AI data centers could require power equivalent to 60 new nuclear reactors in the U.S. by 2027.
Elon Musk has backed the notion of using the energy needed to sustain cryptocurrencies like BTC and data centers as the backing for the value locked up in these sectors, unlike the inconsequential process of central banks printing new fiat currency.
When X user @zerohedge quipped about the “hundreds of new nuclear power plants” the United States would need by 2028 to sustain the ballooning power demands of artificial intelligence in the early hours of Tuesday, October 14, it was meant as sardonic commentary on the AI investment boom.
But it quickly grew into a discussion about energy, monetary systems, and the limits of digital expansion, one that drew in Elon Musk himself.
“The money is not the problem: AI is the new global arms race, and capex will eventually be funded by governments (US and China),” Zerohedge wrote in a follow-up post. “If you want to know why gold, silver, and bitcoin are soaring, it’s the ‘debasement’ to fund the AI arms race. But you can’t print energy.”
Musk replied simply: “True. That is why Bitcoin is based on energy: you can issue fake fiat currency, and every government in history has done so, but it is impossible to fake energy.”
The AI power race
The International Energy Agency projects that global electricity consumption by data centers, AI, and cryptocurrency could more than double by 2030. In the United States alone, research by RAND suggests AI data centers could demand an additional 10 gigawatts of capacity by 2025 and up to 68 GW by 2027, roughly equivalent to the output of 60 large nuclear reactors
Tech giants are already seeking long-term solutions, especially those active in the AI race who understand the power needs of their enterprise.
Late last year, Google agreed to buy power from Kairos Power, a company developing small modular nuclear reactors (SMRs), to sustain its AI workloads. Microsoft and Meta have signed separate 20-year deals with Constellation Energy to utilize its nuclear power plants to get uninterrupted power.
A report by the Center for Strategic and International Studies warned that electricity supply is the binding constraint on U.S. AI scaling, with grid delays and permitting issues threatening to slow innovation. Experts project that AI data center power demands could grow exponentially to 123 gigawatts by 2035.
Given all these developments and the global AI-dominance race, Zerohedge’s question comes up again, “how many hundreds of new nuclear plants will the U.S. need?”
Musk’s energy-backed money thesis
Musk’s assertion that “Bitcoin is based on energy” taps into an idea long cherished among crypto proponents, which is that Bitcoin’s proof-of-work system and some cryptocurrencies, such as the Musk-supported Dogecoin, require significant electrical expenditure to mint new coins, tying issuance to a tangible, finite resource.
The logic contrasts with fiat currencies, which governments can expand at will through monetary policy.
In the context of surging government spending to support AI infrastructure and industrial policy, Musk’s comment suggests a parallel, and that is, energy may be the ultimate store of value in an era of fiscal excess.
Bitcoin’s energy consumption, while measurable, does not necessarily equate to intrinsic worth. Mining activity often clusters in regions with cheap or surplus electricity, meaning its supposed energy “backing” fluctuates with market incentives rather than physical scarcity.
However, miners are also now sourcing power from stranded natural gas, hydroelectric dams, and renewable grids.
Both the AI buildout and Bitcoin mining highlight the same foundational constraint, which is real energy.
Zerohedge’s observation that “you can’t print energy” is a reality that policymakers have to contend with as they balance digital expansion with physical infrastructure limits.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It’s free.
