•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
AI’s boom is accelerating worldwide, but infrastructure is lagging behind, creating delays in large-scale data-center projects in the United States and stretching the expansion plans of major tech companies. As demand for the “oil refinery” of the data economy grows faster than deployment, the AI ecosystem faces a growing bottleneck that slows the path from research and deployment to commercialization.
Data centers are increasingly unable to come online on schedule, limiting how fully companies can use the compute capacity they have already paid for. When facilities cannot operate as planned, it disrupts the chain from model development to product rollout, delaying the transition from investment to revenue and profit.
The issue is not confined to individual firms. Delays in U.S. data centers can create a domino effect across the AI supply chain, reducing the speed at which technology is deployed industry-wide.
The delays are particularly relevant for AI companies such as Meta and OpenAI, which have invested billions in model development and software infrastructure. However, limited data-center availability means they must wait longer to access enough compute resources, slowing the conversion of investment into revenue and profit.
The Financial Times reports that core construction trades for data centers—such as electricians and plumbers—are in very short supply. Building data centers requires high technical standards, stable electrical systems, and complex cooling infrastructure. With many projects underway in certain regions, competition for workers intensifies as laborers move between sites for higher wages, stretching schedules and raising costs.
Electricity pressure is another major factor. Some new U.S. data centers are said to consume electricity equivalent to the output of a nuclear power plant, creating challenges for local grids. While companies may invest in upgrading power infrastructure, implementation takes time.
Equipment shortages also contribute. Essential components such as gas turbines and transformers have long lead times, causing projects to stall even after initial preparation work is completed.
OpenAI is building a facility of about 1,200 acres (roughly 485 hectares) in the United States. Because land costs in urban areas are too high, companies often choose remote locations, which can increase labor costs by as much as 30% due to the need to attract talent to those areas.
Large-scale data-center investment requires substantial capital, and delays can make lenders more cautious. Some U.S. banks have withdrawn from financing commitments for Oracle-related projects as OpenAI has long-term commitments up to 1.4 trillion USD.
Wes Cummins, CEO of Applied Digital, said mobilizing capital at this scale is “very difficult,” reflecting growing financial pressure across the sector.
Geopolitical factors are also affecting progress. Investments in data centers in the Middle East have stalled due to tensions between the United States and Iran.
While some companies involved in data-center building and operation, including Nebius and Applied Digital, remain optimistic about meeting timelines, field reports indicate labor and material shortages are widespread. About 40% of sites are at risk of delays.
Overall, the delays do not necessarily stop projects, but they extend completion times. Investors who have put trillions into AI face concerns that profits will not be realized in the short term, as infrastructure constraints slow the monetization of compute resources.
In this environment, the biggest bottleneck in AI is no longer algorithms or processing hardware, but physical infrastructure. With constraints in electricity, equipment, land, labor, and financing, AI growth will continue to depend on solving infrastructure challenges in the period ahead.
Some firms have considered on-site generation such as gas turbines, but environmental permitting hurdles and supply-chain constraints with long delivery times could push timelines to 2028–2030.
Source: FT
Premium gym chains are entering a “golden era” that is ending or already in decline, as rising operating costs collide with shifting consumer preferences toward more flexible, community-based ways to exercise. Long-term memberships are shrinking, margins are pressured by higher rents and facility expenses, and competition from smaller, more personalized…
