As global investment in artificial intelligence climbs beyond $500 billion, a growing number of crypto mining operators are repositioning themselves for a new phase of infrastructure demand: AI data centers. The logic is straightforward. Many miners already control sites with power access, cooling setups, operational expertise, and experience running power-intensive computing equipment around the clock. As AI workloads become one of the hottest segments in technology, that installed base is increasingly being viewed as a launchpad for a broader compute business.
So far, much of this transition has been concentrated in North America and parts of Western Europe. These regions remain the default destination for large-scale AI infrastructure because they offer relatively stable political systems, mature capital markets, and established network connectivity. For investors and operators looking to deploy quickly, those factors still matter. But the article argues that the same markets now face an increasingly serious obstacle: power.
AI Compute Is Far More Energy Intensive Than Mining
The energy gap between bitcoin mining and advanced AI workloads is becoming central to the infrastructure debate. According to the source material, bitcoin mining may run at roughly 8–10 kW per rack, while modern AI data centers built around top-tier GPUs can reach 50 kW per rack and in some cases exceed 100 kW per rack. That kind of density creates not only much larger electricity demand but also significantly more heat, making cooling and grid availability major constraints.
Those pressures are already showing up in established AI hubs. The report points to highly visible strain on local grids in places such as Northern Virginia and Europe’s FLAPD regions. Utilities in those areas are being pushed to accelerate generation projects just to keep up with data-center demand. In other words, the best-known AI markets still offer access to capital and infrastructure, but they are no longer frictionless environments for expansion.
Nick Rose’s Case for the Developing World
Web3 veteran and investor Nick Rose believes the next major opportunity may lie outside the standard map of AI infrastructure. Through his latest venture, Orion Compute, Rose is challenging the assumption that developing markets are too risky or too immature to become serious data-center locations. His thesis is that these regions are not simply cheaper alternatives; in some cases, they may offer a structural advantage that Western markets increasingly lack.
Rose argues that many developing economies are both marginalized and underused when it comes to digital infrastructure investment. While they are often overlooked because of perceived regulatory uncertainty, they may possess abundant power resources that remain underutilized. In his view, the traditional narrative that operators must prioritize the West because of internet speed or market prestige is becoming less convincing as power shortages and rising electricity costs become more important.
The article highlights Rose’s claim that developing markets are producing “tons and tons of power” that is extremely cheap and not fully absorbed by local demand. That dynamic can translate into minimal curtailment and better continuity of supply. For AI operators or miners, reliable access to low-cost power is one of the most important variables in long-term profitability. Rose’s argument is that this factor may matter more than being located in a fashionable technology corridor.
Why Power Economics Matter More in the AI Era
The shift from mining to AI is not just a change in end market; it is a change in how infrastructure value is measured. In bitcoin mining, energy has always been a decisive cost. In AI, it becomes even more critical because the hardware is expensive, utilization matters enormously, and interruptions can severely undermine economics. If a site can deliver stable electricity at low marginal cost, it gains an advantage that extends beyond a single market cycle.
That is where Rose sees a disconnect between investor perception and operating reality. Building data centers in established economies may look safer on paper, but he contends that it is not necessarily the most effective way to maximize shareholder value over time. If energy is constrained, expensive, or politically contested, the long-term return profile can deteriorate even in otherwise attractive jurisdictions.
Orion Compute’s Phased Risk Strategy
Rose does not dismiss the risks associated with developing markets. Critics of his thesis argue that weaker regulatory frameworks can increase uncertainty, especially when operators are considering the deployment of high-value AI equipment such as Nvidia H100 GPUs. That concern is especially relevant for capital-intensive infrastructure where legal predictability, import conditions, contract enforcement, and policy continuity all affect deployment decisions.
To address those concerns, Orion Compute is not pursuing an all-in strategy. Instead, the company plans to grow in tandem with local market development. According to Rose, Orion will start by deploying lower-cost AI hardware, such as A100 GPUs, rather than immediately scaling into more expensive H100 systems. This phased approach is designed to reduce capital expenditure exposure while allowing the company to establish an energy footprint and operational presence in targeted regions.
The idea is to build local capabilities first, then upgrade technology as economic conditions, infrastructure quality, and policy frameworks improve. In theory, that gives Orion more flexibility: it can participate in early-stage opportunities without taking on the full risk of top-end hardware deployment from day one. Rose presents this as a practical middle path between ignoring emerging markets entirely and overcommitting too early.
On-Grid and Off-Grid Infrastructure
Another notable aspect of Orion Compute’s strategy is its dual-purpose infrastructure model. The company intends to develop both on-grid and off-grid deployments, with a core emphasis on ultra-low-cost energy. That reflects a broader trend already familiar to parts of the mining industry, where access to stranded, remote, or otherwise underused energy sources can become a competitive differentiator.
Rose also pointed to collaboration with Terra Solis and its low-cost energy technologies, which he said are not limited to a single geography. The broader message is that Orion is trying to build around cost discipline rather than simply chasing AI hype. In a sector where enthusiasm can lead to overbuilding or inflated assumptions, controlling the cost base may prove more important than headline growth alone.
A Broader Industry Transition
The article ultimately frames the crypto miner-to-AI transition as part of a larger restructuring of digital infrastructure economics. Miners are moving beyond a single-revenue model and exploring how their sites, power relationships, and operational expertise can support more diversified compute businesses. At the same time, the AI boom is exposing how fragile power availability can be in even the most advanced markets.
Rose’s view is that the next winners in data-center expansion may not be determined solely by financial depth or established reputation. Instead, they may emerge from regions where electricity is cheap, abundant, and underused—and where operators are willing to build patiently rather than pursue maximum deployment at maximum risk. Whether that thesis proves right will depend on execution, regulation, and demand growth, but the argument reflects a meaningful shift in how the market is starting to think about compute infrastructure.
For now, one conclusion is clear: as AI compute scales, the competition is no longer just about chips, software, or capital access. It is increasingly about energy. And in that race, emerging markets may be better positioned than many investors have assumed.