Can AI Outcompete Bitcoin Miners for Cheap Power?

Proving that bitcoin mining remains a profitable activity is getting tougher at the largest and most expensive facilities.

The network hashrate, a metric for the total computing capacity protecting Bitcoin, topped 1.1 ZH/s in October 2025 but thereafter dropped toward 900 EH/s on multiple occasions. Mining difficulty fell 11.16% in February 2026 and another 10.09% in June.

Put simply, enough miners went offline that the Bitcoin system had to reduce the challenge of mining for those that remained active.

Meanwhile, several of the biggest mining firms are discovering more profitable opportunities elsewhere. Core Scientific posted a negative 56% gross margin from self-mining during the second quarter, whereas its data-center colocation segment brought in almost $80 million in gross profit.

At TeraWulf, HPC leasing contributed roughly 71% of quarterly revenue. Consequently, leasing high-powered computing equipment for artificial intelligence and cloud computing is yielding greater returns.

So, is AI edging out bitcoin miners, and what becomes of mining if that happens?

The Struggle for Prime Power

AI hardware and bitcoin mining equipment cannot be swapped. Graphics processors designed for AI are typically not cost-effective for bitcoin mining, and bitcoin ASICs are unable to run extensive AI models. The rivalry focuses instead on chip-production capacity, capital, land, infrastructure and, most critically, dependable electricity.

For AI firms, a location that already has substations, grid capacity and fiber connections is much more valuable than raw land next to a power plant. AI infrastructure needs rapid deployment, yet large power projects commonly require years to finish.

Numerous mining companies acquired suitable land and grid hookups before AI heightened competition for such sites. Those locations can now be worth more as AI data centers than as mining operations. Therefore the industry shift is not just about selling electricity. It involves monetizing power access that is already in place.

That benefit does not extend to every energy source.

📊 Miner Poll

What’s the biggest challenge for Bitcoin miners right now?

Vote below and tell us why you chose your answer. Your insights help the entire mining community learn from each other. 👇

— ViaBTC (@ViaBTC) July 30, 2026

AI training and inference usually demand stable, highly available electricity. Bitcoin mining can function with greater flexibility. Mining machines can start up when extra power is available, cut consumption when supply drops and turn off when the grid faces strain.

For instance, a factory equipped with rooftop solar can use surplus midday generation to run a small set of mining machines once its regular production needs are fulfilled. The machines do not have to stay on continuously. Their role is to create value from electricity that would otherwise be wasted or sold back to the grid at a low rate.

The same idea applies at a bigger scale. Energy company ENGIE has stated it is considering battery storage or bitcoin mining at its Assú Sol solar project in Brazil, where transmission limits stop all available generation from being used.

Intermittent solar and wind power can feed AI, but generally only when paired with storage, grid electricity or another steady source. That additional infrastructure drives up expenses.

Mining is more suited to consuming electricity that is affordable precisely because it is intermittent, remote or hard to transmit.

Bitcoin mining is seeing declining profits, per Wintermute.

Many miners have moved into AI hosting or begun using their Bitcoin reserves as working capital just to keep operations running. pic.twitter.com/vPD56FlTkm

— BeInCrypto (@beincrypto) March 14, 2026

Hashrate Will Shift, Not Vanish

As big mining firms convert prime sites to AI, some of their machines are expected to enter the secondhand market. A rig that is unprofitable in an expensive data center may still be viable at a location with cheap hydropower, excess solar or stranded energy.

Lower equipment costs cannot offset costly electricity, but they cut upfront capital needs and shorten payback periods. Older, less efficient machines may remain economical where power is very cheap and continuous operation is not required.

This could reshape the mining industry's structure. Publicly traded firms will stay significant, but future hashrate growth might increasingly originate from private operators, smaller miners and energy producers that have direct access to underused power.

Bitcoin's difficulty adjustment also aids the network's response. When miners go offline, blocks initially take longer to appear. Difficulty then declines, letting the remaining machines earn more bitcoin for the same computing effort. Some equipment that was previously unprofitable may then come back online.

A lower hashrate still matters because it reduces the expense of attacking the network. Yet a temporary drop does not automatically indicate a security emergency. The system keeps moving toward a new equilibrium influenced by bitcoin's price, electricity costs and machine efficiency.

AI will drive up costs for premium power sites and make some mining models unprofitable. But it is improbable that AI will wipe out bitcoin mining altogether.

Rather, it is splitting two markets: dependable, infrastructure-heavy power will increasingly go to AI, while mining will gravitate toward cheaper and less conventional energy.

As long as underutilized electricity exists, miners will keep seeking ways to put it to use.