Bitcoin mining and artificial intelligence are moving in different directions in how computing power is distributed, according to Galaxy Research head Alex Thorn. Bitcoin began with home computers and now largely runs in large industrial facilities packed with specialized machines. AI, by contrast, is concentrated in massive data centers today, yet Thorn argues it could gradually shift toward personal devices and local computing.
Speaking on Sunday, Thorn said Bitcoin mining has moved far from its early home-based roots. Large warehouses and dedicated hardware now dominate the activity. AI may take the opposite path. He said open-source progress is narrowing the gap at a time when major models are running into memory and data constraints, creating more room for smaller local systems.
His view is straightforward: if local models continue getting smaller, cheaper, and more efficient, AI could become increasingly personal and run directly on consumer devices. That argument lines up with the growth outlook for Edge AI, where processing happens on the device itself instead of relying on a centralized cloud.
Edge AI market projected to reach $119 billion by 2033
Grand View Research estimates the global Edge AI market will reach $119 billion by 2033. The same market is expected to stand at roughly $25 billion in 2025. The expansion is tied to the growing number of connected devices and demand for instant data processing without waiting for a faraway server.
GVR analysts linked that momentum to the spread of the Internet of Things. They also pointed to rising demand for data privacy and localized intelligence at the network edge. For companies, that means some automated functions can stay local, reducing the need to send sensitive data back to a central hub.
Mining hardware gets less accessible to individuals as operations spread globally
A separate report released Friday by crypto exchange KuCoin said Bitcoin mining hardware has become harder for individuals to own, even as the physical locations of those machines are spreading across more parts of the world. Power costs are a major factor. In some parts of the United States, the cost of producing one Bitcoin can exceed $100,000, pushing mining economics under pressure.
Operators are responding by looking for cheaper electricity in countries such as Ethiopia and Paraguay, where hydroelectric power is more available. That does not mean mining is returning to a small-scale model. The business remains capital-intensive and industrial. What is changing is the map: large mining sites are being placed in a wider range of countries rather than remaining concentrated in only a few.
KuCoin said that spreading mining power across continents strengthens the network because it becomes less exposed to the political environment or grid conditions of any single nation. In its wording, this geographic decentralization makes the network less vulnerable to political or environmental shocks centered in one country.
Set side by side, the two sectors show different versions of decentralization. Bitcoin mining is becoming more industrial in ownership and operation while gaining broader geographic distribution. AI infrastructure is still heavily centralized, but model deployment may shift downward to local devices. The direction is not the same, though both point to a changing structure for where computation happens.