Ethereum co-founder Vitalik Buterin reignited a discussion within the crypto ecosystem today, highlighting a key difference in cost structures between Bitcoin mining and AI inference.
Buterin referred to this as an interesting fact and briefly explained that both Bitcoin and AI rely on computational resources, but in different ways.
Buterin discusses Bitcoin mining costs
Buterin pointed out that computing hardware and electricity play different roles in Bitcoin and the AI industry, and explained the cost structures of both.
According to Buterin, the cost of Bitcoin mining is roughly split evenly between investment in computing hardware—capital expenditure (CapEx)—and electricity consumption.
Therefore, miners must make substantial investments in specialized equipment and pay significant energy bills to maintain operations.
On the other hand, AI inference refers to using a trained AI model to generate responses or perform tasks, and this process has a different cost structure.
Buterin believes that, unlike Bitcoin, AI inference costs are almost entirely dominated by computational hardware investment, as they do not rely on high electricity consumption.
Comparison of Bitcoin and AI
Notably, Buterin’s comparison clearly shows that Bitcoin miners require powerful hardware and a stable power supply to compete for block rewards, making energy costs a significant portion of their production expenses.
In contrast, AI relies heavily on computing infrastructure capable of handling a large volume of requests.
This means that the cost of acquiring and maintaining such infrastructure represents a significant portion within the broader economic structure supporting AI systems.