This article was compiled and organized by BlockWeeks
The explosion of artificial intelligence is pushing data centers into an unprecedented battle for power and capital. Hyperscale cloud providers are continuously ramping up investment to add new computing power, but traditional data centers, constrained by limited power supply capacity and 2-4 year construction cycles, are already struggling to keep pace. Bitcoin mining companies that hold large-scale land, cooling water sources, dark fiber, reliable power, power approval permits, and long-lead-time critical equipment are in a position to multiply the value of their assets.
How big is the AI data center pie really
The widespread adoption of generative AI in 2024 ignited the entire industry chain. Pitchbook data shows that since 2016, cumulative funding into AI and machine learning startups has exceeded $680 billion, involving more than 100,000 deals, with $120 billion invested in 2024 alone.
The power consumption intensity of AI and high-performance computing (HPC) far exceeds that of traditional internet businesses. According to the International Energy Agency, a single ChatGPT query consumes about 2.9 watt-hours of electricity, while a single Google search consumes only about 0.3 watt-hours. This energy-intensive business model directly drives up demand for data centers.
Goldman Sachs Research estimates that U.S. data center demand will reach 21GW in 2024, a year-over-year increase of 31%; from 2022 to 2033, U.S. data center demand growth is expected to be a 15.8% compound annual growth rate. Based on the strong year-over-year increase in 2024, the institution predicts that U.S. data center demand will rise to 45GW by 2030, at which point it will consume up to 8% of total U.S. power capacity.
Behind the demand is massive capital expenditure by hyperscalers. Hyperscale data center companies like Google Cloud and AWS are positioning themselves by committing to invest more than $100 billion over the next 10 years to build AI-oriented data centers. JP Morgan Asset Management estimates that by the end of 2024, $163 billion will be invested in hyperscaler business expansion, a year-over-year increase of 28%. By 2038, hyperscaler AI capital expenditure is expected to reach $370 billion, an increase of 127% compared with the estimated level in 2024.
The continued expansion of AI and HPC is reshaping the form of data centers: from traditional general-purpose computing facilities to advanced AI infrastructure hubs, providing the foundation for autonomous driving, cutting-edge medical research, and next-generation AI applications. The future of digital innovation depends to a large extent on the continued evolution and expansion of these facilities.
Why traditional data centers cannot handle this wave of demand
The most direct bottleneck is power. The surge in interconnection applications for 300MW or even 1000MW+ scale has already overwhelmed local power grids at such a rapid pace, and interconnection and construction cycles are generally extended to 2-4 years.
The second is power density. Traditional data centers do not have large-capacity power supply capabilities and cannot support high-density computing. In the past, the power limit per rack was about 40kW, but today cutting-edge systems like GB200 NVL72 require a single rack to support more than 132kW. This means traditional facilities are already out of the game at the physical level.
As a result, hyperscalers that want to expand capacity on demand and on schedule have very limited options.
Converting mining farms into AI data centers: where are the barriers
Not all mining farms can be converted into AI data centers. AI/HPC workloads have specific requirements for cooling, networking, and redundancy, and only mining companies with suitable assets and engineering capabilities can truly capture this wave of benefits. The core conversion difficulties are concentrated in four areas:
1. Cooling systems. Heat dissipation in mining farms mainly revolves around the machines themselves, with less attention paid to supporting infrastructure; AI data centers require more advanced cooling solutions, such as direct-to-chip liquid cooling for the latest generation of high-power-density NVIDIA servers, combined with air cooling systems to support networking and mechanical infrastructure.
2. Redundancy levels. The redundancy requirements for AI data centers are far more stringent than for mining. Mining loads themselves are elastic and do not require strong backup power; AI data centers typically require at least N+1 redundancy throughout the entire process, and critical components such as core networking and storage also require higher levels of redundancy to ensure uninterrupted operation, or at least to complete reasonable caching and checkpoint saving when equipment fails. This means that every piece of critical infrastructure (such as cooling equipment) must have a backup: when maintaining one cooling unit, another must be online to maintain continuous operation. This level of redundancy is extremely rare in mining farms.
3. Form factor reconstruction. AI data centers use rack-mounted servers, which are vastly different from the "shoebox" form factor of ASICs used in Bitcoin mining. To accommodate AI hardware, the physical infrastructure inside the facility must be completely redesigned to support rack-mounted systems and their specific cooling, networking, and electrical requirements.
Taken together, converting mining farms into AI/HPC data centers is essentially an engineering and design challenge, and its enhanced infrastructure requirements also make the capital expenditure of AI/HPC data centers significantly higher than that of Bitcoin mining facilities.
Only a few mining companies can get a ticket to entry
Traditional data centers cannot meet the power requirements of modern AI workloads by retrofitting existing facilities. This market vacuum has precisely created an opportunity for Bitcoin mining companies—they hold exactly what AI/HPC companies urgently need: large-scale sites with accelerated energization timelines. Hyperscalers that want to expand quickly in the short term to keep up with the explosive demand of AI/HPC businesses have very limited choices, making Bitcoin mining companies a logical and viable option.
But this is a highly selective generational opportunity. Only a small number of mining companies possess the scarce infrastructure and capabilities necessary to support modern AI/HPC workloads. Mining companies that truly hold these scarce assets and want to maximize their value will turn to AI/HPC data centers.
The key to value release lies in valuation arbitrage: mining companies currently trade at about 6-12x EV/EBITDA, while leading data center operators generally trade at 20-25x. With the predictable cash flows of AI/HPC businesses, active financing markets, and significant valuation upside, this transformation is extremely attractive for mining companies with the right assets and has a clear accretive effect.
Chain effects on mining itself
Some critics argue that Bitcoin mining companies diverting into AI/HPC businesses will reduce the computing power used for block generation, thereby weakening network security. But this shift may instead benefit the entire mining ecosystem: mining companies that cannot meet AI/HPC site requirements may actually improve profitability due to better hashprice. As more miners go offline and Bitcoin prices rise, the upward movement in hashprice will significantly improve the profit margins of all Bitcoin mining companies. Bitcoin has risen by as much as 143% during the year, and with a Bitcoin-supportive president entering the White House, U.S. Bitcoin mining is entering its strongest era.
The clearest implementation path in crypto × AI
The combination of crypto and AI can be said to be one of the hottest crypto tracks in 2024. As of December 2024, the total market capitalization of crypto projects with circulating tokens that are building AI projects is about $33 billion; Galaxy Research estimates that more than $382 million in venture capital flowed to early-stage crypto AI startups in 2024.
Although most crypto AI projects have not yet found product-market fit, the intersection of Bitcoin mining and AI/HPC business growth is clearly visible. The reason Bitcoin mining's entry into AI stands out among many "crypto × AI" narratives is that it has the potential to provide at scale the most core element for AI/HPC companies—energy. Therefore, Bitcoin mining companies that hold assets convertible to AI/HPC may be one of the very few crypto × AI investment targets in the current industry that combines both purity and scale potential.
