Cryptomining Reinvents Itself: $500M for AI Data Centers

SWI Group, through its subsidiary AiOnX, has acquired a 77% stake in cryptocurrency miner Genesis Digital Assets (GDA) for $500 million. The deal includes 15 crypto mining data centers located in North Carolina, South Carolina, Texas (USA), and two sites in Sweden. The main strategic asset: 1.3 gigawatts (GW) of electrical power already contracted and available. This move is not isolated; it responds to a global trend where energy has become the most critical resource for the expansion of artificial intelligence (AI).

What happened?

The transaction, reported by TechRadar and DataCenterDynamics, grants SWI Group operational control of GDA, although the remaining 23% stays with the original founders. The 15 data centers total an energy capacity of 1.3 GW, enough to power approximately one million US homes. Of these, 13 are in the United States (concentrated in Texas, a state with energy surplus and favorable regulations) and two in Sweden, a country with abundant hydroelectric and wind power. The infrastructure includes connections to power grids, substations, and already approved environmental permits, drastically reducing construction timelines compared to greenfield projects.

SWI CEO Max-Hervé George stated: 'Energy connectivity is the most valuable resource in digital infrastructure today, and converting legacy cryptocurrency mining infrastructure into high-performance computing (HPC) and AI is the best and highest use of these assets.' The company has been investing in energy connectivity since 2020, according to TechRadar.

Why it matters

Energy demand for AI data centers is growing exponentially. According to a study by the International Energy Agency (IEA) cited by TechRadar, data centers could consume up to 20% of global electricity by 2030, up from the current 1-2%. Another report from Gartner indicates that 40% of AI projects could face energy constraints by 2025. Cryptomining, on the other hand, faces shrinking margins: the Bitcoin halving in April 2024 reduced the block reward to 3.125 BTC, compressing miners' profits. Additionally, Bitcoin price volatility (which fell 50% in 2022) makes the business risky.

Converting existing mining infrastructure is faster and cheaper than building from scratch. While a typical AI data center can take 3 to 5 years to plan and build, repurposing a crypto mining center can be completed in 12-18 months, according to industry estimates. Moreover, AI capacity lease contracts offer margins of up to 85% with multi-year revenue, according to CoinDesk. In contrast, Bitcoin mining margins range between 30% and 50%, depending on electricity prices.

However, the ASICs (application-specific integrated circuits) used in cryptomining are not reusable for AI. The real value of the acquisition lies in the contracted energy, physical facilities (cooling, security, network connectivity), and permits. SWI plans to replace ASICs with NVIDIA GPUs (such as H100 or B200) and HPC hardware, according to industry sources.

Market consequences

This repurposing trend benefits both industries: miners gain liquidity and an exit to a more stable sector; AI providers get immediate power. More mining companies are expected to follow this path. Previous examples include Core Scientific, which signed an agreement with CoreWeave in 2023 to lease 200 MW of its mining capacity for AI. Another case is Hut 8, which converted part of its site in Alberta, Canada, to host GPUs. According to a J.P. Morgan analysis, up to 20% of global mining capacity could be repurposed for AI by 2026.

SWI's operation also reflects a structural shift: energy has become the most valuable resource in digital infrastructure. Companies like Amazon, Microsoft, and Google are directly buying nuclear and solar power plants to feed their data centers. In this context, the acquisition of GDA gives SWI a significant competitive advantage over other operators struggling to secure energy.

However, the concentration of power in large players like SWI could raise costs in the long run. If few operators control available energy, prices for startups and small businesses could increase. Additionally, massive repurposing could reduce the computing capacity available to the Bitcoin network, affecting its security and decentralization. According to a CoinMetrics report, Bitcoin's hash rate has already dropped 15% since the halving, and this trend could accelerate.

What readers should know

For startups and companies needing computing power, this repurposing can accelerate the availability of AI cloud services, especially in regions with energy constraints. However, competition for these resources will remain intense. Readers are advised to monitor alliances between miners and AI providers, as well as energy regulations in their countries. SWI's operation is a milestone marking the beginning of a new era where energy is the new oil of the digital economy.