Bitcoin Miners Squeezed as Difficulty Rises and AI Pivots Accelerate

Bitcoin miners are getting squeezed by brutal difficulty, weak BTC pricing, and almost nonexistent transaction fee income, while a growing number are chasing AI and high-performance computing revenue just to stay alive.

• Mining difficulty is near 139 trillion hashes
• All-in BTC production cost is around $85,000
• BTC price sits near $62,000, leaving many miners underwater
• Transaction fees made up less than 1% of miner revenue this month
• AI/HPC pivots are becoming survival moves, not hype
• Cryptojacking and rogue AI agents are also stealing GPU power

Bitcoin mining economics are back in ugly territory. Difficulty has climbed to just under 139 trillion hashes, making it harder than ever to find a block and earn the reward. At the same time, the average all-in cost to produce a single BTC is estimated at about $85,000, while Bitcoin trades near $62,000. That is not a healthy spread. That is a profitability grinder with the blade set to “obliterate.”

For readers new to the mechanics: mining difficulty measures how hard it is for miners to solve the puzzle needed to add a new block to the Bitcoin blockchain. Block subsidy is the newly created BTC miners receive for doing that work. ASICs are specialized mining machines built for one thing only: Bitcoin mining. And hashrate is the total computing power pointed at the network. When difficulty rises and BTC price weakens, only the most efficient operators can keep the lights on without praying to the spreadsheet gods.

A difficulty adjustment is scheduled for June 14, with the network expected to drop to about 123.7 trillion hashes, roughly an 11% reduction from current levels. That should give miners a bit of breathing room. But breathing room is not the same as rescue. The deeper problem is that Bitcoin miners are still heavily dependent on the block subsidy, while transaction fees are contributing almost nothing. This month, fees accounted for less than 1% of total miner revenue.

That fee weakness matters. Bitcoin’s long-term security model was always meant to transition gradually from block subsidies to transaction fees as subsidies decline over time. Satoshi Nakamoto’s original design assumes fees eventually become meaningful enough to support the network. Instead, the fee market is still barely showing up to work. The result is a mining business that remains highly sensitive to BTC price, energy costs, and hardware efficiency, with little cushion when any one of those turns sour.

In plain English: if the price of BTC is below the cost to produce it for a large share of miners, something has to give. Usually that means higher-cost miners shut down, sell coins they would rather keep, or take on more debt to survive until conditions improve. And yes, some of them are effectively running a capital-intensive business on hope, leverage, and a lot of PowerPoint optimism.

That pressure is forcing a major shift toward AI and high-performance computing infrastructure. HPC stands for high-performance computing, which usually means renting out data center power, cooling, and GPU capacity for AI training, cloud workloads, scientific computation, and similar tasks. It’s not glamorous, but it is often more predictable than pure Bitcoin mining. Same buildings, same power hookups, different customer, better margins.

That’s why so many miners are suddenly talking like data center companies. The sector has become obsessed with the phrase “if you’re not pivoting to AI, you’re not paying attention,” and for once, the cliché isn’t entirely wrong. AI demand can be steadier than BTC mining revenue, and in a world where energy, land, and cooling are the real scarce resources, miners are trying to monetize those assets however they can.

The financing numbers make that trend pretty obvious. Hut 8 raised $4.35 billion in senior secured notes. IREN closed a $3.65 billion GPU financing facility. Cipher Digital priced an $810 million offering, while Keel Infrastructure and Bitfarms completed a $458 million offering. Those are not tiny survival rounds. They’re serious capital raises aimed at turning mining infrastructure into something closer to an AI and HPC platform.

There’s a practical logic to it. Serving AI and other high-performance computing operators offers a more predictable revenue per megawatt-hour than pure BTC mining. That matters because miners don’t just need cheap power; they need consistent cash flow. Bitcoin can be brutally cyclical. AI workloads are still competitive, but the demand profile is often easier to plan around. For miners, that is the difference between building a business and building a stress test.

Bitdeer shows both sides of that pivot. The company launched its SEALMINER DL1 Hydro and broke ground on a 100MW data center in Alberta near Fox Creek, with operations expected by Q2 2027. On paper, that’s a long-term infrastructure play with real scale. The company said it was “executing a disciplined strategy to strengthen our mining foundation while advancing into AI infrastructure.” Translation: we need more than one way to make money, and we need it before the next margin squeeze gets uglier.

But Bitdeer is also dealing with some messy internal churn. Within 11 days, three top executives departed: COO Chao Suo, CBO and former CEO Linghui Kong, and, earlier, CFO Jianchun Liu. Michael Potter has been named the new CFO. Leadership turnover at a capital-heavy mining and infrastructure company is never a great look. It doesn’t automatically mean disaster, but it does make investors wonder whether the “disciplined strategy” is solid or just corporate wallpaper over a rough patch.

Cango is another example of how ugly the math can get. The company reported May self-mining hashrate of 23.3 EH/s, down from nearly 28 EH/s in March, and mined 236.5 BTC in May. In Q1, it posted $102 million in revenue but also a $261 million net loss, with costs hitting $356.3 million. That’s what happens when top-line revenue exists, but the cost structure is chewing through it faster than the machines can mint fresh coins. A business can brag about revenue growth all day long, but if expenses are sprinting ahead with a flamethrower, the bragging rights vanish fast.

BitFuFu is also living in the gap between operational scale and profitability. It reported a $35 million net loss in Q1, with $72.7 million in mining revenue, $57.5 million in cloud mining revenue, and just $11.4 million in self-mining revenue. The company held 1,794 BTC in treasury at quarter-end and 1,855 BTC by the end of May. That’s a serious stack of coins, but treasury accumulation is not the same as a healthy mining margin. You can be sitting on BTC and still be getting kneecapped by operating costs.

Hive Digital is trying to straddle the same divide. In its final fiscal quarter, it posted $71.8 million in revenue, including $67.2 million from mining and $4.6 million from HPC. Full-year revenue came in at $297.8 million, but the company still logged a $148.4 million loss. Hive ended the period with a hashrate of 25.1 EH/s. Chairman Frank Holmes called the company “at the intersection of two powerful technology trends: Bitcoin and AI.” That’s a neat way of saying the firm wants exposure to both the original crypto rails and the compute gold rush that followed.

This is where the larger strategic shift becomes hard to ignore. Bitcoin mining is no longer just about stacking ASICs, plugging into cheap power, and waiting for the next bull run. It is increasingly becoming a real estate, energy, and infrastructure business. Miners are trying to become hybrid operators: self-mining, cloud mining, GPU hosting, AI compute, and whatever else can turn unused megawatts into cash flow. That may be smart. It also tells you the old “digital gold” story alone is not enough to keep the lights on for everyone.

The uncomfortable part is concentration. When difficulty is high and margins are thin, the winners are the miners with the newest ASIC rigs, the cheapest energy, and the strongest balance sheets. Everyone else gets forced out or scaled down. That means Bitcoin network security can become more concentrated even as the system remains decentralized at the node level. Thousands of nodes can still verify the rules, but mining power itself can still cluster around a handful of big operators. Economics does not care about ideology. It follows the cheapest electrons.

And now miners have a second enemy: compute thieves. China’s state-run media recently reported warnings about malicious AI agent skills packages that hijack GPU resources for BTC mining. In plain language, some bad actors are sneaking into AI tooling and using it to steal GPU power for their own gain. Microsoft has also warned about a cryptojacking campaign that combines SEO poisoning, trojanized system utility installers, and remote monitoring tool abuse to mine crypto through compromised GPUs.

Cryptojacking means secretly using someone else’s hardware to mine cryptocurrency. It’s basically digital parasitism with a power bill attached to somebody else’s name. The irony is rich: honest miners are struggling to make BTC mining profitable, while criminals and malware authors are hijacking GPUs to do the same thing on stolen compute. The machine is eating its own tail, and the tail is probably paying for the electricity.

There is also a darker lesson here about the value of compute. When GPUs become a prize worth stealing, it means the infrastructure underneath AI and crypto is now genuinely strategic. That’s good news for companies that own land, power, cooling, and access to hardware. It’s bad news for anyone assuming easy margins, passive income, or some magical “free” security model where fees will someday save everything without the market doing any heavy lifting.

Bitcoin itself is not broken. But Bitcoin mining economics are under real strain, and that strain is pushing the industry toward consolidation, hybrid business models, and a much harsher version of survival of the fittest. If BTC price strengthens, difficulty eases, or the fee market starts doing actual work, some of this pressure can reverse quickly. Until then, the miners with cheap power and efficient machines are the ones standing upright while everybody else gets bent into modern art.

What is happening to Bitcoin mining economics?

They are deteriorating again. Difficulty is high, BTC is trading below the estimated average all-in cost to produce a coin for many miners, and transaction fees are still too small to carry much of the network’s security budget.

Why are miners pivoting to AI and HPC?

Because AI and high-performance computing can generate more predictable revenue than pure Bitcoin mining. The same data center infrastructure that mined BTC can often be repurposed to host GPUs and sell compute power instead.

Why do transaction fees matter so much?

Bitcoin’s subsidy schedule keeps shrinking over time. Fees are supposed to become a bigger part of miner revenue eventually, but right now they are less than 1% of total revenue, which leaves miners highly exposed to BTC price swings.

Is Bitcoin mining becoming more centralized?

That’s the direction the economics are pushing. Miners with older hardware or expensive power get squeezed out, while the biggest and most efficient operators keep expanding. The network still has distributed validation, but mining power can concentrate pretty quickly.

Which companies are showing the strain most clearly?

Bitdeer, Cango, BitFuFu, and Hive Digital are all showing the pressure in different ways, from losses and executive turnover to reduced hashrate and major infrastructure pivots.

What is cryptojacking?

Cryptojacking is when software or malware secretly uses someone else’s hardware to mine crypto. In this case, GPU resources are being hijacked through AI-related tools, trojanized installers, and remote monitoring abuse.

What’s the biggest irony in all of this?

Bitcoin miners are rushing toward AI infrastructure for better margins, while criminals are using AI-adjacent tools to steal GPU power and mine BTC anyway. Everyone wants the same compute, and nobody is getting it cheap.

The bottom line is simple: Bitcoin mining is still essential to the network, but the business is getting tougher, more capital intensive, and more concentrated. The easy-money era is long gone. The survivors will be the miners who can monetize electrons better than everyone else, whether that means mining Bitcoin, renting GPUs, or turning old warehouses into AI cash machines. The rest? They’ll learn the hard way that hashpower is only valuable when the economics stop acting like a slap in the face.