Mining costs in some parts of the United States have surpassed $100,000 per Bitcoin, forcing operators to pack up and move. Paraguay and Ethiopia have emerged as top destinations offering surplus hydropower that keeps electricity bills low.
According to cryptocurrency exchange KuCoin, change is already underway, with hash rates actively moving toward what analysts call the “Global South.”
KuCoin argues that this geographic spread actually strengthens the Bitcoin network by reducing its exposure to political or energy shocks in any one country.
This is a different kind of decentralization than what Satoshi Nakamoto imagined. But decentralization is the same.
Opposite paths of the two technologies
While Bitcoin mining may grow more intensive in terms of hardware and industry scale, artificial intelligence may move in the opposite direction.
Galaxy research director Alex Thorn made this case Sunday, noting that AI began in large, corporate-controlled data centers.
Bitcoin mining started out decentralized (CPU, GPU) and then became centralized (ASIC, industrial scale farms).
AI may follow the opposite path. Although it is starting to be centralized on large hosting clusters, frontier models are slow (due to lack of data, context limitations, and memory bottlenecks)… pic.twitter.com/J2indQsTt8
— Alex Thorne (@intangiblecoins) April 12, 2026
Now that frontier models are constrained by data shortages, memory limitations, and context bottlenecks, open source alternatives are gaining ground. Smaller models are becoming more affordable and more capable. Some already run directly on phones and laptops.
“If local models continue to become smaller, cheaper and more efficient, AI can become increasingly personalized and used within devices,” Thorn said.
Bitcoin mining started in the opposite direction. Ordinary people once mined coins on their home computers. Those days are long gone.
Mining today requires access to specialized ASIC hardware or industrial-scale facilities. The gap between casual participants and serious miners has never been wider.
A $119 billion market is forming.
The push toward on-device AI processing has a name: edge computing. This means running AI models locally on the device itself, rather than routing data to a remote server.
According to data, the global edge AI market has been valued at approximately $25 billion in 2025. This figure is expected to approach $120 billion by 2033, according to projections from Grand View Research. This is an increase of almost 300% in eight years.
Growth is being driven by the proliferation of connected devices, demand for real-time processing, and growing concerns about data privacy. Industries that cannot afford delays, such as manufacturing, healthcare, and logistics, are among those driving adoption.
In the case of Bitcoin, the concerns run in a different direction. The increasing concentration of mining power raises questions about long-term network security.
A network where a few large players control the majority of the hash rate is more vulnerable to outages than a network spread across thousands of independent operators.
Geographically moving from the United States can alleviate some of this pressure. Whether that is enough remains an open question.
Featured image from Unsplash, chart from TradingView
editing process for focuses on providing thoroughly researched, accurate, and unbiased content. We adhere to strict sourcing standards and each page undergoes diligent review by our team of top technology experts and seasoned editors. This process ensures the integrity, relevance, and value of the content for readers.
