SMIC, China’s largest contract chip manufacturer, is seeing an unexpected opportunity amid the global AI craze. The company states that more foreign clients are shifting orders to China because fabrication plants in other markets are increasingly busy with higher-margin products: AI accelerators, memory, advanced packaging, and components related to high-performance servers.
This development should not be mistaken for a direct technological victory over TSMC or Samsung at the most advanced nodes. SMIC remains constrained by U.S. export restrictions, which hinder access to key equipment necessary to compete openly in leading-edge processes. What is happening is different: while major foundries prioritize AI and cutting-edge chips, part of the demand for mature nodes is seeking available capacity elsewhere. According to SMIC, China is one of the few markets where manufacturing capacity still exists.
Zhao Haijun, co-CEO of SMIC, noted that many products formerly manufactured outside China can no longer find space in those fabs, prompting some clients to return orders to the country. Reuters reports that this trend is favoring SMIC, which benefits from the “overflow” of demand in an industry where capacity is once again a strategic asset.
AI tightens advanced nodes and boosts the value of mature ones
In recent years, much of the semiconductor conversation has focused on 3 nm, 2 nm, and upcoming lithography generations. However, the true economics of chips are broader. Cars, appliances, sensors, medical equipment, robotics, industrial devices, wearables, IoT, power controllers, microcontrollers, and analog circuits still rely on mature nodes.
These chips rarely make headlines but are indispensable. An electric vehicle may include advanced computing or driver assistance components, but it also requires a large quantity of power chips, sensors, controllers, and energy management circuits manufactured using less advanced processes. The same applies to data centers for AI: alongside GPUs or ASICs, there are PMICs, controllers, network components, storage, and auxiliary electronics.
TrendForce points precisely to this reorganization. The average utilization of 8-inch wafer capacity among the top ten foundries worldwide is projected to approach 90% in 2026, up from around 80% in 2025, and stay above 80% through the first half of 2027. The firm also notes that multiple fabs are reallocating mature 8- and 12-inch capacity toward low-power processes, AI servers, and edge AI, which could lead to orders shifting toward Chinese suppliers.
This shift explains why SMIC can improve its position without surpassing TSMC in the race for finer nodes. The company doesn’t need to win in 2 nm to secure orders for products manufactured at 28 nm, 40 nm, 55 nm, or similar processes. Its current advantage lies elsewhere: available capacity, proximity to the Chinese industrial chain, and a customer base that needs to ensure production.
More clients, more capacity, and increased margin pressure
SMIC has been preparing for this scenario. In February, the company announced it would add approximately 40,000 equivalent 12-inch wafers per month in 2026, following an addition of 50,000 wafers per month in 2025. This expansion aims to meet stronger demand but comes with financial costs: SMIC expects its depreciation expenses to increase by around 30% this year, which could pressure margins even as revenues grow.
The latest financial snapshot shows this tension. In Q1 2026, SMIC reported a profit of $197.4 million, up 5% year-over-year, though below analyst forecasts. Revenue increased by 11.5%, reaching $2.5 billion. The company also expressed a more optimistic outlook for the year due to stable orders and demand from clients.
Growth, therefore, is not without friction. Installing new lines, adding equipment, improving manufacturing performance, and increasing utilization require investments. If capacity comes online before being filled with profitable orders, costs weigh heavily. Reuters also reports that SMIC added 9,000 equivalent 12-inch wafers in Q1 2026 and that utilization slightly dipped to 93%, affected by lower demand for mid- and low-end smartphones and the startup of new factories.
| Indicator | Key Data |
|---|---|
| SMIC revenue in Q1 2026 | $2.5 billion |
| Net profit in Q1 2026 | $197.4 million |
| Revenue YoY growth | 11.5% |
| Capacity utilization | 93% |
| Wafers shipped in Q1 | 2.5 million 8-inch equivalents |
| China’s share of revenue | 89% |
| U.S. share of revenue | 9% |
| Projected depreciation increase | Around 30% |
China consolidates its dominance in legacy nodes
The big question is whether this situation will be temporary or structural. The answer seems to lie somewhere in between. The current capacity saturation is heavily driven by investment in AI, but China’s expansion in mature nodes has been ongoing for years. According to Semiconductor China data cited by Reuters, China’s share of global capacity for legacy nodes from 22 nm to 40 nm is projected to rise from 32% in 2025 to 37% this year and 41% in 2027.
This progress clearly has an industrial message. While parts of the world focus on investing in advanced nodes and high-performance packaging, China is building a broad base in mature processes. Though not the most glamorous segment, it is one of the most widespread. Moreover, it may be more resilient from a demand perspective, as it supports diverse sectors.
This movement does carry risks. Accelerated expansion of mature capacity could lead to oversupply if demand softens or if clients revert to traditional foundries once pressure on TSMC, Samsung, or UMC stabilizes. Additionally, some international buyers still evaluate geopolitical risks, export controls, compliance issues, and China’s role in their supply chains.
There is also a commercial tension. Chinese foundries have been competitive on price, but profitability depends on maintaining high occupancy, process improvements, and avoiding a price war. If too many Chinese players compete for the same mature orders, volume growth might not translate into proportional profits.
Nevertheless, SMIC has reasons to be confident. AI is drawing resources from the industry’s most advanced layers, creating gaps in products that seemed less attractive. These gaps represent business opportunities—not necessarily in making the world’s most cutting-edge accelerators but in producing millions of chips that power cars, factories, routers, appliances, medical devices, and servers.
SMIC’s case reminds us that technological sovereignty isn’t only about nanometer nodes. It’s also about having the capacity to manufacture enough chips, reliably and affordably, for daily economy needs. China hasn’t won the 2 nm race, but it is strengthening a position that could give it significant influence in the less visible, more widespread parts of the market.
Frequently Asked Questions
Why is SMIC receiving more foreign orders?
Because AI demand is saturating capacity in foreign fabs, especially for advanced and higher-margin products. Some mature-node chips are seeking available capacity in China.
Does SMIC already compete head-to-head with TSMC on 3 nm or 2 nm?
No. SMIC remains limited by export restrictions and isn’t on the same footing as TSMC or Samsung at leading nodes. Its current opportunity mainly lies in mature processes.
What are mature nodes?
They are less advanced manufacturing processes, yet widely used in analog chips, controllers, sensors, PMICs, microcontrollers, automotive, IoT, industrial, and consumer electronics.
Why are mature nodes important for China?
They enable building a broad industrial base, supporting essential sectors, and reducing dependence on external sources for components that don’t require the most advanced processes but require high-volume manufacturing.