Commodities March 25, 2026

AI Demand Spurs Rapid Expansion of China’s Chip Manufacturing Capacity

Surging AI infrastructure needs drive capital spending and strain supply chains even as foreign suppliers remain critical for high-end segments

By Ajmal Hussain
AI Demand Spurs Rapid Expansion of China’s Chip Manufacturing Capacity

China’s semiconductor sector is accelerating as global demand for AI infrastructure raises capital expenditure and pressures production capacity. Executives at Semicon China 2026 in Shanghai report faster-than-expected growth, rising requirements for testing and interconnects, and a near-term squeeze on raw materials and high-end components. Industry participants say China’s manufacturing base is relatively well positioned to scale, but foreign firms continue to play a vital role in specialized, higher-end parts of the supply chain.

Key Points

  • AI infrastructure demand is driving faster-than-expected growth and higher capital spending across China’s semiconductor industry, affecting manufacturers, equipment suppliers and materials providers.
  • China’s output share for chips produced on 22nm-40nm process nodes is projected to rise to 42% of global capacity by 2028 from 37% in 2026, reflecting significant capacity expansion in mature-node production used in cars, smartphones and electronics.
  • The AI-driven shift increases requirements for testing, packaging and optical interconnects; foreign suppliers remain important in higher-end segments and in after-sales and technical services.

China’s chip industry is experiencing an accelerated growth phase driven by a surge in global demand for AI infrastructure, industry executives said at Semicon China 2026 in Shanghai. The ramp-up in AI-related workloads has led manufacturers to increase capital spending and expand production capacity to meet rapidly rising orders.

On the show floor, several executives described the current growth as outpacing earlier expectations. As one industry executive put it, growth "is coming faster than expected" this year. That momentum is prompting companies across the value chain to add capacity and invest in equipment that supports higher-performance, more complex chips.


Capacity shifts in mature process nodes

Industry estimates presented at the event indicate a notable shift in global manufacturing distribution for mature process nodes. Production capacity for chips built on 22nm to 40nm process nodes - technologies commonly used in automobiles, smartphones and consumer electronics - is expected to rise to 42% of worldwide output by 2028, up from 37% in 2026, according to comments made at the conference.

This change reflects heavy capital investment aimed at meeting demand in segments that continue to rely on these mature nodes.


Rising complexity: testing, packaging and interconnects

The spread of AI workloads is also altering technical requirements across the semiconductor supply chain. As chips become more performance-intensive, requirements for testing, packaging and high-speed interconnects have grown.

"AI has significantly increased computing power requirements, and that in turn has raised the requirements for semiconductor testing," said Terry Feng, China sales director at U.S. chip testing firm Teradyne.

One area where the impact is particularly visible is optical interconnects - a layer that links chips within data centre systems. Firms that provide precision assembly equipment for optical modules reported heavy order books. "Our order backlog is already booked out into next year," said Zhou Limin of Mycronic’s MRSI unit, which supplies high-precision equipment used in assembling optical modules.


Supply chain strains and materials demand

The rapid uptake of AI-capable infrastructure has begun to stretch parts of the semiconductor supply chain, notably raw materials and high-end components, as manufacturers race to keep pace with demand. Executives at the event noted shortages and longer lead times in some segments.

Given the scale of its manufacturing base, several industry participants expressed confidence that China's domestic industry is relatively well placed to respond to the surge. "We are very optimistic about the memory (chip) cycle - there will be large-scale capacity expansion," said Bai Yu, vice president at Suzhou Origins Materials Technology. The company, which supplies materials to leading domestic memory manufacturers including ChangXin Memory Technologies, Yangtze Memory Technologies and Semiconductor Manufacturing International Corporation, plans to begin construction of a new production base next month.


Role of foreign suppliers

Despite the expansion of domestic capabilities, foreign firms continue to occupy a central role in higher-end segments of the chip supply chain. Conference participants noted that specialized expertise, specific materials and service capabilities keep foreign suppliers relevant even as local producers increase scale.

"There’s still a place for foreign firms because it is a very specialized industry - you have to have the expertise, the materials and the understanding," said Cameron Johnson, senior partner at Tidal Wave Solutions.

He added that foreign companies often maintain an advantage in after-sales support and technical services, areas where domestic competitors are still building capabilities.


Outlook

Executives at Semicon China 2026 painted a picture of brisk demand-driven growth across multiple segments of the semiconductor industry, from mature-node manufacturing to the specialized equipment and materials that support advanced packaging and optical interconnects. While capacity and materials constraints are creating short-term supply pressures, firms with strong manufacturing footprints and established supplier relationships are preparing to scale up operations to meet the AI-driven surge.

Risks

  • Supply chain strain in raw materials and high-end components as manufacturers struggle to meet surging AI-related demand - impacting materials suppliers, equipment makers and chip producers.
  • Order backlogs for precision assembly equipment, such as those for optical modules, could delay deployment timelines for data centre interconnects and related infrastructure - affecting system integrators and cloud providers.
  • Continued dependence on foreign firms for specialized expertise, materials and after-sales services could create vulnerabilities in the higher-end segments of the supply chain as domestic players scale up.

More from Commodities

Bernstein Warns of 1970s-Style Supercycle Forming in Middle East Oil Services Mar 25, 2026 BlackRock CEO Says Oil Could Reach $150 a Barrel if Iran Remains a Regional Threat Mar 25, 2026 U.S. Moves to Accelerate Return of Minesweeper-Capable Ships Undergoing Maintenance in Singapore Mar 25, 2026 India Accepts First Iranian LPG Shipment in Years After U.S. Eases Sanctions Mar 25, 2026 UBS: Modern Economy Far Less Sensitive to $100 Oil Than 1974 Shock Mar 25, 2026