Bitcoin Mining and AI Head in Opposite Directions, Says Alex Thorn

"If local models keep getting smaller, cheaper, and more efficient, AI may become increasingly personal and on-device," he said.

Bitcoin  News

Bitcoin mining has grown steadily more centralized over time, while artificial intelligence may be moving the other way, according to Galaxy Research head Alex Thorn.

Thorn noted that BTC mining started with individuals running software on personal computers but has since shifted to industrial-scale operations requiring ASIC hardware and large energy infrastructure. AI, by contrast, began in

data clusters and may decentralize as open-source models become lighter and cheaper to run locally. "If local models keep getting smaller, cheaper, and more efficient, AI may become increasingly personal and on-device," he said.

The divergence matters for crypto's foundational promise. Thorn said that if BTC mining continues to concentrate among fewer, larger operators, it could raise concerns about the network's long-term resilience to political or regulatory shocks.

Crypto exchange KuCoin reported on Friday that mining BTC has become increasingly unviable in parts of the United States, with costs exceeding $100,000 per coin in some regions due to rising energy prices. This is driving hash rate toward lower-cost jurisdictions, with Paraguay and Ethiopia emerging as leading destinations because of surplus hydroelectric power.