← Back to stories Detailed view of a green circuit board featuring capacitors and microchips.
Photo by Pixabay on Pexels
SCMP 2026-05-22

China’s top chip foundries forecast second-quarter growth amid AI boom

Growth outlook

China’s biggest contract chipmakers, SMIC (中芯国际) and Hua Hong Semiconductor (华虹半导体), have signalled they expect sequential revenue gains in the second quarter as demand from artificial-intelligence projects lifts orders for processors and power-hungry accelerators. It has been reported that both firms cited stronger inbound demand from cloud providers and AI start-ups; analysts say customers are stocking up ahead of autumn product cycles. Short supply and sustained appetite for AI-capable silicon are giving even mature-node manufacturers a rare tailwind.

Geopolitical constraints and supply realities

But growth comes with limits. China’s foundries remain constrained by export controls that curtail access to the most advanced lithography tools, leaving them reliant on more mature process nodes that still dominate many AI hardware stacks. Reportedly, this means firms are winning business for inference chips and power-management devices rather than cutting-edge 3nm logic. So can domestic players translate demand into long-term competitiveness when key equipment and IP are restricted? The broader backdrop of US-China tech frictions and Beijing’s push for semiconductor self-reliance frames the answer.

Why it matters

For Western readers unfamiliar with China’s semiconductor landscape: foundries fabricate chips designed by others and are central to the global electronics supply chain. A bump in output from SMIC and Hua Hong reduces pressure on local supply but does not erase strategic vulnerabilities. Investors will watch whether short-term revenue gains lead to sustained investment in capacity and R&D, or whether restrictions and capital intensity keep China tied to legacy nodes for years. Either way, the AI boom is reshaping demand patterns—and the politics of supply chains—at the same time.

AISemiconductors
View original source →