The mining of the first cryptocurrency is becoming increasingly centralised, while the field of artificial intelligence is moving in the opposite direction, according to Alex Thorn, head of Galaxy Research.
bitcoin mining began decentralized (CPUs, GPUs) and became centralized (ASICs, industrial-scale farms)
AI may follow the opposite path: it started centralized in giant hosted clusters, but as frontier model gains slow (from data scarcity, context limits, and memory bottlenecks)… pic.twitter.com/J2indQsTt8
— Alex Thorn (@intangiblecoins) April 12, 2026
He noted that in the early stages, mining the first cryptocurrency was accessible to any personal computer owner. However, with the growth in network difficulty and the advent of specialised equipment, digital gold mining has come under the control of large industrial operators with access to cheap electricity.
Thorn suggested that AI might follow a different scenario. According to him, open models will gradually catch up in performance with closed ones, becoming more compact and affordable. Then, artificial intelligence will operate directly on the user’s device rather than in cloud clusters.
Edge AI
Deploying artificial intelligence models on local devices is known as Edge AI.
According to Grand View Research, by the end of 2025, the volume of this sector reached $24.9 billion. Experts predict that this year the figure will grow to $29.9 billion, and by 2033 it will exceed $118 billion.
The market is actively developing due to the proliferation of IoT technologies, according to the analysts’ report. This creates a steady demand for real-time data processing and stimulates the implementation of AI automation across various industries. There is also increasing attention to data privacy and edge computing technologies.
Back in late March, CoinShares analysts recorded the capitulation of 20% of bitcoin miners. The decline in industry revenues amid market challenges is forcing miners to increasingly adapt to the needs of AI.