Miners welcome a new beginning

For more than a decade, Bitcoin mining farms have been heavily criticized in the energy and tech sectors. Their massive electricity consumption has triggered congressional hearings, ESG rating downgrades, and ongoing public criticism. Yet now, these farms have signed 15-year leasing agreements with companies like Microsoft, Google, and Anthropic. The farms themselves have hardly changed. In fact, if there’s a commonality among these farms over the past ten years, it’s the crisis itself. So, what exactly has happened?

Regarding crises, there’s a very interesting adage: “The best opportunities often come from the most severe crises.” The experience of Bitcoin miners is exactly that. From July 2016 to April 2024, they experienced three halving events. Each halving halves the block reward, forcing miners to seek cheaper electricity in increasingly remote corners of the U.S. grid, including West Texas, rural Georgia, and the plains of North Dakota.

The weak have been eliminated. Some companies have pivoted in time. Others learned their lessons later.

In today’s story, I will explain how the surge in artificial intelligence infrastructure investments aligns with miners’ growing computational power and processing capacity, helping them find new vitality.

Let’s continue.

Halving — The First Turning Point

The first survival test for Bitcoin miners occurred in April 2024, during the most recent Bitcoin halving event. Each halving is a stress test. But each time, the reward is cut in half, and the challenge doubles.

The April 2024 Bitcoin halving reduced the block reward from 6.25 BTC to 3.125 BTC. Within a week after the recent halving, the hashrate price dropped from $0.12 per terahash to $0.047. Hashrate refers to the expected earnings per unit of mining power. By the first quarter of 2026, the hashrate price had fallen to its lowest in five years, at $0.023 per terahash per day.

Currently, the average cost to produce one Bitcoin is about $81,000. Including other non-production costs necessary to keep miners operational, the total mining cost per Bitcoin would far exceed $115,000. Bitcoin’s current trading price is $70,760. Over the past three months, its price has never exceeded $80,000. Do the math yourself.

The Bitcoin mining industry can only keep pursuing lower mining costs, while it has no control over Bitcoin’s price.

Those miners whose main income comes from the spread between mined Bitcoin and its sale on the open market suddenly reported losses. So, they shifted to holding the Bitcoin they mined. Their idea was to wait until Bitcoin’s price rose enough to turn a profit.

This strategy worked until Bitcoin’s price surged. But market volatility is cyclical. Every bull market is followed by a bear market and retracements. The cryptocurrency market is no exception.

10/10 — The Second Turning Point

October 10, 2025: A day of terror for the crypto industry, witnessing the largest crypto liquidation in history. Since then, crypto prices have plummeted to record lows, marking the start of a bear cycle. This completely shattered miners’ “mine and hold” strategy.

Some companies hesitated whether to change their approach. But others announced strategic shifts within 24 hours of the liquidation event.

On October 11, Bernstein released a report redefining the role of Bitcoin miners, no longer as hashrate producers but as holders with access to gigawatt-scale secure power grids. Analysts called these miners “a key link in the AI value chain.” They unanimously regarded IREN (formerly Iris Energy) as the top choice for successfully transitioning from Bitcoin mining to focusing on AI cloud infrastructure.

Digital asset leader and AI infrastructure provider Galaxy Digital announced it had raised $460 million to convert its Helios mine in Texas into a high-performance computing (HPC) campus for CoreWeave, with a 15-year lease and expected annual revenue exceeding $1 billion.

Following the 10/10 event, a series of systemic balance sheet liquidations ensued, which had previously defined the industry’s identity through the “mine and hold” strategy. Miners spent at least 18 months accumulating Bitcoin as reserves, viewing unsold Bitcoin as a confidence signal.

Under the heavy weight of the bear market, Bitcoin’s price dropped about 40% from its peak of roughly $126k in just 45 days, shaking this stance. Some publicly listed miners who had never sold Bitcoin before also began to sell. US’s third-largest Bitcoin holder, Marathon Digital (MARA), broke its record of holding Bitcoin continuously, selling 15,133 BTC in three weeks.

The CEO of this company has long supported and drawn inspiration from Bitcoin, with strategic reserves being its largest Bitcoin holdings. Less than two years ago, MARA’s CEO and chairman Fred Thiel announced Bitcoin as its strategic reserve asset.

Just last month, Fred did a 180-degree turn, admitting that selling Bitcoin “enhanced financial flexibility and increased strategic options, as we expanded from pure Bitcoin mining into digital energy and AI/high-performance computing infrastructure.”

But I won’t blame him. Difficult times require tough decisions. And MARA is not the only one abandoning Bitcoin as a permanent strategic asset.

While some investors increased their Bitcoin reserves after the liquidation, others slowed their accumulation or publicly stated they no longer see Bitcoin as a strategic reserve.

Bitfarms’ CEO candidly admitted: “We are no longer a Bitcoin company.” Ben Gagnon added that Bitfarms will focus on “building the infrastructure for future computing.” CleanSpark took a different approach, viewing its over 13,000 Bitcoin holdings as productive capital, and hedging with multi-layer covered call options.

Even if Bitcoin hasn’t disappeared from their balance sheets, they see it as a resource to strategically drive their infrastructure transformation.

Sow the Seeds of Fortune in Adversity

Transforming Bitcoin mining farms into AI infrastructure is no easy feat. The cost per megawatt for conversion ranges from $126k to $8M, including new liquid cooling systems, triple power redundancy, high-bandwidth fiber optics, and network upgrades for GPU training clusters.

However, mining infrastructure—cooling, power, and computing capacity—is closer than any other industry to meeting the needs of AI and data centers. Bernstein analysts pointed out in their report that existing miner infrastructure could reduce deployment times by up to 75%.

It’s not just analysts who hold this view. The deals struck by these mining companies in recent months also confirm it.

IREN signed a $9.7 billion contract with Microsoft to provide GPU cloud hosting services at its Texas Chillicothe campus, marking the largest single deal between miners and large-scale data centers to date. Hut 8 reached a $7 billion deal with Google-backed Fluidstack and Anthropic. Cipher Mining signed an $8.5 billion contract with AWS and Fluidstack. By Q4 2025, revenue from Core Scientific’s AI hosting business (renting data center space for IT equipment) will rise from 9% to 39% over four quarters.

The Surprising Fortress

But why are hyperscale data center operators paying miners for space?

Time is the key. To survive each halving, miners have had to chase cheaper electricity. To survive, they’ve adopted various measures: negotiating long-term power supply agreements, acquiring industrial land in low-cost energy corridors, building dedicated substations, and ensuring direct interconnection with the grid. Modern farms are equipped with dedicated high-voltage transformers, redundant power supplies, and thermal management systems designed to operate at full capacity around the clock.

Perhaps it’s not all planned in advance—you might say miners are just lucky. But who can strike gold while struggling to survive?

Currently, public miners have about 6.3 gigawatts of operational capacity, with another 2.5 gigawatts under construction. In most U.S. markets, data center interconnection wait times are 5 to 7 years. Microsoft’s internal forecasts show that its data center resource constraints will persist into 2026 and beyond.

That’s why hyperscale data center operators overlook miners’ lack of expertise in AI infrastructure. Instead, they foot the bill for substation costs, land use permits, utility relationships, and grid connections—things that often take years elsewhere to implement.

Mining companies can gradually improve performance by applying existing equipment to AI applications. MARA recently announced a $1.5 billion acquisition of energy infrastructure, boosting its total power capacity to over 2.2 GW. This allows MARA, at a cost unmatched by other AI infrastructure builders, to convert an already depreciated facility into AI infrastructure.

CEO Fred Thiel describes these assets as ready-made infrastructure that would otherwise take up to 10 years and $2–3 billion to build independently. He states, “Power is a scarce input in AI. With the planned acquisition of Langridge Energy, we will control an efficient, contracted energy platform.”

The Closed Door

There’s a trap in this story. Every megawatt shifted from Bitcoin mining to AI infrastructure subsidizes the economic interests of those still engaged in mining. This reduces mining difficulty, making it cheaper for Bitcoin miners to mine a block.

Some may still choose to allocate part of their equipment to Bitcoin mining as a hedge against price drops. But only those capable of bearing the costs of equipment replacement or reservation can do so. Not everyone can. The reason is that those who repurpose mining equipment for AI infrastructure cannot switch back repeatedly. Mining is an interruptible process. When electricity prices are high, you can shut down your miners. But AI and high-performance computing cannot do that. Once you lease or commit your computing capacity, you can’t temporarily cancel the agreement to mine Bitcoin with those devices.

For most miners, this isn’t feasible. They have only a short window to switch tasks, and such good luck doesn’t come often.

Everything has progressed so smoothly that it’s almost unbelievable. The Bitcoin halving squeezed mining’s economic benefits to the limit. The subsequent 10/10 liquidation event forced miners to face reality: holding Bitcoin during a bear cycle is not a viable strategy. But the booming AI infrastructure industry arrived just in time, giving miners both the motivation and the assets to pivot.

This scenario is unlikely to repeat. The mining companies signing contracts today will enjoy the economic benefits of the next decade, while latecomers will miss out on these gains.