Bitcoin vs. AI: Who will win the race for electricity?

​Bitcoin and artificial intelligence have become two of the defining technological phenomena of recent years. The former offers an alternative financial system, while the latter promises to transform work, science and everyday life. But in an era of constrained power supply, which one will take priority?

Both technologies require enormous computing capacity. Yet while Bitcoin's energy consumption has faced years of criticism and restrictions, the construction of AI infrastructure has long been viewed as a necessary cost of technological progress.

According to the Cambridge Centre for Alternative Finance, the Bitcoin network consumes around 138 TWh of electricity annually, equivalent to roughly 0.5% of global consumption. At the same time, the expansion of AI has become one of the main drivers of rising data center power demand in recent years.

View post on X Loading...

The tweet was deleted by the author.

But we saved everything 🙂.

As a result, two fundamentally different technologies are finding themselves in the same race for energy. However, public and political attitudes toward their power needs remain markedly different.

So is the real issue how much electricity they consume? Or is the attitude toward Bitcoin and AI determined by the value people see in each technology?

Two competitors for the same power supply

From a technical perspective, it would be wrong to equate a cryptocurrency mining farm with a modern AI data center.

Miners primarily use specialized ASIC hardware, while AI infrastructure is built around GPUs, servers, data storage systems and high-speed networks.

At the physical level, however, the interests of the two industries overlap. Both mining and AI convert enormous amounts of electricity into computing power, even though they use different hardware and serve different purposes.

The International Energy Agency (IEA) expects global data center electricity consumption to more than double by 2030. According to the agency, demand will reach around 945 TWh, slightly more than Japan's total electricity consumption today.

Artificial intelligence will be the main driver of this growth. Under the IEA's base-case scenario, electricity consumption by AI-optimized servers is expected to increase by around 30% annually.

U.S. data center electricity consumption in 2014–2028. Source: LBNL.

In the U.S., data centers already accounted for around 4.4% of total electricity consumption in 2023. According to estimates from the U.S. Department of Energy, their share could rise to 6.7%–12% by 2028.

At the same time, Bitcoin's energy profile is changing. According to Cambridge research, sustainable energy sources now account for 52.4% of miners' electricity consumption, including 42.6% from renewables and 9.8% from nuclear power. Coal's share has fallen from 36.6% in 2022 to 8.9%.

Yet even as Bitcoin mining has become greener through greater use of alternative energy sources, it continues to compete with the AI sector for available electricity.

One megawatt, different value

The difference in attitudes toward the two technologies becomes clearer when considering not only how much energy they consume, but also what society receives in return.

The utility of artificial intelligence is relatively easy to demonstrate. It is used to write code, analyze documents, translate content, conduct research and automate work.

Bitcoin offers a fundamentally different product. Its core function is to provide a financial network that allows users to store and transfer an asset without a central issuer. The number of coins is limited by its protocol, while transactions do not depend on the decisions of any individual bank or government.

View post on X Loading...

The tweet was deleted by the author.

But we saved everything 🙂.

For someone with a stable currency and access to a developed banking system, the value of such an option may seem abstract. By contrast, the benefits of an AI system that helps complete a work task are almost immediately apparent.This contrast is particularly visible in government policy.

When sustainability depends on the technology

In recent years, cryptocurrency mining has repeatedly faced restrictions because of its high energy consumption. China imposed sweeping restrictions on the industry, some U.S. states considered moratoriums, and European regulators raised concerns about the environmental footprint of Proof-of-Work.

Laos provides a notable example. The country's authorities decided to stop supplying electricity to cryptocurrency miners and redirect capacity toward other industries, including AI data centers.

The government concluded that digital asset mining created too few jobs and generated little additional economic activity across related industries. The issue, therefore, was not simply how much electricity was consumed, but the economic return generated by each megawatt.

At the same time, mining facilities can reduce their electricity consumption relatively quickly during periods of high grid demand. When power prices rise sharply, shutting down equipment can be more profitable than continuing to mine, freeing up capacity during peak periods.

In practice, mining's energy profile depends on several factors at once: the source of electricity, the location of the equipment, pressure on the local grid and companies' ability to temporarily curtail operations.

AI loses its environmental free pass

For years, the expansion of artificial intelligence faced little comparable resistance. But as more data centers are built, that is beginning to change.

In the U.S., rising demand is forcing utilities to invest in power generation, transmission lines and substations. As a result, the question of who pays for grid expansion is increasingly becoming a political issue.

View post on X Loading...

The tweet was deleted by the author.

But we saved everything 🙂.

In Pennsylvania, Governor Josh Shapiro has pushed for large data centers to pay for the energy infrastructure required to support their operations. Similar debates are emerging in other states.

The issue also extends beyond electricity. Data centers require resources to cool their equipment, so construction in regions with limited water supplies is drawing opposition from local communities.

AI is gradually facing the same questions that were previously directed at the cryptocurrency industry: Is such resource consumption justified? Who captures the economic benefits? And are the costs being shifted onto other residents?

However, attitudes remain different, and the AI industry has yet to face resistance on the same scale as the protests surrounding Bitcoin.

Resources are being reallocated

Meanwhile, the market is already bringing the two industries together.

Hut 8 is a notable example. The company historically developed infrastructure for Bitcoin mining but is now using its experience with large-scale energy facilities to expand into AI data centers.

In May 2026, it signed a 15-year agreement worth $9.8 billion to lease 352 MW of IT capacity at its Beacon Point site in Texas.

View post on X Loading...

The tweet was deleted by the author.

But we saved everything 🙂.

Deals like this show that processors are not the only scarce resource in the new computing economy. Land, grid connections, transformers, cooling systems and guaranteed access to large amounts of electricity are becoming just as important.

That gives companies with experience building large energy-intensive facilities an advantage as they expand into AI data centers.

Electricity demand, meanwhile, continues to rise. In the U.S., the expansion of AI and cryptocurrency infrastructure has already become one factor driving power consumption, which the EIA expects to reach new records in 2026 and 2027.

What matters more — Bitcoin or AI?

There is probably no definitive answer. Comparing Bitcoin and AI reveals a different issue: the amount of electricity consumed says little by itself about whether society considers that consumption justified.

If cryptocurrency is viewed purely as a speculative asset, its energy use may appear excessive. If Bitcoin is instead viewed as a global, independent payment and savings network, the assessment looks very different.

The same applies to artificial intelligence. As long as the technology improves productivity, accelerates scientific discoveries and transforms entire industries, massive investment in energy infrastructure may appear justified.

But if a significant share of computing power is ultimately devoted to services with limited social value, the old environmental arguments will inevitably return in a new context.

Ultimately, the key question is not which technology consumes more electricity. What matters more is who decides — and by what criteria — which computations are worth the energy they consume.