Vitalik Buterin Explains How AI Could Push Bitcoin Miners Out

Why BitcoinBTC-0.13% Miners and AI Networks Are Fighting for Power

EthereumETH-0.46% co-founder Vitalik Buterin shared an interesting observation on social platform X on October 17, 2026. He pointed out a fundamental difference in how Bitcoin miners and artificial intelligence (AI) companies spend their money.

According to Buterin, Bitcoin mining costs are split almost equally between buying hardware and paying for electricity. In contrast, AI inference operations spend almost their entire budget on high-tech compute hardware.

This cost difference matters right now because both industries are increasingly competing for access to the exact same power grids around the world.

Breaking Down the Cost to Mine One Bitcoin

To illustrate his point, Buterin shared a detailed cost breakdown table. The data showed that the total expense to mine a single Bitcoin added up to roughly $93,700.

Here is where every dollar goes in that mining cost model:

• Direct Electricity: $40,600 per coin, accounting for 43.3% of the total cost.
• Additional Electricity: $2,400 per coin, bringing total power expenses to roughly 46%.
• Compute Capex: $37,600 per coin (about 40% of costs), which covers hardware, cooling gear, power infrastructure, and financing.
• Operational Overhead: The remaining 14% covers facility rent, labor, maintenance, insurance, taxes, and bandwidth.

This breakdown clearly demonstrates that electricity and hardware capital expenditures are nearly neck-and-neck for Bitcoin mining firms.

Comparing Mining Costs Against Market Reality

Interestingly, the $93,700 total cost in Buterin's table sits above recent estimates from Wall Street firm JPMorgan, which calculated the average Bitcoin production cost at around $85,000.

Because the total cost estimate was higher than Bitcoin's market price at the time, an average miner under these parameters would not cover their full operational expenses.

As a reminder, past returns or historical digital asset production costs do not guarantee future performance or profitability.

Why AI Plays by Different Power Rules

While miners spend nearly half their budget on electricity, AI inference completely flips that ratio. AI inference involves running trained artificial intelligence models to process user requests.

Because running AI models relies heavily on extremely expensive specialized chips, hardware costs dominate the bill. Consequently, electricity makes up a much smaller percentage of an AI operator's expenses.

This financial dynamic gives AI companies a unique edge. They can easily absorb higher electricity prices than Bitcoin miners can because power is a much smaller fraction of their total bill.

The Great Tug-of-War for Energy Sites

This spending gap is already actively reshaping the market. Currently, many Bitcoin mining operations hold pre-approved grid capacity permits that AI developers desperately need.

Because local regulations and restrictions on new data centers are spreading, AI firms cannot build new facilities fast enough. This has led some industry experts to ask whether AI will eventually price miners out of expensive power locations entirely.

However, others are looking to combine the two industries. BitMEX co-founder Arthur Hayes, for example, is building a Bitcoin-style AI mining network that pays GPU owners for providing processing power.

Ultimately, cheap electricity remains the primary advantage for Bitcoin miners. Meanwhile, access to cheap capital favors AI companies. The upcoming contest for energy sites may come down to who can finance hardware the fastest.