#Gate广场五月交易分享

‍# Cryptocurrency Mining Companies Accelerate Deployment of AIDC
Hut 8 finalizes a 15-year, $9.8 billion data center lease; IREN signs a $3.4 billion AI cloud contract with NVIDIA and a 5GW strategic partnership; multiple mining companies simultaneously shift toward AI computing services. Behind this is not just a simple case of “collective job hopping”:

1. When mining coins is less profitable than “selling electricity”: The underlying logic of mining companies shifting to AI

In the Bitcoin mining circle of 2026, a real-life “escape from mining” is unfolding. Mining bosses who once made a fortune mining coins are now packing their mining rigs into warehouses and turning into “electricity landlords” for AI giants. The trigger for all this is that Bitcoin mining has become a “loss-making business”—currently, the cost to mine one Bitcoin is about $87k, while the coin price hovers around $70k, losing $17k per coin mined, which is like “losing a car for each Bitcoin mined.”

What’s even more despairing for mining companies is the “hash power prisoner’s dilemma”: when the coin price drops, miners dare not shut down easily, or they risk losing network share; but continuing to mine means burning money continuously. This “dilemma” coincides with the “super opportunity” of AI computing demand. As large model training and inference needs explode, the global AI computing gap is expected to reach 44 gigawatts by 2028, with electrical infrastructure becoming the biggest bottleneck—building new data centers takes 5 years from grid connection application to commissioning, whereas upgrading mining farms only takes 18-24 months.

Mining companies suddenly realize that their years of stockpiling electricity, land, and cooling facilities have become “hot commodities” in the AI era. As one miner joked: “We used to be ‘coal bosses’ mining coins, now we’re ‘landlords’ in the AI world, earning just by collecting electricity fees.”

2. How crazy is the transformation? From “selling coins to survive” to “getting overwhelmed with AI orders”

In this wave of transformation, mining companies are acting more and more aggressively. Core Scientific sold 1,900 Bitcoins in one go, cashing out $175 million, then signed a 12-year, $10.2 billion AI hosting contract with CoreWeave; Hut 8 simply leased its entire Louisiana data center to Google, with a $7 billion contract pushing its mining business into a side gig; the most extreme is IREN, which paused mining expansion, spent money buying 4,200 NVIDIA chips, and its stock price soared 600% this year—an outright “AI concept stock.”

The confidence of these mining companies comes from the “super profits” of AI business: AI hosting revenue per megawatt is over three times that of mining, with operating profit margins reaching 80%-90%. For example, IREN’s AI cloud service has a gross margin of 86%, a stark contrast to the “loss-making” mining business—truly “from hell to heaven.” Even the traditionally conservative Bit Deer quietly upgraded some of its mines into AI data centers, claiming it’s “attack and defend at the same time,” but in reality, they don’t want to miss out on this computing power feast.

The capital market also votes with its feet: by 2025, the average stock price increase of mining companies is 300%-400%, far surpassing Bitcoin’s 10% rise. Investors now talk about mining companies in terms of “AI computing contracts” and “electricity reserves,” and anyone mentioning “mining output” feels embarrassed to claim expertise.

3. Pitfalls on the transformation road: Not all mining companies can become “AI landlords”

However, behind this frenzy of transformation, many hidden currents exist. First is the technical barrier: ASIC chips used in Bitcoin mining are incompatible with GPUs needed for AI; transforming a mining farm requires rewiring and cooling upgrades—equivalent to “turning a coal kiln into a high-tech factory,” which is costly. Small and medium miners can’t even afford GPUs, let alone the initial investment for upgrades, only watching as top-tier firms take the lead.

Second is customer resources: AI giants choose partners based not only on electricity but also on operational capabilities and compliance qualifications. CoreWeave chose Core Scientific because of its 1.3 GW power capacity and mature operations; small miners with no reputation, even if they have electricity, find it hard to secure long-term contracts with big firms and can only do some “piecemeal jobs” to earn a little.

More critically, the AI computing market is changing rapidly. Currently, “electricity is king,” but once grid construction catches up and GPU capacity increases, the advantage of mining companies may diminish. As an industry analyst said: “Mining companies are making money just by providing electricity now, but the AI industry is changing too fast—maybe one day they’ll be replaced.”

4. Impact on the crypto industry: Bitcoin network “bleeding,” ecosystem restructuring

The most direct impact of mining companies shifting to AI is the decline in Bitcoin hash rate. Early 2026, the total network hash rate peaked over 1000 EH/s, but it has started to fall back, and network difficulty has been reduced by 10.7%. While this hasn’t yet affected network security, if more miners exit, Bitcoin’s “hash rate moat” could be weakened.

A deeper impact is the restructuring of the mining ecosystem. Previously, miners were “faithful believers” in Bitcoin, holding large amounts of Bitcoin and acting as the “ballast” of the market; now, as they liquidate Bitcoin and turn to AI, it’s akin to “decoupling” from Bitcoin. This could intensify price volatility and gradually weaken Bitcoin’s “hash rate consensus”—after all, when mining is unprofitable, who will pay for network security?