Ramp-Up of Chinese Foundry Capacity in 2025 May Intensify Price Competition in Mature Process Nodes

[Daniel Nenni]

Richard CHENRichard CHEN • | SemiconductorsMarketing Analyst |

China's wafer foundry players are continuing to scale up capacity, especially in mature process nodes—a move that could put increasing pricing pressure on global foundries relying on legacy nodes.

Key Highlights:
SMIC & Nexchip Expansion (2024)
SMIC's Semiconductor Manufacturing South China (SMSC) is slightly increasing its advanced-node capacity (14nm and below), while Nexchip N3 Fab is ramping up to produce 100,000 wafers/month at 55nm and 40nm nodes, with a focus on high-end CIS (CMOS Image Sensors).

New Fab Developments (2024–2025)
1. SMIC Beijing (SMBC): Two-phase construction. Once completed, monthly 12-inch wafer capacity will hit 100,000 units.
2. Huahong Wuxi Phase-2: Targeting automotive chips. Production starts post-December 2024, scaling to 20,000 wafers/month in Q4 2025 and 83,000/month by 2027.

Strategic Moves by Huahong Group
Huahong has acquired GlobalFoundries’ Chengdu fab. Production is expected to start by the end of 2026, aiming for 30,000 wafers/month under subsidiary HLMC.

As China builds out massive mature-node capacity, global foundries may face intensifying price competition, particularly those with limited ability to differentiate or absorb cost pressures.

Let’s keep an eye on how this shapes the global semiconductor supply chain going forward.

 

[Paul2]

The COVID played a trick with the industry: by far, the most modern 200mm inch fabs are in the mainland China, because China was the last country to jump on the semiconductor train.

Most of them were making really outdated chips from their own mainland China's captive clients, like SOEs, or companies who just bought their design as a physical mask-set in a box, while Taiwanese, and European 200mm fabs were serving far more richer, and bigger clients like car chip makers.

Car-centric microcontroller makers withdraw orders, and 200mm foundries outside of China all try to fill their order queue with anything to keep them running. Then comes the economic bounceback, and 200mm service price explodes, and Chinese 200mm fabs make huge money, as the only ones with spare cap, and flexibility to move their own products aside for newly coming $10000 per 200mm wafer orders.

 

[Daniel Nenni]

Very true. I was in China when COVID hit and I spent a lot of time there prior to the pandemic. China has gone from very little semiconductor manufacturing to a lot of semiconductor manufacturing. From my experience, however, China is a brute force manufacturer. So much money was spent to get to where they are today and there was so much waste but the ends just might justify the means.

I'm at CDNLive, Jensen Huang is speaking, his view of China may be seen as self serving but I agree. We should be partnering with China not blocking them. A semiconductor independent China, if that is even possible, is not a good thing for the world. The semiconductor ecosystem is missing huge amounts of revenue and innovation opportunities.

 

[Paul2]

partnering with China not blocking them

TSMC was pretty much blocking them until they started chasing latest nodes.

TSMC's offers were extremely competitive even on legacy nodes. Even with complete zero tax, and even net positive subsidy, mainland fabs weren't able to approach TSMC on cost-service-quality. They weren't even approaching UMC. in other words, TSMC was blotting out the sky, even at nodes with very low profitability, and that was not letting mainland fabs to self-sustain, and catch up.

But everything changed when TSMC started to prioritise latest node customers at all costs, including the retreat from mature nodes, which was protecting them from the prospect of low-end competitors growing, and catching up to them.

 

[MKWVentures]

I would expect overcapacity across the board. The question will be the quality of the foundry service