Currence report finds electricity demand, construction bottlenecks, public perception is slowing data center growth
A recent market outlook from energy intelligence firm Currence estimates that between 30% and 50% of the large-scale data center capacity expected to come online in 2026 will likely be delayed, due to mounting power constraints, construction bottlenecks and increasingly unreliable project timelines.
The report, which tracks 777 announced AI factories and large data centers exceeding 50 megawatts, shows an industry announcing projects far faster than it can build them. Since 2024, developers have announced approximately 190 gigawatts of future capacity, with approximately 16 GW scheduled to begin operations during 2026.
However, Currence found that only about 5 GW of that capacity is currently under construction, leaving roughly 11 GW remains in the announced stage with no visible construction progress, despite typical build timelines of 12–18 months.
“The announced pipeline suggests another year of explosive growth,” the report said, “but 30–50% of that pipeline is unlikely to come online before the end of the year.”
The findings highlight the growing reality that there are a lot of challenges involved in data center construction, from local opposition to finding skilled construction workers. Building them is proving a lot harder than planning them.
According to Currence, project schedules have already proven increasingly unreliable. During 2025, more than one quarter of expected capacity missed projected completion dates, while another 10% of projects quietly postponed their commercial operation dates. It’s not looking like things will get any better in 2026.
A primary obstacle is electricity. After a number of instances where local citizens saw their electric bill skyrocket after a data center opened up shop in their neighborhood, there has been tremendous pushback from cities and states on large scale data centers. In some instances, operators are being required to provide their own power rather than get power from the public grid, according to Currence.
Although projects powered entirely by on-site generation or hybrid systems account for fewer than 10% of announced facilities, they represent nearly half of the total announced capacity, according to the report.
Mindful of their public image, hyperscalers are responding quickly to these demands. Google has expanded its strategy by acquiring a large renewable energy development pipeline, while Amazon has increased direct investments in solar generation and battery storage.
Currence says the distinction between announced and executable projects is becoming increasingly important for utilities, equipment suppliers, infrastructure investors and developers attempting to forecast future demand.
The nature of accounting makes it difficult to determine how much data center debt hyperscalers have accrued, but there are some estimates out there. Bloomberg estimates that there’s over $500 billion in outstanding AI data center debt.
And Nikkei Asia reports this week that Meta, Google, Amazon, Microsoft and Oracle have accrued around $1.65 trillion in “hidden debt” in the last five years, with an additional hundreds of billions of dollars’ worth of off-balance sheet debt, on top of the roughly $1.35 trillion in debt reflected on their balance sheets. The data includes some estimates.
Meta’s off-balance sheet debt is particularly high at about $420 billion, nearly triple its recorded debt. But even though Meta has excess capacity and is looking to sell it as a cloud service provider, it still wants more. BlackRock is currently raising $12 billion to build a data center for Meta.
