AI expansion drives record US power demand and data center cooling shifts

AI data centers electricity demand liquid cooling power grid constraints capital expenditure

The rapid expansion of artificial intelligence is driving unprecedented demand for electricity and specialized infrastructure. In the United States, the Energy Information Administration (EIA) projects that power consumption will reach record highs in 2026 and 2027, fueled by AI data centers, cryptocurrency, and widespread electrification. Commercial power sales are expected to hit historic levels, while the share of renewable energy generation is projected to rise to 27% by 2027.

To manage the intense heat generated by high-density AI chips, the industry is shifting from traditional air cooling to liquid cooling technologies. Liquid cooling can remove heat up to 20 times more efficiently than air, which is critical as rack densities climb toward 100 kW and beyond. However, this transition requires significant capital investment; retrofitting existing facilities for liquid cooling can cost between $30,000 and $50,000 per rack when including necessary electrical and piping upgrades.

Major technology providers are facing a compute shortage. Amazon, Alphabet, and Meta have all signaled massive capital expenditures to expand capacity, with Amazon expecting to fall short of demand through 2027. Meanwhile, the financial sector is navigating the complexities of AI adoption. While many wealth management firms are investing heavily in AI, some report a weak correlation between spending and measurable business value, noting that many tools currently assist with tasks rather than fully integrating into workflows. In Europe, companies are also being warned about the rising costs of fragmented software spending, with potential savings of billions through better rationalization.

Entities

Alphabet · Amazon · Capital Power · CoreWeave Inc. · Dell Technologies Inc. · Ecolab · Financière Marjos · IDC · International Energy Agency · Meta Platforms · Microsoft · Nvidia

Claims

What the coverage asserts, and how many sources carry each claim.

[○ 1 SOURCE] A survey of leading wealth management firms found a weak correlation between AI spending and measurable business value. www.europesays.com
[○ 1 SOURCE] Water is capable of carrying approximately 3,500 times more heat than an equivalent volume of air. capacityglobal.com
[○ 1 SOURCE] Blackwell-class AI systems have increased rack power density to approximately 150 kW. www.datacenterknowledge.com
[● 2 SOURCES] The International Energy Agency projects global data center electricity use will reach approximately 950 TWh by 2030. www.automatedbuildings.com · podpora.postoj.sk
[○ 1 SOURCE] Financière Marjos plans to invest in AI data center operations across Europe, North America, and Asia Pacific. bmmagazine.co.uk
[○ 1 SOURCE] Corporate AI spending increased from roughly 1.7 percent of revenue in late 2025 to 3.3 percent. businessday.ng
[○ 1 SOURCE] Global data center electricity consumption is projected to rise from 397 TWh in 2024 to 915 TWh by 2028. biztechmagazine.com
[○ 1 SOURCE] Capital Power signed a long-term agreement to supply 250 MW of power to Meta Platforms. www.fool.ca
[● 2 SOURCES] US power consumption is projected to reach record highs in 2026 and 2027 due to AI data centers and electrification. boereport.com · www.cryptopolitan.com
[● 2 SOURCES] US power demand is projected to rise from 4,195 billion kWh in 2025 to 4,288 billion kWh in 2026 and 4,356 billion kWh in 2027. boereport.com · www.cryptopolitan.com
[● 2 SOURCES] Commercial power sales in the US are projected to reach a historic high of 1,549 billion kWh in 2026. boereport.com · www.cryptopolitan.com
[● 2 SOURCES] Data centers could account for between 9.5% and 15.3% of total US electricity use by 2030. www.automatedbuildings.com · www.cryptopolitan.com