Crypto Legacy and AI Future in Georgia: Can Energy Infrastructure Become a Platform for Supercomputing?

Georgia was once one of the world’s leading hubs for cryptocurrency mining. How did a small Caucasian country, carrying the heavy energy legacy of the Soviet era, manage to secure second place on the global crypto map? And today, as the boom of artificial intelligence (AI) ushers in a new era of data centers, can Georgia’s accumulated infrastructure and expertise transform into a platform for AI supercomputing? This is not only a technological challenge; it is also a question of energy policy, economic development, and digital sovereignty.

The Era of Crypto Mining: Georgia’s Path to “World Top-2”

In the mid-2010s, Georgia became a “paradise” for crypto investors. According to NPR, low hydroelectric tariffs and lax regulation propelled the country to second place globally in mining profits by 2018, just behind China. At the heart of this boom were Bitfury’s facilities in Gori and Tbilisi: a 20 MW center opened in Gori in 2014, followed by a 40 MW site in Gldani in 2015. Both were notable for their pioneering cooling system — two-phase immersion — which delivered a world-class efficiency index of 1.02. The success was not limited to large investors. By 2018, Cointelegraph estimated that around 200,000 Georgians were engaged in crypto mining, often by building small “home farms” with graphics cards. The government’s liberal tax regime fueled this expansion, exempting individual crypto income from VAT and income tax.

Energy Consequences: When Mining Becomes a “Big User”

This boom placed a heavy burden on Georgia’s energy system. The surge revealed both opportunities and risks. On the one hand, Georgia is still branded as a “green energy” country, since most of its electricity comes from hydropower. On the other, seasonal deficits and reliance on winter imports mean that supplying power for mining — or for future AI data centers — could push the grid to its limits.

Global Context: From Crypto Farms to AI Data Centers

Bitcoin’s 2024 halving slashed profitability, undermining the economics of many mining operations. At the same time, demand for AI training infrastructure has soared to unprecedented levels. Analysts at IDCnova note that many miners now stand at a strategic crossroads: repurposing their energy-intensive facilities for AI. The logic is simple. Mining farms already have what AI centers need — high-capacity electricity connections, advanced cooling, and remote management experience. For example, U.S. operator Core Scientific transferred part of its 5 GW portfolio to CoreWeave, an AI cloud provider, in a multibillion-dollar deal. Yet the differences are substantial. Bitcoin ASIC chips are useless for AI, which requires powerful GPUs, low-latency networking, and specialized software stacks. Cooling systems designed for ASICs often prove inadequate for AI chips, which demand cleaner airflow and advanced thermal management. Moreover, high-speed optical fiber is critical to AI, and without it even cheap electricity cannot make a location competitive.

Georgia’s Position: Advantages and Obstacles

For Georgia, this is a particularly delicate dilemma. The advantages are clear. The Gldani center is one of the most energy-efficient sites worldwide. Georgia’s reliance on hydropower aligns with global trends toward low-carbon data centers. Owning domestic AI centers could strengthen digital sovereignty, allowing Georgian institutions to process sensitive data locally and enabling startups and universities to access GPU time for research. But the obstacles are serious. Transitioning to AI requires thousands of high-end GPUs costing hundreds of millions of dollars. Georgia’s fiber optic infrastructure lags behind regional competitors, and completion of the Black Sea Digital Corridor project will take years. AI servers are far more power-hungry than ASIC miners, raising new risks of shortages. The domestic labor market lacks the specialized expertise in high-performance computing and AI infrastructure. And finally, demand is limited: the local market for large-scale AI training is small, while neighboring Turkey and Azerbaijan are already building 100 MW “AI-ready” facilities — dwarfing Georgia’s 20–40 MW legacy sites.

Possible Scenarios

Several paths lie ahead. Georgia could pursue a hybrid model, repurposing parts of existing facilities into GPU infrastructure for local demand. It could partner with global firms, replicating deals like Bitfury–Nvidia or CoreWeave–Core Scientific, in exchange for technology transfer and market access. Or it could choose to prioritize other sectors — agriculture, tourism, education, renewable energy — where investments may yield faster and broader economic returns.

Conclusion: An Unresolved Path Requiring Long-Term Vision

Georgia stands at a transitional moment. Its crypto boom built infrastructure and financial instruments that once placed the country among global leaders. Yet the AI wave demands an entirely new level of connectivity, skills, and vision. Ultimately, the nation’s choice will depend on how it balances economic benefits with energy security and what priorities it sets in its development strategy. Georgia may not match Turkey or Azerbaijan in scale, but with careful planning, it could still become a niche “AI-ready” hub. The open question is whether the country will seize this legacy as a springboard into the AI era — or let it fade into a short-lived chapter of history.