Almost Half of US Data Centers That Were Supposed to Open This Year Slated to Be Canceled or Delayed

Why It Matters

These supply‑chain constraints could slow the rollout of AI‑driven services and erode the United States’ edge in high‑performance computing, affecting both investors and tech firms.

Key Takeaways

• 12 GW announced 2026; only ~4 GW under construction.
• 2027 plans: 21.5 GW announced, 6.3 GW building now.
• 2028‑2032: 37 GW planned, just 4.5 GW started.
• Imported electrical components cause major build delays.
• Supply‑chain bottlenecks threaten US AI compute expansion.

Pulse Analysis

The U.S. data‑center market has been riding a wave of AI‑driven demand, with developers announcing over 70 GW of new capacity through 2032. Yet the reality on the ground tells a different story: construction pipelines are thin, and many projects remain in a pre‑production limbo. Analysts from Sightline Climate highlight that only a fraction of announced gigawatts are truly under way, suggesting that the sector’s growth forecasts may be overly optimistic if current trends persist.

A critical choke point lies in the electrical component supply chain. Although batteries, transformers and circuit breakers represent less than 10% of a data‑center’s capital cost, they are indispensable for powering the massive compute loads AI workloads demand. With domestic production insufficient, firms are forced to source these parts from abroad—primarily Canada, Mexico, South Korea and China—introducing longer lead times, higher logistics costs, and exposure to geopolitical risks. The resulting delays ripple through project schedules, turning a single missing component into a stalled megawatt‑scale facility.

For investors and policymakers, the implications are clear: without a concerted effort to domesticize critical electrical components or streamline import pathways, the United States risks ceding its AI infrastructure advantage to regions with more resilient supply chains. Strategies may include incentivizing local manufacturing, creating strategic stockpiles, or revising tariff structures to reduce bottlenecks. Companies that can secure reliable component flows will likely capture a larger share of the burgeoning AI market, while laggards may see projects shelved and capital eroded.