The global contract chip manufacturing industry is experiencing a new leap forward, driven by a powerful combination: the demand for Artificial Intelligence, the rapid cycle of consumer electronics, and the rise of “specialized” semiconductors (power, automotive, sensors) that don’t require the most advanced nodes but do need high capacity and healthy margins.
In this context, the map of pure-play foundries (manufacturers that produce chips for third-party clients) is more concentrated than ever. According to data cited by DIGITIMES based on ChipInsights, the combined revenue of 29 pure-play foundries reached 1.149 trillion yuan (CNY) in 2025—a 25.46% year-on-year increase—surpassing the one trillion yuan mark for the first time in this metric.
An industry that’s growing, but with an increasingly dominant leader
The fact that the market is growing isn’t surprising. What’s notable is how it grows: the investment cycle is driven by AI computing (training and inference), which demands advanced chips and high-level packaging; at the same time, mature nodes continue to be essential for screens, controllers, automotive, energy management, and sensors.
Practically, this dynamic tends to favor those with scale, advanced technology, streamlined supply chains, and a customer base capable of reserving production in advance. In this landscape, TSMC remains the benchmark: by 2025, it had already made significant differences in market share and quarterly revenue, with peaks of dominance reflecting the boost from AI and cutting-edge nodes.
Nevertheless, growth isn’t evenly distributed. Counterpoint, for example, already predicted in 2025 that industry pure-play foundries would surpass 165 billion USD that year—representing a 17% year-over-year increase—and indicating the scale of this cycle.
The political-industrial data: Four Chinese foundries now in the Top 10
The other major headline is geopolitical: China has managed to place four players in the Top 10 of the pure-play foundry ranking. An article from the Shanghai Stock Exchange (SSE)—also echoed via other channels—mentions that SMIC, Hua Hong, Nexchip, and United Nova Technology (UNT) are ranked 2nd, 5th, 9th, and 10th, respectively, among the “exclusive wafer foundry companies” (pure-play foundries).
This is no minor detail: it confirms that China’s growth in semiconductors is no longer limited to investment announcements or future capacity, but is reflected in significant positions within the global classification.
Summary table: the four Chinese foundries in the Top 10
| Company | Position (SSE ranking) | Profile and industry focus |
|---|---|---|
| SMIC | 2nd | Main Chinese foundry; strong focus on mature nodes and expansion of domestic capacity. |
| Hua Hong | 5th | Specialized in 8-inch and 12-inch manufacturing, with power/discrete semiconductors as key areas. |
| Nexchip | 9th | Closely tied to display driver ICs (DDI), diversifying into automotive and power devices; production mainly in 12-inch mature nodes (90–150 nm). |
| United Nova Technology (UNT) | 10th | Foundry and packaging/modules (including SiC and MEMS); originated as a joint venture with support from SMIC and local government. |
Why it matters to buyers (and anyone dependent on chips)
To the end customer, this “league” of foundries might seem distant. But the effects are tangible:
- Availability and lead times: When capacity is tight, lead times lengthen and planning becomes more rigid. This impacts electronics manufacturers, automotive, and industrial sectors.
- Pricing and margins: During high-demand cycles, foundries with negotiating power pass some costs to clients, squeezing lower-margin segments.
- Concentration risk: Over-reliance on a handful of suppliers makes any disruption (energy, logistics, regulatory restrictions, incidents) a systemic risk.
Add to that the fact that AI growth not only pressures top-tier nodes but also drives “collateral” demand (for controllers, connectivity, energy management, sensors), which are often produced in mature nodes. TrendForce already described this mix in 2025: the foundry market was improving due to AI demand in HPC and consumer launches, while Chinese foundries gained business from supply chain diversification.
A deeper insight: “mature” doesn’t mean “irrelevant”
The public debate often focuses on 2 nm, 3 nm, or 5 nm nodes, but a large part of industrial volume—and China’s muscle—is in mature nodes. These are chips that don’t make headlines but power everything from modern cars to electrical infrastructure. As several analyses have pointed out, “mature” doesn’t mean stagnant: they are constantly refined for new demands.
With global revenues already exceeding a trillion yuan and four Chinese companies firmly in the Top 10, the question for 2026 is no longer whether the market will grow. The real issue is who controls capacity—and under what conditions—in the most expensive and strategic cycle the industry has ever experienced.
Frequently Asked Questions
What is a “pure-play foundry” and why does it matter?
It’s a company dedicated to manufacturing chips for third parties (not selling its own products). It’s important because much of the innovation in fabless design and AI hardware depends on their capacity, prices, and lead times.
What does it mean that four Chinese foundries are in the Top 10?
It shows that China is no longer just competing in future capacity but has actors with real volume and a presence in global contract manufacturing, especially in mature nodes and specialized semiconductors.
Why are mature nodes still critical in 2026?
Because many essential chips (power devices, automotive, controllers, sensors) don’t need the most advanced node but do require a stable industrial supply chain. “Mature” doesn’t mean obsolete: these nodes are continuously optimized for new applications.
Does AI only drive the demand for top-tier chips?
No. Besides advanced GPUs/NPUs, AI also boosts demand for “supporting” components (power, connectivity, control), often fabricated on mature nodes, adding pressure to total capacity.