The boom in artificial intelligence (AI) is not a bubble, but a structural shift that will last 20–30 years. This assessment was given by Alpine Fox founder Mike Alfred at the Energy Investors Forum in Dallas. However, for Bitcoin miners seeking to diversify into this area, the key question is choosing a model: remain an infrastructure provider or take the risk and invest in the GPUs themselves.
Alfred emphasized that demand for AI infrastructure will remain resilient even during periods of corrections and economic downturns. But miners will have to choose between two fundamentally different strategies. The first is owning GPUs on their own premises. This promises higher revenue and control, but comes with serious risks: the need to finance the purchase of expensive equipment, rapid chip obsolescence, and the obligation to ensure their constant utilization. "If you decide to own them, you must have a compelling reason," the expert noted.
The second, more conservative model is colocation. The miner provides the client with space, electricity, cooling, and water, while the client independently places servers and GPUs. Alfred compared this scheme to real estate, making it much more attractive for raising financing. According to him, about 80% of Alpine Fox's investments are concentrated in companies IREN and Cipher Digital, which started as Bitcoin miners and then switched to AI. Among the recent deals are IREN's five-year contract with Microsoft for $9.7 billion to supply cloud infrastructure on Nvidia GB300 chips and Cipher's ten-year agreement with Fluidstack to host 168 MW of equipment in Texas.
Mining Sites Are Not Ready for AI
Most existing cryptocurrency mining facilities are difficult and expensive to retrofit into AI data centers. Panelists at the forum discussed the concept of mining as a flexible load: it can temporarily monetize excess electricity during the construction phase of an AI campus but is shut off when power is needed for more expensive computing. However, as SATOKIE CEO Jay Zapata noted, when competing for a suitable substation, mining will always lose to AI in business attractiveness.
Investors are increasingly evaluating miners through the lens of available power and project portfolio, rather than Bitcoin production volume. Examples include TeraWulf's 20-year agreement with Anthropic for $19 billion and Galaxy Digital's 15-year deal with CoreWeave for $4.5 billion. Both cases involve a complete pivot of facilities toward AI, rather than a sustainable coexistence of the two models. The cost of building a mining facility is $200,000–300,000 per 1 MW, while for AI infrastructure it is over $10 million per MW, excluding chips. The intermediate scheme only works with cheap electricity (up to $0.06–0.08 per kWh); at higher prices, the developer is better off going straight to AI or simply holding onto power access.
Typical mining facilities are located in remote areas, built to simplified standards, and lack necessary redundancy. AI tenants require redundant fiber optic lines, backup generation, complex cooling, and service availability at the 99.999% level. Compass Mining's Director of Mining Shanon Squires stated that the vast majority of new sites are unrealistic for Tier 3 data centers. The overlap between the two models ends at the land, substation, and step-down transformers.
Panelists identified sites with their own generation as a more sustainable option. Gas power plants often have excess capacity that mining can absorb while the AI facility uses guaranteed load. However, expensive GPUs are poorly suited for such flexibility, and long-term contracts carry concentration risk. According to panel estimates, before the next Bitcoin halving in 2028, at least one major tenant, lender, or landlord in the AI segment could default.
My comment: The AI and mining markets are converging, but not as quickly as investors might hope. Miners considering AI as a lifeline from declining profitability should remember: this is not just a change of business model, but a move into a completely different league, where competition for capital and technology is significantly higher. Those who can adapt their infrastructure and choose the right strategy will gain an advantage, but most risk being left behind.