At the recent Energy Investors Forum in Dallas, a significant signal was sent to Bitcoin miners considering diversification into artificial intelligence. Alpine Fox founder Mike Alfred warned that the key choice for such companies is not whether to enter AI, but which operational model to choose.

According to Alfred, the boom in AI infrastructure is not a bubble, but a long-term trend that will persist for 20–30 years, despite possible corrections and economic downturns. However, miners should be extremely cautious in choosing their strategy.

Two Paths: GPU Ownership or Colocation

The expert highlighted two main models. The first is purchasing and owning graphics processing units (GPUs) at their own facilities. This path promises higher revenue and control, but carries 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," Alfred emphasized.

The second, more conservative model is colocation. In this case, the mining company provides the client with space, electricity, water, and cooling, while the client places their own servers and chips. Alfred compared this scheme to real estate, noting that it is much easier to attract financing for it.

Most Mining Sites Are Not Ready for AI

Participants in the thematic panel at the forum concluded that a significant portion of existing cryptocurrency mining facilities are difficult and expensive to retrofit into AI data centers. SATOKIE founder Jay Zapata noted that when competing for a suitable substation, mining will always lose to AI in terms of business attractiveness.

Building a mining facility costs $200,000–300,000 per 1 MW, while AI infrastructure costs over $10 million per MW, excluding chips. At an electricity price of $0.06–0.08 per kWh, the economics of mining become challenging, and it is more profitable for a developer to immediately pursue AI or simply retain access to capacity.

Hybrid Model: Mining as Flexible Load

The forum discussed a concept where mining acts as a temporary consumer of electricity during the construction phase of an AI campus or absorbs excess generation, shutting down when power is needed for more expensive computations. However, speakers noted that expensive GPUs are poorly suited for such flexibility, and long-term AI contracts carry concentration risk. According to participants' estimates, before the next Bitcoin halving in 2028, at least one major tenant or lender in the AI segment could default.

Expert Opinion from Cryptalist: The market is clearly signaling that miners should not chase the AI "hype" by purchasing GPUs. It is much wiser to focus on developing infrastructure—energy capacity, cooling, and logistics. Those who can offer reliable and scalable colocation sites will be in a winning position, minimizing risks and generating stable rental income.