Nearly 40% of US AI Data Centers Face Construction Delays, Satellite Data Reveals

Satellite Data Exposes Construction Delays in US AI Data Centers

Silicon Valley’s rapid expansion of AI data centers—projects requiring electricity equivalent to hundreds of thousands of US homes—faces mounting construction and power challenges, compounded by growing local opposition. New satellite imagery now reveals that nearly 40% of these projects may fail to meet their scheduled completion timelines in 2024.

Analysis Highlights Widespread Delays

The Financial Times analyzed satellite imagery from SynMax, a geospatial data analytics firm, to assess land clearing and foundation progress for data center projects. This data was cross-referenced with public statements and permit documents from IIR Energy, an industry research group. The findings indicate that major projects led by companies such as Microsoft, Oracle, and OpenAI are likely to miss completion deadlines by more than three months.

Key Challenges Behind the Delays

Interviews with over a dozen industry executives cited chronic shortages of labor, power, and equipment as primary causes of the delays. Construction leaders working on OpenAI projects specifically noted a lack of skilled tradespeople—including electricians and pipefitters—to support multiple concurrent data center developments.

Broader Implications for AI Infrastructure

The construction slowdowns threaten to exacerbate existing power grid constraints in regions hosting these facilities. Local communities are increasingly pushing back against the rapid expansion, citing concerns over energy consumption, environmental impact, and strain on public infrastructure. These delays could further delay AI advancements reliant on robust data center capacity.