Artificial intelligence is changing the Bitcoin mining industry by increasing demand for electricity, data centers, land and investment. But ViaBTC CEO Yang Haipo says AI will not push Bitcoin mining out of the market.
Instead, the two industries will compete for resources while Bitcoin mining moves toward cheaper and more flexible power sources.
In a recent analysis, Yang said Bitcoin mining has already gone through a major adjustment. Network hashrate rose above 1.1 zettahashes per second (ZH/s) in October 2025 before falling to more than 900 exahashes per second (EH/s) several times this year.
Mining difficulty also recorded major declines, falling 11.16% in February and another 10.09% in June. At the same time, large mining companies are moving into AI and high-performance computing (HPC).
The economics of mining are driving this shift. Core Scientific reported a negative 56% profit margin from its own Bitcoin mining in the second quarter, while its data-center hosting business generated nearly $80 million in profit. TeraWulf also reported that HPC leasing made up about 71% of its revenue during the same period.
These numbers show why mining companies are turning toward AI infrastructure.
However, AI is not directly taking Bitcoin mining machines. Bitcoin ASICs are built to run the SHA-256 algorithm and cannot simply be converted into AI servers. GPU-based Bitcoin mining is also largely uneconomical today.
The real competition is for electricity, land, grid connections, capital, chip supplies and data-center infrastructure.
Yang says the most valuable asset for many mining companies is no longer their mining hardware. It is the infrastructure they have already built.
Mining companies have spent years securing land, building substations and obtaining grid connections. AI companies now place a high value on these resources because building new power infrastructure can take years.
This gives mining companies with ready-to-use power infrastructure an opportunity to sell or repurpose their sites for AI workloads. But AI cannot use every type of electricity efficiently.
Bitcoin mining has an advantage when electricity is cheap, remote or unreliable. Mining machines can use surplus electricity from solar farms, hydropower projects, associated gas and other energy sources that are difficult to sell or transport. Miners can also reduce their power use or shut down when electricity becomes expensive.
AI operations need much more reliable power because computing workloads require steady energy. Bitcoin mining is more flexible.
For example, surplus solar power that cannot reach the grid can be used to mine Bitcoin instead of being wasted. This makes mining a flexible buyer of electricity.
As large companies move some operations toward AI, older Bitcoin mining machines can enter the secondary market. A machine that is too expensive to run in a high-cost data center can still make money when bought cheaply and moved to a region with very low-cost electricity.
This creates opportunities for smaller miners, energy companies and private operators. Over time, Bitcoin mining may become more spread out geographically, with different types of machines operating in areas where electricity costs are low.
Meanwhile, Bitcoin’s fifth halving in 2028 will cut the block subsidy by another 50%. This will put more pressure on miners and make cheap electricity, efficient equipment and strong cash flow even more important.
Yang’s main argument is that AI is not simply taking Bitcoin mining’s computing power. It is reallocating resources.
