As global investment in artificial intelligence rises beyond $500 billion, a growing number of cryptocurrency miners are repositioning themselves as AI infrastructure operators. The shift is not simply a response to hype. It reflects a broader effort to reuse existing power-heavy infrastructure, operational expertise, and facility footprints to serve the expanding demand for AI compute.
According to Web3 veteran and investor Nick Rose, this transformation may ultimately favor regions that many investors still overlook. While most large-scale AI data center development remains concentrated in North America and parts of Western Europe, Rose argues that emerging markets may offer a more durable cost and power advantage, particularly for companies that care about long-term efficiency rather than short-term narrative momentum.
Why miners are moving toward AI infrastructure
The economics behind the pivot are increasingly clear. Crypto miners already understand how to build and operate energy-intensive facilities, source power, manage uptime, and deploy hardware at industrial scale. Those capabilities overlap with what is required for AI data centers, even if the performance and cooling demands of AI hardware are far more intense.
Rose’s view is that the migration from bitcoin mining into AI hosting is less about abandoning one business for another and more about maximizing the utility of existing assets. In an environment where the race for AI compute capacity is accelerating, miners with access to power infrastructure are seeking new ways to monetize those assets.
Still, the requirements for modern AI are substantially different from those of mining. The report notes that bitcoin mining may run at around 8–10 kW per rack, while advanced AI deployments using cutting-edge GPUs such as Nvidia’s H100 can reach 50 kW per rack and often exceed 100 kW per rack. That jump in power density creates major challenges not only in electricity supply but also in cooling, heat dissipation, and grid reliability.
Why North America and Western Europe still dominate
For now, the largest share of AI data center expansion is still flowing into established markets such as the United States, Canada, and selected countries in Western Europe. These jurisdictions remain attractive because they offer deep capital markets, clearer regulatory frameworks, and mature digital infrastructure. For investors seeking immediate deployment, those factors make the West look like the safest destination.
But Rose argues that the apparent safety of those markets comes with mounting structural constraints. Electricity prices are rising, local grids are under pressure, and utilities are being pushed to accelerate new generation projects just to keep up with growing demand from AI workloads. Well-known hubs such as Northern Virginia and the FLAPD region in Europe have already experienced strain linked to the rapid concentration of high-density data center activity.
That means the traditional Western advantage may be narrowing. While policy stability and financing access remain important, they do not eliminate the fundamental bottleneck of power availability. In AI infrastructure, access to dependable energy at scale is not a secondary issue. It is often the deciding factor.
Rose’s case for emerging markets
Rose believes emerging markets are being undervalued because investors focus too heavily on perceived political and regulatory risk while overlooking a much more practical variable: energy. In his view, many developing economies have abundant electricity that is both inexpensive and underutilized, creating a favorable environment for compute-heavy operations.
He argues that one of the biggest advantages of the developing world is that these markets are still marginalized and untapped in the current data center buildout. Unlike oversubscribed Western hubs, some of these regions reportedly have “tons and tons of power” that is “extremely cheap.” Because local demand is lower, they can also offer minimal curtailment and a more continuous supply profile for operators that need consistent uptime.
For AI operators and mining companies alike, that combination matters. Cheap power improves margins. Underused grid capacity can reduce competition for electricity. And lower curtailment risk means infrastructure can operate more predictably. Rose’s central thesis is that these conditions may make emerging markets a more rational long-term base for compute infrastructure than established economies that look stronger on paper but face worsening energy constraints.
Managing the risk of developing-market expansion
Critics of this thesis argue that lower electricity costs do not cancel out the risks associated with weaker regulatory systems or less mature policy environments. That concern becomes even more serious when the hardware involved is expensive. Deploying premium AI assets such as Nvidia H100 GPUs in uncertain jurisdictions may be difficult to justify from a capital preservation standpoint.
Rose says Orion Compute is addressing that problem with a phased strategy rather than an aggressive all-in deployment. Instead of placing its highest-value hardware into developing markets from day one, the company plans to enter gradually and expand in parallel with the evolution of local economies, infrastructure, and policy frameworks.
At the early stage, Orion intends to use lower-cost AI hardware such as Nvidia A100s rather than immediately deploying H100s. The idea is to establish a regional footprint, build energy infrastructure, and limit upfront capital expenditure exposure while testing operating conditions. If the local environment improves over time, the company can then migrate toward more advanced systems.
This staged approach is designed to balance opportunity and caution. Rose presents it as a way to preserve investor capital while still capturing the upside of entering underdeveloped compute markets before they become crowded.
A dual-use model across AI and mining
Another notable part of Orion’s strategy is its intention to build dual-purpose infrastructure that can support both AI compute and mining-related operations. Rose said the company is focused on ultra-low-cost energy deployment and plans to develop both on-grid and off-grid infrastructure.
That emphasis is significant because it points to a deeper convergence between digital asset infrastructure and AI infrastructure. Both sectors depend heavily on power procurement, site development, equipment logistics, and uptime management. As a result, operators that can secure low-cost energy may gain optionality across multiple compute-intensive business lines.
Rose also indicated that Orion is working with Terra Solis on low-cost energy technologies that are not tied to a single location model. In practical terms, that suggests a strategy centered on flexibility: go where the power economics work best, then adapt hardware deployment to the maturity and risk level of each market.
The broader industry implication
The larger takeaway from Rose’s argument is that the next phase of competition in AI data centers may be less about who can attract the most attention and more about who can secure the best energy base. In recent years, public markets and private investors have placed enormous value on AI exposure. But the ability to operate profitably over time will likely depend on fundamentals such as electricity cost, grid access, and operational resilience.
That matters especially for crypto miners, many of whom are searching for ways to diversify their business models. AI represents a compelling adjacent opportunity, but only if operators can avoid overpaying for power or building in already congested regions. From that perspective, emerging markets may offer more than a speculative frontier. They may become a practical answer to the supply constraints now appearing in traditional data center hubs.
Rose’s thesis does not eliminate the real political and regulatory risks associated with developing economies. However, it does challenge a widely held assumption in infrastructure investing: that safety and profitability naturally align in established Western markets. If power becomes the dominant bottleneck of the AI era, then regions with cheaper, underused, and more reliable electricity could play a much larger role than current investment patterns suggest.
For the crypto mining industry, this could mark the beginning of a more strategic transformation. The future may belong not only to those who own machines, but to those who can place them where energy is abundant, costs are structurally lower, and infrastructure can scale with demand. In that scenario, emerging markets may move from the margins to the center of the global compute map.