Half of Planned U.S. Data Centers for 2026 Delayed or Canceled Amid Grid, Supply Chain, and Community Pushback
April 23, 2026 – A wave of delays and cancellations is reshaping the U.S. data center buildout for 2026, with industry analysts estimating that roughly half of planned facilities will not come online on schedule. The slowdown is raising concerns across the artificial intelligence sector, which depends heavily on rapid expansion of computing infrastructure.
According to industry tracking cited by analysts at Sightline Climate, data centers representing as much as 12 gigawatts of new power demand were announced for completion this year. However, only about one-third of those projects have actually broken ground, underscoring a widening gap between ambition and execution.
A system hitting multiple bottlenecks
The reasons behind the slowdown are stacking up. Developers are facing prolonged supply chain delays for critical electrical and cooling components, many of which are still heavily dependent on overseas manufacturing. Firms report sourcing equipment from countries including Canada, Mexico, South Korea, and China, adding complexity and time to already long construction cycles.
At the same time, the domestic electrical buildout has struggled to keep pace. Power transmission upgrades, substation expansions, and interconnection approvals are all emerging as major chokepoints in bringing new AI-focused facilities online.
Crusoe, a data center infrastructure company, noted that grid constraints can determine whether projects succeed or stall. “They can make or break a project,” a company executive told Bloomberg, pointing to long lead times for electrical infrastructure as a growing limiting factor.
Local resistance and resource pressure
Beyond industrial bottlenecks, many projects are encountering resistance at the local level. Rural communities that have become targets for large-scale data center campuses are increasingly pushing back over land use, water consumption, noise, and strain on local power grids.
Data centers require significant physical resources—not just land and electricity, but also large volumes of water for cooling systems. In some regions, these demands are colliding with existing agricultural and municipal needs, fueling opposition and regulatory scrutiny.
Strategic concerns over AI infrastructure
The slowdown comes at a sensitive moment for the AI industry, which is in the midst of an infrastructure race. Some analysts warn that if domestic capacity continues to lag, companies may face growing reliance on foreign supply chains for critical hardware, potentially slowing U.S. competitiveness in artificial intelligence development.
At the same time, policymakers and industry leaders have pushed for expanding domestic manufacturing capacity for grid equipment and semiconductor-adjacent infrastructure. Progress, however, has been uneven, with limited near-term relief for construction bottlenecks.
A turning point for AI expansion?
The divergence between projected demand and actual buildout is leading some in the industry to question whether AI expansion is hitting a structural ceiling—at least in the near term. While demand for compute continues to surge, physical infrastructure is proving far harder to scale.
For now, the sector faces a tension between rapid technological growth and the slower realities of land use, energy systems, and global supply chains. Whether that gap narrows or widens over the next year may shape the pace of AI development well beyond 2026.
