How artificial intelligence is changing the Bitcoin mining industry

The rapid development of artificial intelligence (AI) has given the Bitcoin mining industry a direct competitor when it comes to electricity consumption. Both technologies require enormous computing power, which leads to rapidly increasing electricity consumption.

According to the International Energy Agency, AI data centers currently consume 1 to 1.3 percent of the world's electricity - around three times the power consumption of mining facilities. Some analysts expect this discrepancy to grow, with Bitcoin mining companies converting 20 percent of their electricity capacity to AI by the end of 2027.
The Electric Power Research Institute estimates that AI data centers in the US will double their electricity consumption by 2030, using up to 9 percent of all US electricity.

This increasing energy demand not only puts pressure on power grids, leading to higher costs and potential environmental impacts, it also presents a new challenge for Bitcoin mining companies to maintain their core business in certain regions.

According toreports, the demand for electricity in the US for AI data centers is so high that the most well-funded technology companies such as Amazon, Microsoft and Google are constantly on the lookout for new locations and energy sources. Money is no object, which also has a direct impact on Bitcoin mining companies competing for the same energy facilities and contracts. In contrast to Bitcoin mining companies, the price of electricity is secondary for AI data center operators, so they offer more financial incentives to energy providers.

Talen Energy, for example, has decided to neglect its mining activities and to award Amazon the contract for a nuclear-powered data center in Pennsylvania instead of Marathon Digital. Although Marathon is the world's largest listed Bitcoin miner, it has no chance against the financial resources of Amazon, which has a market capitalization 350 times greater than Marathon Digital.

Similarly, bitcoin mining services company EZ Blockchain had to forgo a 10MW project with a utility because the utility ultimately opted for a 100MW contract with a global, well-funded AI company that EZ Blockchain could not compete with.

While some mining companies are not gaining or even losing access to power due to better supply from AI companies, leasing or selling existing infrastructure to tech companies represents a new attractive revenue stream for the mining industry.

In June, Core Scientific became the first bitcoin mining company to announce an agreement to lease 200 megawatts (MW) of its own high-performance computing facilities from Nvidia-backed cloud provider CoreWeave. The contracts run for twelve years and have an estimated value of 6.7 billion US dollars.

Other large mining companies can also imagine reducing their mining activities and marketing their sites to AI companies or helping to develop more AI data centers. According to Kerry Langlais, the CSO of TeraWulf, Amazon and Google have already expressed interest in a location in Upstate New York.

According to a study by Morgan Stanley, the conversion of mining facilities to AI data centers could quintuple the value of the facilities and greatly reduce the waiting times that are normally required to connect a new facility to the power supply in the USA.

Buying or leasing space from a miner with at least 100 MW of capacity can cut the wait time to get a data center up and running by about 3.5 years and save tech companies billions.
_{Morgan Stanley in an interview}

The acquisition of mining facilities is therefore an attractive alternative for well-funded tech companies in the USA due to the waiting times. However, not every mining plant is suitable for conversion to AI.

However, converting from a mining facility to an AI data center comes with some challenges. The two types of data centers have several differences that could cause problems for some mining companies during the conversion.

According to CleanSpark CEO Zach Bradford, an AI facility takes about three years to build, while a mining facility only takes six to twelve months. When converting a mining plant, other infrastructure and special cooling systems are required, which drive up costs and therefore cannot simply be financed by many mining companies. In addition, facilities far away from urban centers are not really suitable for AI calculations as they rely on low latencies.

For electricity grid operators, who primarily use intermittent renewable energy sources, Bitcoin mining facilities are more attractive due to their flexible nature. Unlike AI data centers, they are suitable for demand response programs as they can be ramped up and down quickly, stabilizing power grids and prices. AI calculations are not as flexible as Bitcoin mining and require additional power generation for peak demand, whereas mining facilities often only utilize the unused surplus energy. This flexibility ultimately also leads to lower emissions. Contrary to the claims of the International Monetary Fund (IMF), Bitcoin mining facilities are therefore better for the environment than AI data centers - Blocktrainer.de reported in detail.

Artificial intelligence is influencing the Bitcoin mining industry through the financial backing of large technology companies. The high profit margins of AI and the long waiting times for connection to the power supply mean that the infrastructure of mining facilities is increasingly being used or converted for AI calculations. This could also have an impact on the hashrate of the Bitcoin network in the future.

Nevertheless, mining facilities have numerous advantages over AI data centers. They are more flexible, environmentally friendly and universally applicable, while AI data centers are location-dependent, still largely inflexible and have a greater impact on the environment.

When it comes to utilizing unused energy in the most remote locations and reducing emissions, Bitcoin mining facilities remain the better choice. This applies not only to large mining companies, but also to small projects scattered around the world.