AI-Driven Price Surge Spreads Across Semiconductor Supply Chain, Creating Bottlenecks

The AI-driven price surge in the semiconductor supply chain is spreading beyond graphics processing units (GPUs) and memory chips to upstream materials and manufacturing inputs, creating new bottlenecks that could slow the buildout of global artificial intelligence infrastructure.

Suppliers of once less-visible parts are gaining leverage as customers compete for limited capacity.

These range from power chips and capacitors that regulate electricity inside AI data centres, to copper-clad laminates and glass fabric that form the base of printed circuit boards (PCBs), as well as industrial gases, valves and ceramic parts used in chipmaking tools.

“Capacitors and power semiconductors are now both moving into a price-increase cycle,” said Liu Gaochang, an analyst at Sinolink Securities, who noted that the adjustments have spread from selected products to wider categories.

AI servers use three to 10 times as many capacitors as traditional servers, and orders for power components were “fully loaded”, Liu said, adding that costs had also risen for aluminium foil, chemical materials and electricity.

Japan’s Murata Manufacturing, the world’s largest maker of multilayer ceramic capacitors, or MLCCs, will raise prices for products used in AI servers and high-end automotive electronics by 10 to 40 per cent from July, Shanghai Securities News reported.