Divergent Paths: Bitcoin Mining Centralizes as AI Shifts Toward the Edge

Alex Thorn, head of Galaxy Research, posits that Bitcoin mining and artificial intelligence are following opposite trajectories regarding decentralization. While Bitcoin began as a peer-to-peer activity conducted on personal computers, it has evolved into an industrial operation dominated by specialized ASIC miners and massive data farms, raising questions about long-term network resilience. Conversely, AI—which launched in centralized, high-cost hosted clusters—is seeing a push toward 'Edge AI.' This involves deploying and running models directly on local devices to bypass memory bottlenecks and data scarcity, potentially allowing open-source models to close the gap with proprietary frontier models. The financial scale of this shift is significant. According to Grand View Research, the global AI edge market is projected to grow from approximately $25 billion in 2025 to $119 billion by 2033. This growth is being driven by the rapid expansion of the Internet of Things (IoT), a demand for low-latency processing, and an increased focus on localized data privacy. Simultaneously, Bitcoin mining is facing economic headwinds in the United States. In some regions, surging energy costs have pushed the cost to mine a single BTC above $100,000, triggering a geographic migration of hash rate toward the 'Global South.' Paraguay and Ethiopia have emerged as primary destinations due to their surplus of hydroelectric power. While the industrialization of mining suggests a trend toward centralization, this geographic dispersion across different continents may actually enhance the network's security by making it less vulnerable to the political or environmental shocks of any single nation.