Global data center capacity demand is projected to reach 163 gigawatts (GW) by 2030, roughly double today’s level, according to Bain & Company’s latest baseline forecast. The firm’s outlook—anchored in continued strength in AI-driven demand and gradual easing of supply chain and power constraints—suggests that while growth remains robust, discipline and innovation in energy sourcing will define the next phase of the buildout. 

Power Demand Soars with AI as the Catalyst 

By 2030, U.S. data center electricity consumption could double to 409 terawatt-hours (TWh), with AI workloads accounting for most of the increase, Bain projects. The firm estimates that U.S. data centers could consume about 9% of the nation’s total electricity—more than double their current share and roughly 150 TWh higher than the U.S. Energy Information Administration’s baseline outlook. 

“We expect there will be sufficient energy supply to meet demand,” said Aaron Denman, leader of Bain’s Americas Utilities and Renewables practice. “However, power access is now the critical gatekeeper of growth. Even as GPU and construction constraints ease, more flexible and independent sources of power will be needed. As such, behind-the-meter (BTM) power generation has become the go-to source, shifting timelines and decision-making.” 

Behind-the-Meter Power and Flexible Infrastructure 

Meeting this unprecedented electricity demand will require tight coordination among utilities, regulators, and developers, as well as a mix of short- and long-term solutions, Bain noted. In the near term, flexible demand programs, battery storage, and BTM generation—such as natural gas turbines, rooftop solar arrays, and small nuclear restarts—will be essential to offset grid strain and smooth load volatility. 

Longer-term relief will depend on grid modernization, renewable energy integration, and transmission expansion, which are critical to sustaining AI-era demand growth. Bain notes that BTM generation can also support smaller, distributed data center clusters suited for AI inference workloads, which are expected to dominate compute demand later this decade. In contrast, mega-scale facilities exceeding one gigawatt of capacity will become standard for frontier AI model training. 

Strategic Investment and Regional Shifts 

Despite predictions that hyperscalers would pull back on expansion in 2025, Bain finds that major cloud and AI players have instead adopted a more deliberate, capital-efficient approach. 

“The general prediction that hyperscalers would scale back investments didn’t happen in 2025,” said Padraic Brick, co-leader of Bain’s data center perspectives. “However, we are seeing more deliberate investments as they scale capacity—focusing on capital efficiency and becoming more selective on locations for new deployments, particularly those tied to AI workloads.” 

By 2030, North America is expected to retain roughly half of global data center capacity, fueled by hyperscaler capital expenditures. Meanwhile, sovereign AI mandates and enterprise digital adoption are accelerating investment across Europe and Asia-Pacific, as companies seek geographic diversification to balance latency, data sovereignty, and energy sourcing priorities. 

Construction Delays Add to Industry Strain 

Beyond power, the physical construction of data centers has emerged as a growing bottleneck. Bain highlights escalating execution risks, including permitting delays, equipment lead times of eight to 24 months, and electric utility interconnection lags of up to five years. Labor shortages, particularly in skilled electrical and mechanical trades, further compound delivery timelines. 

According to Bain, four proven actions can reduce construction timelines by up to a year: Identify and pre-secure the right markets and site portfolios; Adopt modular design and prefabricated equipment to accelerate deployment; Leverage cross-functional design and supply chain expertise to optimize workflows; and collaborate closely with suppliers and pre-purchase critical equipment in bulk to mitigate supply risk. 

The New Data Center Mandate: Scale with Precision 

“The AI data center race is no longer just about scale,” said Peter Hanbury, leader of Bain’s global operations work for technology clients. “Winners are taking deliberate and careful approaches to capacity investments, while at the same time, actively securing fit-for-purpose power generation and mitigating build delays.” 

As the AI revolution continues to reshape digital infrastructure, Bain’s outlook suggests the industry’s next growth phase will hinge less on speed and more on execution discipline—where success will depend on the ability to secure reliable, sustainable power and deliver capacity efficiently in an increasingly constrained ecosystem.