, with the first second-phase data hall expected to begin delivery in the second quarter of 2028.

The ability to command higher premiums for the same power capacity when serving AI stems from fundamental differences in risk structure and capability requirements. Bitcoin mining can tolerate higher interruption rates, with miners able to voluntarily shut down or relocate equipment during electricity price spikes. AI training and inference workloads, however, demand extremely high power supply stability, network bandwidth, cooling capacity, and system redundancy. Therefore, what actually commands a premium is not simply "having electricity" but rather a combination of four capabilities: energized power capacity or capacity with clear interconnection arrangements, engineering capability to complete construction on schedule, financing capacity to absorb massive upfront capital expenditures, and customer creditworthiness to secure long-term leases.

Long-term leases transform miners from commodity-like producers with volatile revenue into stable cash flow generators resembling data center developers or infrastructure asset operators. A lease backed by an investment-grade hyperscaler enables miners to secure financing at lower costs, and the same 100MW capacity will carry significantly different valuations if the tenant's credit quality is weaker. Wall Street has begun pricing miners by megawatts rather than BTC. In VanEck's June 2026 AI infrastructure valuation framework for miners, Gross Energized Power serves as the primary comparison metric. Based on data as of June 4, 2026, companies that have signed AI or HPC leases generally command valuation multiples above 10 times under this metric, while companies with little to no contracted capacity and only forward-looking power reserves trade at multiples of just 2 to 6 times.

These multiples are not traditional P/E ratios, EV/Revenue, or EV/EBITDA. The market is strictly distinguishing among four capacity states: planned capacity remains at the development concept stage, secured power capacity has agreements in place but is not yet energized, energized capacity has power supply conditions ready, and delivered and billing capacity is already generating revenue. As more projects come online, CoreScientific, TeraWulf, Hut8, and Cipher (CIFR.US) are increasingly resembling data center development and operation platforms, while MARA and CleanSpark (CLSK.US) maintain higher exposure to Bitcoin mining. The mining stock sector is experiencing a clear logic divergence.

However, the AI transition also brings execution barriers and funding gaps far larger than those of traditional mining. Based on VanEck estimates as of June 4, 2026, the relevant companies had delivered only about 25% of their leased capacity at that time, with a funding gap of approximately $50 billion between near-term capital expenditure requirements and existing cash. Furthermore, long-term capital expenditure needs for these companies approach $221 billion. The GW-scale figures and total contract values announced in press releases still correspond to construction plans scheduled for 2027 or even 2028 and beyond. Projects remain subject to grid upgrade delays, equipment delivery timelines, construction costs, financing conditions, regulatory approvals, and community resistance. For miners lacking experience in high-density data center construction, any delay or cost overrun could strain cash flows.

Customer concentration risk also exists, as a single tenant cutting capital expenditure or next-generation chips driving design changes could leave heavy assets stranded with limited alternative uses. Additionally, raising capital through equity issuance, convertible bonds, project loans, and customer prepayments may bring shareholder dilution, higher leverage, and complex financing constraints.

The industry's endgame points toward the evolution of miners into power infrastructure companies. Bitcoin mining has not lost its value, as it still enables rapid monetization of electricity and provides high elasticity returns during price rallies. Unlike traditional data centers, mining facilities can actively shed load and participate in grid demand response programs, generating transitional revenue for uncontracted capacity. But for a group of listed miners, Bitcoin is shifting from being the sole core business to just one monetization channel for power infrastructure. What is truly difficult to replicate is the acquired land, interconnection qualifications, transmission facilities, and large-scale power arrangements.

The statement that "the most valuable asset is electricity" requires a qualifier: it is not planned GW that matters, but electricity that can be interconnected on schedule, financed, built into high-density data centers, and leased long-term by reliable customers. Miners are transitioning from seeking cheap power to produce BTC toward monetizing the time value of scarce grid connection resources and infrastructure through leases. Bitcoin mining has thus arrived at a new fork in the road. Some companies will continue betting on coin price and hash rate cycles, maintaining the traditional mining model, while others may completely shed the "miner" label and become a new breed of AI-era power landlords.

This transformation is not merely an adjustment of business focus but a fundamental reshaping of asset attributes and valuation systems. As the 2027 to 2028 delivery window approaches, the market will test whether these companies can convert their power blueprints into actual revenue, determining whether they ultimately become overleveraged failures or establish themselves as a new generation of infrastructure giants. This represents another major paradigm shift in the convergence of energy and computing, following the rise of internet infrastructure.