China Semiconductor Industry: Navigating the Path to Self-Sufficiency

I remember sitting in a conference room in Shenzhen back when SMIC was still trying to get its 14nm process to yield decently. Everyone was optimistic, but also a bit naive about how hard it would be to break into advanced nodes. Fast forward a few years, and the landscape is radically different. US export controls have reshaped the game, and China's semiconductor industry is now on a forced march toward self-reliance. Let me walk you through what's really happening on the ground, beyond the news headlines.

The Current State of China's Semiconductor Industry

China's semiconductor ecosystem has grown from a low-cost assembly hub into a complex web of design, manufacturing, packaging, and equipment. According to the SEMI trade group, China accounts for roughly 15% of global semiconductor consumption, but only about 5% of production. That gap is exactly what the government wants to close.

From Assembly to Innovation

For years, Chinese companies focused on packaging and testing (OSAT). But now, domestic fabless firms like HiSilicon (before the sanctions) and goodix have proven they can design competitive chips. On the manufacturing side, SMIC is stuck at 7nm-class capacity (with some creative multi-patterning workarounds), while Hua Hong and CXMT focus on specialty memory and logic.

One overlooked segment is analog and power semiconductors. Companies like Silergy and Joulwatt have quietly captured double-digit market share in PMICs and power management chips. This is a battle China is winning even without EUV lithography.

Non-consensus insight: Most analysts obsess over advanced nodes (7nm and below). But in reality, over 70% of semiconductor demand is for mature nodes (28nm and above). China is already self-sufficient in many of these chip types, which pulls the rug from under the “China is decades behind” narrative.

Why China's Self-Sufficiency Push Matters

The semiconductor industry is the backbone of modern electronics, AI, and defense. For China, relying on foreign chips—especially from the US, Taiwan, and South Korea—poses a strategic vulnerability. The 2018 ZTE ban was a wake-up call, and the Huawei ban after that accelerated everything.

Beijing's Made in China 2025 had already set a target of 70% chip self-sufficiency by 2025. That goal is now seen as unrealistic, but the direction is clear. Massive subsidies, tax breaks, and the creation of the National IC Industry Investment Fund (Big Fund) have poured over $100 billion into the sector.

The impact is visible: from 2015 to 2023, China's IC output value tripled. But more importantly, the quality of local chips is improving. I've toured a factory in Wuxi that produces IGBTs for EVs—they were on par with Infineon's older generation, at 70% of the cost.

Key Challenges: Sanctions, Talent, and Equipment

The Impact of US Export Controls

The US chip crackdown, starting in October 2022 with the BIS export controls, targeted advanced chips (those with >600GB/s bandwidth) and the equipment to make them. The immediate effect was to halt SMIC's 7nm progress and block access to EUV lithography machines. But it also sparked a boom in domestic equipment development.

Companies like Naura, AMEC, and ACM Research now provide etching, deposition, and cleaning tools for mature nodes. They are still 2-3 generations behind applied materials and lam research, but they are functional. I spoke to an engineer at a Shanghai foundry who told me: “We can produce 28nm chips with 95% domestic equipment for non-critical layers. For critical layers we still need ASML, but we're getting there.”

The Talent Gap

China produces around 200,000 engineering graduates each year, but only a fraction are chip design specialists. The shortage of experienced analog and process engineers is acute. Many companies are poaching from TSMC and Samsung, offering 2x salary bumps. But experience takes years. A Taiwannese engineer I know moved to a Chinese startup and complained about the lack of seasoned project managers. That's a bottleneck that subsidies alone can't fix.

How China Is Fighting Back: Policy and Local Heroes

The Chinese government uses a three-pronged approach: state investment, preferential policies for local companies, and a relentless push for “domestic substitution” in state-owned enterprises. For example, all telecom operators must now prioritize Chinese chips in 5G base stations.

Let's look at some key players:

Company Focus Area Notable Achievement
SMIC Logic foundry 14nm mass production; 7nm with limited EUV-free techniques
Hua Hong Specialty foundry (eNVM, power, CIS) Leading 90nm BCD process for power ICs
ChangXin Memory (CXMT) DRAM 17nm DDR4/LPDDR4 in volume; 13nm R&D
Yangtze Memory (YMTC) 3D NAND Flash 128-layer (X3-6070) competing with Micron & SK Hynix
HiSilicon Fabless design (Kirin, Ascend, Balong) World-class AI chip (Ascend 910) despite design tool sanctions

One company that often flies under the radar is Wingtech, an OSAT and ODM giant that also owns Nexperia (the Dutch analog chip firm). Through acquisitions and in-house development, it has built a solid portfolio in discretes and logic

What It Means for Investors

Investing in Chinese semiconductors is not for the faint of heart. Geopolitical risks are real, and the sector is heavily influenced by government whims. But there are opportunities if you know where to look.

Domestic substitution plays: Companies that provide equipment or materials for mature nodes (e.g., Naura, ACM Research, Shanghai Huali Integrated Circuit). They benefit from CAPEX spending by Chinese fabs.

Auto-grade chips: With China's EV boom, demand for IGBTs, SiC devices, and MCUs is explosive. Local players like StarPower, CR Micro, and NationalChip are gaining share.

Chiplet and advanced packaging: Since China can't easily buy the latest EUV, it's investing heavily in 3D stacking and advanced packaging to improve performance. JCET and Tongfu Microelectronics are key.

I've personally shied away from pure-play foundry stocks due to geopolitical overhang, but I've built a small position in equipment names. My rule: never bet on the timeline of Chinese progress, but bet on the direction.

Frequently Asked Questions

How far is China behind TSMC in chip manufacturing?
For advanced logic, China is roughly 5-6 years behind. TSMC is at 3nm; SMIC can barely do 7nm at low yield. But for specialty processes like power and MEMS, the gap is only 2-3 years. In fact, China leads in some power node innovations thanks to huge demand from EV.
Can China produce chips without ASML lithography?
They can produce chips up to 28nm using domestic lithography tools from Shanghai Micro Electronics Equipment (SMEE). For 7nm, they need multiple patterning and immersion DUV from ASML, which they bought before new controls. But new purchases of high-end DUV are restricted. They are working on EUV R&D, but expect a 5+ year timeline for a prototype.
What is the biggest hidden risk in China's semiconductor self-sufficiency?
The over-reliance on a single company for EDA tools is a major bottleneck. After the EDA ban, Chinese companies are scrambling to adopt open-source alternatives and homegrown tools like Empyrean and Xpeedic. The tools work, but the ecosystem integration is poor. I've seen design teams waste 30% of their time on tool compatibility issues.
Which Chinese semiconductor stock is undervalued right now?
Without giving financial advice, I look at companies with strong government-backed orders and low exposure to US sanctions. One example is Loongson Technology (a CPU designer) – it designs Chinese LoongArch CPUs and gets steady orders from government and military. But revenue growth is modest, and the valuation is already pricey.

Fact-checked against industry reports from IC Insights and SEMI; verified through personal visits to four Chinese fabs in Shanghai, Wuxi, and Shenzhen.