Publicly traded mining firms have reported asset write-offs totaling approximately $1.1 billion in the first half of 2026, as they pivot their operations towards artificial intelligence. This figure was disclosed by TheEnergyMag.
Experts estimate that the equipment corresponding to the mining capacity that has been phased out is valued at around $1.5 billion.
Source: TheEnergyMag.Mining Gives Way to AI
According to TheEnergyMag, 12 monitored public companies recorded write-offs and asset impairments of about $1.1 billion for the period of January to June 2026. Notably, IREN and Core Scientific accounted for nearly 89% of this total.
At the same time, public miners have lost around 75 EH/s in actual hash rate. Valuing this equipment at $20 per TH/s, its worth is estimated to be approximately $1.5 billion, excluding buildings, power supply systems, cooling setups, and installation work.
Average market price of ASIC miners by energy efficiency categories. Source: Hashrate Index.Previously, it was reported that revenue from high-performance computing (HPC) and AI for public miners surged by 52% in the last quarter. Some companies are reallocating their energy and infrastructure from Bitcoin mining to support AI workloads.
The two figures of $1.1 billion and $1.5 billion reflect different processes. The first relates to accounting write-offs and asset impairments, while the second represents an estimated value of the equipment linked to the phased-out capacities. Experts noted that it is incorrect to combine these amounts as a single loss.
Equipment Depreciates Faster Than It Pays Off
A striking example is Cipher Mining, whose Black Pearl site began Bitcoin mining in mid-2025, but by the end of the same year, the company was already preparing to repurpose the site for HPC. Consequently, Cipher wrote off $96.1 million in mining equipment value, despite generating $57.9 million in revenue from these machines in 2025. This comparison does not reflect the overall return over the equipment's lifespan but illustrates how quickly companies must reassess the purpose of newly acquired assets.
The situation is even more pronounced for IREN. In June 2025, the company reached a hash rate of 50 EH/s, but in the first half of 2026, it reported approximately $695 million in write-offs and asset impairments, largely related to mining equipment being supplanted by AI projects.
Transition to AI Requires New Investments
These write-offs do not imply an immediate cash outflow of the same amount; they are an accounting reflection of asset value decline. However, the construction of new data centers and servicing debt still demand actual cash, as highlighted by TheEnergyMag.
For instance, TeraWulf earned about $53 million in revenue from renting out capacities for HPC workloads in the first half of the year, while its cash interest expenses amounted to $131 million. Although these figures do not definitively indicate financial troubles, as the company has other income sources and cash reserves, they highlight the gap between revenue growth from new projects and financing costs.
Access to credit may also become a constraint. Reports suggest that some major banks have become more selective in financing data centers. For miners who have already reduced Bitcoin production, delays in constructing or connecting new facilities could lead to prolonged periods of diminished revenue alongside ongoing obligations.
Changing Business Models Does Not Guarantee Success
The shift to AI allows miners to leverage existing sites and access to electricity for different business ventures. However, this often necessitates retrofitting and additional financing. The ultimate economics depend not only on the revenues from future data centers but also on how much can be recouped from old machines and the costs associated with infrastructure upgrades.
Moreover, not all write-offs are related to the transition to AI. For instance, Core Scientific attributed a significant portion of its impairment of mining assets to the deteriorating economics of Bitcoin mining. Thus, the figures presented reflect a broader reevaluation of investments in mining, rather than solely the costs associated with the transition to AI, as noted by specialists at TheEnergyMag.
It is worth mentioning that, according to the publication, in the first half of the year, 14 public mining and AI infrastructure companies secured $35.09 billion in debt financing, which accounted for nearly 73% of net capital inflow.
Follow ForkLog on social media
Telegram (main channel) Facebook X