- Cryptocurrency mining firms are increasingly exploring offering AI processing power.
- They possess vital resources like substantial energy supplies and data center infrastructure.
- However, they require improved networking capabilities and AI-specific knowledge to succeed.
A wave of new participants is poised to enter the cloud computing landscape. Businesses initially established to power cryptocurrency mining are now setting their sights on the artificial intelligence sector, aiming to capitalize on what’s being called the 21st century’s digital gold rush. Crucially, they possess an abundance of a resource that other data center operators are struggling to obtain: electrical power.
“They control access to electricity, a resource even scarcer than Nvidia’s specialized chips,” stated Jack Gold, founder of J. Gold Associates, in a discussion with Fierce.
Roy Chua, Founder of AvidThink, echoed this sentiment, emphasizing in an interview that multi-gigawatt power infrastructures are extremely attractive within the power-constrained AI computing market.
A relevant example is CoreWeave’s recently publicized acquisition of Core Scientific. While CoreWeave, itself a company that transitioned from cryptocurrency mining to cloud computing in 2019, will gain Core Scientific’s infrastructure as part of the agreement, Chua highlighted the importance of the power resources acquired. As detailed in the acquisition announcement, it’s a significant amount.
Core Scientific offers CoreWeave access to 1.3 GW of power capacity throughout its current data center locations, with plans to add an additional 1 GW+ of capacity.
The Crypto-to-AI Transition
CoreWeave represents a prime example of the ongoing shift (or strategic hedge) from cryptocurrency to AI. While CoreWeave has achieved notable success, it’s not the only company making this transition.
Hut 8, for instance, has promoted its newly constructed 205 MW Vega data center campus as being ideally suited for AI training workloads. Furthermore, they recently acquired approximately 600 acres of land in Louisiana, with plans to establish 430 MW of AI data center capacity.
Similarly, Bitdeer, based in Singapore, launched an AI cloud platform in late 2023 and subsequently introduced an AI training platform in mid-2024. Last month, the company disclosed ongoing discussions regarding the construction of “large-scale sites in the U.S.” to support AI and other high-performance computing applications.
Meanwhile, Riot Platforms, a company focused on Bitcoin mining, recently appointed a new Chief Data Center Officer to lead the development of computing infrastructure for hyperscale and enterprise customers.
Chua and Gold emphasized that cryptocurrency companies offer more than just access to power. They also bring pre-existing facilities and operational experience in running data centers. While upgrades to networking, cooling infrastructure, and processors may be required, the availability of physical facilities is a significant advantage, considering that new data centers can take several years to construct.
These companies also possess experience in raising capital, which is a crucial skill in a market that demands significant investment.
While adjustments may be needed to meet the more demanding requirements of AI workloads, Chua noted that this expertise can be acquired through new hires.
Market Consolidation or Downturn?
Chua believes that these transitioning crypto firms will likely compete with traditional cloud computing providers, including hyperscalers, rather than colocation providers. However, demand currently exceeds supply in the near term.
Gold told Fierce that AI is currently a lucrative market. The critical question is how long this will remain true. Additionally, Chua suggested that increasing cryptocurrency prices (for example, Bitcoin reaching $120,000) and favorable regulatory changes could alter the calculus for those considering a shift. Some might choose to hedge their bets, finding profitability in both markets.
Predicting the future with certainty is challenging due to the rapid changes occurring in both the crypto and AI sectors. However, two potential outcomes of the current trends appear likely: consolidation or a market downturn.
Gold stated that consolidation typically occurs as a market matures, usually after three to five years (as seen in the cloud market). Therefore, he does not expect it to happen in the immediate future, but anticipates it over the coming years. Chua concurred with this assessment.
“Consolidation is inevitable,” Chua stated. Companies like CoreWeave, having already transitioned from mining to cloud computing, will likely seek to acquire assets to strengthen their market position. Expanding their scale will help them attract more investment, which can then be used to fund further consolidation.
However, both analysts also acknowledged the possibility of an eventual AI bubble burst.
One possibility is that the market could simply become saturated, with supply eventually surpassing demand. However, issues on the demand side could also arise.
Chua pointed out that the AI industry could potentially cannibalize its own customer base if the industry isn’t careful about economic disruption.
In essence, if AI leads to significant job displacement without providing sufficient alternative employment opportunities at comparable pay levels, consumers will reduce non-essential spending. This would impact not only direct AI products like ChatGPT or Gemini subscriptions, but also other AI-powered services across various industries. This would not just affect cryptocurrency companies transitioning to AI, but also the economy in general.
“As we develop this technology, we need to ensure that everyone benefits and grows together,” Chua concluded, “otherwise the demand side of the AI equation could suddenly disappear.”