Ethereum co-founder Vitalik Buterin posted on X on October 17, using two cost breakdowns to compare Bitcoin mining with AI inference: mining costs are roughly split between buying equipment (capital expenditure on computing power) and electricity, while running AI inference is almost entirely the cost of buying GPUs, cooling and power equipment, along with financing; electricity makes up only a small share.
According to the estimates he posted, mining one $BTC costs about $93,700 in total, including around $40,600 in direct power costs (about 43.3%). Including other power consumption, electricity accounts for about 45.9% in total. Equipment capital expenditure is about $37,600 (around 40%). The total cost of running one GPU for inference for a year is about $26,260, with equipment capital expenditure at around $19,900 (about 76%), direct power costs at only about $1,050 (around 4%), and total electricity costs at about 4.8%.
Both are computing-power businesses, but one is constrained by electricity prices, while the other is constrained by chips and financing. This is also often cited to explain why mining companies in recent years have shifted their power capacity and data centers toward AI and high-performance computing hosting. The post had received more than 260,000 views within a few hours. In the markets, at around 08:57 on the morning of October 18 (UTC+8), $BTC was trading at about $81,479 and $ETH at about $2,404.
#BTC #ETH #挖矿 does not constitute investment advice
According to the estimates he posted, mining one $BTC costs about $93,700 in total, including around $40,600 in direct power costs (about 43.3%). Including other power consumption, electricity accounts for about 45.9% in total. Equipment capital expenditure is about $37,600 (around 40%). The total cost of running one GPU for inference for a year is about $26,260, with equipment capital expenditure at around $19,900 (about 76%), direct power costs at only about $1,050 (around 4%), and total electricity costs at about 4.8%.
Both are computing-power businesses, but one is constrained by electricity prices, while the other is constrained by chips and financing. This is also often cited to explain why mining companies in recent years have shifted their power capacity and data centers toward AI and high-performance computing hosting. The post had received more than 260,000 views within a few hours. In the markets, at around 08:57 on the morning of October 18 (UTC+8), $BTC was trading at about $81,479 and $ETH at about $2,404.
#BTC #ETH #挖矿 does not constitute investment advice