Investments in AI data centers could reach $6.5 trillion by 2030, Bain report says

17 November 2026 Consulting.us 4 min. read

Global spending on data centers that power AI could reach between $5 trillion and $6.5 trillion by 2030, according to the 2026 Technology Report from consulting firm Bain & Company.

The findings suggest spending would add nearly 150 gigawatts of computing capacity, almost tripling the world’s total in just five years. Most of the growth would come from the United States, though capacity is expected to grow by double digits in every region.

The report warns that shortages of electricity, computer chips, skilled workers, and public approval could slow the buildout. Bain says the shortage of computing power is already severe and is unlikely to fix itself, because so many key ingredients are scarce at the same time.

Source: Bain Data Center Model

Five years ago, a facility with 50 megawatts of capacity was considered large. Technology companies are now planning campuses 100 times bigger, each costing between $150 billion and $200 billion.

The five largest cloud companies – Microsoft, Google, Amazon, Meta, and Oracle – could spend $780 billion on equipment and facilities in 2026, nearly five times what they spent three years earlier. The report found that figure could top $1 trillion by 2028.

Power, chips, and workers in short supply

Electricity is among the biggest obstacles. Connecting a new project to the power grid can take four years or more, and key equipment such as transformers is hard to find. Chips and memory are being reserved years in advance, and companies without long-term supply deals are effectively shut out. The industry also lacks enough electricians, mechanical tradespeople, and cooling specialists.

Sources: Bain Data Center Model, July 2026; Bain AI Chips Model, July 2026; TD Cowen; Bain analysis

Public resistance is growing as well. Bain reports that local opposition blocked or delayed at least 75 projects worth $130 billion in the first three months of this year, nearly matching the $156 billion in projects disrupted during all of 2025.

For example, earlier this year, New York became the first US state to pause construction of the largest data centers, and 14 other states have introduced similar legislation. Meanwhile, Ireland, the Netherlands, and Denmark have all limited or paused new grid connections. Data center expansion in France and Spain has been challenged by public opposition and lawsuits.

Workarounds will not be enough

Individual projects have found shortcuts, but Bain says they cannot be copied at scale. One example is the Colossus facility from xAI, which was built in under a year by repurposing a factory and leasing dozens of gas turbines without federal clean air permits, a move that drew a lawsuit. Even on sites that already have power, building a one-gigawatt data center typically takes between one and four years.

Source: Epoch AI

Going forward, closing the gap will require at least one of five major shifts laid out by Bain. Industry consolidation could leave only a handful of companies able to afford the costs. New sources of power, including natural gas, renewable energy, and eventually nuclear plants, could help, although large new nuclear capacity is not expected until the mid-2030s. Data centers in orbit could address land, cooling, and power limits, but it is currently seen as a long-term bet with unproven economics.

Government policy is also playing an increasingly central role. China, for example, is investing $295 billion in data center construction and $740 billion in grid upgrades. The final shift is earning public trust. Bain expects communities to demand binding commitments on water use, noise, emissions, and local hiring before approving new projects.

Who will pay for it all

There is an additional difficult question to address: Whether enough money can be earned to justify the spending. Bain estimates that annual AI infrastructure spending could reach $1.5 trillion by 2031, which would require an AI market of about $6 trillion a year to sustain. Consumer and business products could account for between $1.2 trillion and $1.8 trillion of that, leaving roughly $4.2 trillion that must come from entirely new sources of value.