The expansion of data centers in the United States is slowing down faster than expected. A new report from market analysts indicates that delays in critical hardware deliveries and rising costs are forcing companies to postpone or cancel projects. The strain on supply chains for semiconductors and power infrastructure is now visible, with some facilities facing months-long wait times for essential components.
Major tech firms had planned to increase their AI server capacity by 40% this year. Instead, several have revised projections downward after learning that chip shortages will persist into 2027. A spokesperson for NVIDIA confirmed that orders for high-end GPUs are being prioritized for existing clients, leaving new projects in limbo. Meanwhile, power grid constraints in Texas and Virginia have added another layer of uncertainty, with local utilities warning of potential blackouts during peak summer demand.
Software improvements may ease some pressure. New algorithms reduce the computing power needed for large language models by up to 25%. Meta and Google have already begun rolling out these updates, which could lower the hardware requirements for future AI deployments. Analysts at Dell Technologies estimate this shift might cut the need for new data center builds by 15% over the next two years.
Consumer tech prices are also under scrutiny. If demand for AI chips remains unmet, manufacturers of smartphones and laptops may face higher component costs. A senior executive at ASUS told Reuters that some product lines could see price increases of 8-12% by early 2027 unless supply improves. The company has already delayed the launch of two AI-focused laptops scheduled for this fall.
Source: itavisen.no