Bitcoin Miners Bet Big on AI, But May Regret It Within a Year

Bitcoin miners are facing a significant downturn in profitability due to low prices and high network competition. In response, some of them are redirecting their resources towards artificial intelligence (AI) infrastructure. Data from CryptoSlate shows that Bitcoin's price has dropped by nearly 50% since its October peak, reaching around $64,000. This decline, combined with weak transaction fees, is squeezing miner revenue.

Despite the challenges, AI infrastructure has become an attractive option for miners due to the higher prices they can command for reliable electricity and long-term capacity. Some of the industry's largest operators are already converting their facilities and signing multi-year computing contracts. However, André Dragosch, head of research at Bitwise Europe, is warning that this shift may be premature.

Dragosch believes that expectations for AI compute demand could take longer to materialize than current investment implies. He also thinks that Bitcoin is approaching the end of its downturn and that a recovery in prices could restore the economics of mining. This creates a potential timing problem, where miners may commit capital and power capacity to AI just as a recovery in Bitcoin prices makes mining profitable again.

The shift towards AI infrastructure involves repurposing access to electricity, grid connections, land, and data-center infrastructure. However, this transition can be expensive, with estimates suggesting that AI facilities cost between $8 million to $15 million per megawatt, compared to $700,000 to $1 million per megawatt for Bitcoin mining infrastructure.

Dragosch's contrarian view is magnified by the fact that intense competition among technology companies could produce excessive investment in AI-related capital expenditure. If commercial returns fail to keep pace with spending, this could lead to a slowdown in new infrastructure development and weaken demand for additional capacity.