Nvidia Crashed 9.3%. Is the AI Data-Center Boom Starting to Crack?

The November 5, 2026, sell-off was a broad repricing of AI-related investments, not evidence that chip shipments or company cash balances suddenly collapsed. Nvidia fell 9.26% to $181.60, AMD dropped 17.26%, and the Philadelphia Semiconductor Index lost about 12.25%; the Nasdaq slid 5.52%, wiping out an estimated $2.5 trillion in its overall market value. Losses also hit memory makers, AI cloud provider CoreWeave, and power-equipment company GE Vernova, reflecting shared exposure to data-center spending. The pressure followed U.S. midterm elections that raised concerns about scrutiny of facilities’ electricity use, water consumption, and permits, alongside questions about high valuations and whether costly infrastructure will generate returns quickly enough.

The central risk is that data centers may take longer to build, energize, and fill with paying workloads than investors expect. Delays in power access, permits, or financing could shift demand for GPUs, high-bandwidth memory, storage, and networking into later quarters without permanently eliminating it. A UBS estimate that about 30% of expected U.S. data-center capacity additions for 2028 could be at risk describes potential exposure, not confirmed cancellations. To judge whether the boom is weakening, investors should track operating capacity, AI revenue and cash generation relative to capital spending, and whether equipment orders convert into deployments. The sell-off also raises the bar for efficiency: suppliers must show useful computing performance, utilization, and revenue per dollar and watt. The conclusion is that AI demand may remain strong, while infrastructure economics - not headline buildout plans - determine which companies succeed.