Estimates place the Bitcoin network’s annual electricity use at roughly 138–175 terawatt‑hours (TWh). Following the most recent halving, about 164,000 bitcoins are mined each year; at current prices that would correspond to roughly 500 barrels of oil equivalent per bitcoin by one calculation, and — under higher estimates — it could exceed 600 barrels. While substantial, these figures are significantly lower than the energy consumed by data centers and AI infrastructure.
IEA: 415 TWh in 2024, projected 945 TWh by 2030
According to the International Energy Agency (IEA), global data centers drew about 415 TWh of electricity from grids in 2024. Converted to crude‑oil equivalents, this equals around 670,000 barrels of oil per day used solely to power servers. The IEA expects consumption to rise to more than 945 TWh by 2030, which would be roughly 1.5 million barrels of oil equivalent per day.
On an annual basis, the energy used by AI‑centric facilities corresponds to approximately 240–250 million barrels of crude oil. Analysts note this quantity is comparable to the daily output of a medium‑sized oil‑producing country and several times Hungary’s daily crude‑oil consumption of roughly 170–180 thousand barrels.
Grids and utilities are struggling to keep up
Energy suppliers and grid operators are having difficulty scaling capacity at the pace of data‑center and AI investments. The mismatch is particularly acute in the United States, the current leader in the data‑center industry: wait times for grid connections on important local markets already exceed an average of four years in some cases. Market participants report growing transformer shortages, constrained transmission capacity, and difficulties for utilities in serving AI campuses that demand hundreds of megawatts at once.
Investment needs and energy constraints
Jones Lang LaSalle’s 2026 outlook for the AI sector estimates global data‑center capacity could nearly double by 2030, requiring roughly 100 gigawatts of new power and up to $3 trillion in infrastructure and GPU investments. McKinsey projects global spending on AI infrastructure could reach $7 trillion by 2030, with more than $5 trillion directly tied to AI workloads.
However, this investment wave risks running into energy constraints: deployment is not only a financial question but increasingly an issue of access to reliable power, as many regions lack rapidly available grid capacity and connections.
Energy becomes a strategic competitive factor
Technology firms are increasingly competing for reliable, low‑cost, and clean power — a dynamic that resembles past resource races in the oil industry. Rystad Energy’s analysis suggests that digitalization and AI could generate nearly $500 billion of incremental value for oil and gas exploration and production companies between 2026 and 2030. At the same time, the large U.S. tech companies are estimated to be planning around $660 billion of AI investments this year, intensifying competition for energy.
Some firms are pursuing greater energy self‑sufficiency. For example, Microsoft and Meta have sought operating arrangements to secure nuclear power for future facilities, aiming to reduce their indirect dependence on oil and gas.
Who benefits?
Analysts conclude that companies with access to cheap, clean, and scalable energy capacity are likely to gain an edge in the coming technology‑driven growth cycle. As electricity becomes as strategically important as advanced semiconductors, control or reliable access to energy infrastructure is becoming one of the most valuable assets in the digital economy.
In short: the rapid expansion of AI and data centers is creating large and growing power demands that will stress grids and utilities. Upcoming investment decisions will be shaped not only by finance and compute, but also by the availability and cost of electricity — turning energy access into a central competitive factor for the tech sector.
