As investment in artificial intelligence climbs beyond $500 billion, a growing number of crypto mining firms are repositioning themselves as AI infrastructure operators. The logic is straightforward: miners already control power-heavy facilities, operational expertise, and site development capabilities that can be adapted for high-performance computing. According to Web3 veteran and investor Nick Rose, however, the biggest opportunity may not be in the usual destinations such as the United States, Canada, or Western Europe, but in developing markets that remain largely overlooked.
AI demand is reshaping the mining infrastructure playbook
The article describes a strategic shift now underway across the digital asset and data center industries. Rather than treating AI as a separate vertical, some mining operators are repurposing existing infrastructure to serve the explosive demand for compute capacity. This is particularly relevant as training and operating large AI models require far more energy-dense environments than traditional bitcoin mining.
In bitcoin mining, rack density may sit around 8 to 10 kW per rack. By contrast, modern AI facilities equipped with advanced GPUs can run at 50 kW per rack, and in some cases exceed 100 kW per rack. That difference is not merely technical. It has major consequences for site design, cooling, transmission access, and utility planning. In major AI hubs such as Northern Virginia and Europe’s FLAPD corridor, rising power demand is already creating visible stress on local grids.
That strain is helping redefine the economics of where future compute infrastructure can be built. Mature markets still attract most investment because they offer political stability, deep capital pools, and established connectivity. But they are also becoming more expensive and more constrained, especially when it comes to electricity availability.
Why Rose believes developing markets have an edge
Nick Rose argues that the global narrative around data center expansion has been too narrowly centered on Western markets. In his view, many developing economies offer a structural advantage that has not yet been fully appreciated: abundant, low-cost, underutilized energy. He says these regions are often marginalized in investor thinking because of perceived regulatory and operational risk, yet that same neglect can create outsized opportunity.
Rose’s core point is that AI and mining businesses ultimately depend on reliable and affordable power. While Western markets may appear safer on paper, many are now facing a mismatch between rising energy demand and available supply. At the same time, electricity prices continue to rise, and utilities are being forced to accelerate generation buildouts just to keep pace with projected AI demand.
By comparison, Rose says developing markets in some cases produce large amounts of electricity that remain underconsumed. Because local demand is insufficient, the grid may experience little or no curtailment, creating an environment where compute operators can access consistent power at very low cost. For energy-intensive businesses, this can be a decisive advantage.
His argument challenges the long-standing assumption that high-speed connectivity alone is enough to justify concentrating data center growth in the West. As AI infrastructure becomes more power-centric, the location decision may increasingly be driven by energy economics rather than by legacy perceptions of market quality.
Risk remains real, so Orion Compute is taking a phased approach
Rose does not dismiss the risks associated with developing countries. Critics of this strategy point to weaker regulatory frameworks and policy uncertainty, especially when expensive AI hardware is involved. Deploying top-tier chips into less predictable jurisdictions can expose operators to significant capital risk if market conditions or rules change unexpectedly.
To manage that exposure, Rose says his venture, Orion Compute, is not taking an all-in approach. Instead, the company plans to scale gradually alongside the evolution of local economic conditions and policy environments. In the early stages, Orion Compute intends to deploy lower-cost AI hardware such as Nvidia A100s, rather than immediately committing to higher-end H100s.
This phased model is designed to limit capital expenditure while allowing the company to establish an energy and operational footprint in target regions. As infrastructure improves and local frameworks mature, Orion can later upgrade to more advanced compute assets. Rose presents this as a way to balance caution with upside, reducing downside risk for investors while preserving future expansion options.
Energy cost, not hype, is the center of the strategy
Another important part of Orion Compute’s thesis is its focus on cost structure rather than short-term AI hype. Rose says the company is building around ultra-low-cost energy deployment and intends to support both on-grid and off-grid infrastructure models. That flexibility could be significant in markets where transmission networks are uneven or where stranded energy can be monetized more effectively through colocated compute.
Rose also points to collaboration with Terra Solis and its low-cost energy technologies as part of this strategy. The broader idea is to build infrastructure that remains resilient across changing market conditions by keeping variable costs as low as possible. In an industry where revenue assumptions can move quickly, operators with the lowest operating costs often have the most durable business model.
This framing also mirrors a broader convergence now taking place between crypto mining and AI compute. Both sectors are deeply dependent on power access, thermal management, and hardware utilization. As a result, miners with experience in sourcing cheap electricity and deploying industrial-scale infrastructure may be better positioned than many expect to participate in the AI data center buildout.
A wider implication for the industry
The article ultimately suggests that the next phase of competition in AI infrastructure may not be won solely by those with the most capital or the most recognizable geography. It may be won by those that can secure the best energy economics over the longest period of time. That creates an opening for developing regions that have historically sat outside the center of global data center investment.
For crypto miners, this trend is especially important. What began as a niche digital asset infrastructure business is increasingly intersecting with one of the largest technology investment themes in the world. If miners can successfully convert their expertise into AI-ready facilities, and if firms like Orion Compute can prove that developing markets can host reliable, scalable compute, then the industry’s next growth chapter may look very different from the one centered on North America and Western Europe.
In that sense, Rose’s thesis is not simply about geography. It is about rethinking what matters most in the age of AI: not only capital access and policy stability, but the ability to secure continuous, low-cost power in places the market has yet to fully value.