The Country Willing to Burn $70 Billion to Make Its Own Chips — Why China's 15th Five-Year Plan Should Keep You Up at Night
Summary
China just crossed 35% semiconductor equipment self-sufficiency and is now betting $70 billion on a plan that could render US tech containment obsolete. If this works, the entire global chip market gets rewritten.
Key Points
$70 Billion — The Largest Semiconductor Incentive Package in History
China's 15th Five-Year Plan allocates up to $70 billion in incentives for the semiconductor industry alone. This surpasses the US CHIPS Act ($52.7 billion) and operates separately from the existing Big Fund III ($47.5 billion). Combined, over $120 billion is being concentrated on a single sector. Comparable to the Cold War-era Apollo program, this investment represents not mere industrial promotion but a national response to US-led tech containment, carrying fundamentally different strategic weight.
SMIC Hitting 7nm Without EUV
SMIC has achieved 7nm-class chip mass production using DUV multi-patterning technology without access to ASML's EUV equipment. While yield issues remain, this technological achievement exposes fundamental limitations of US export controls. It proves that creative engineering can achieve remarkable semiconductor manufacturing with restricted equipment. The fact that ASML sold 70% of its DUV systems to Chinese entities in 2024 reveals the scale of regulatory loopholes.
Equipment Self-Sufficiency Breaks 35% — Beating Targets Ahead of Schedule
By end of 2025, China's semiconductor equipment self-sufficiency rate hit 35%, surpassing the original 30% target. This represents a 10 percentage point jump from 25% in 2024 in just one year. Etching and thin-film deposition exceeded 40%, with local equipment makers like ACM Research, AMEC, and Naura seeing rapid market share growth. A policy mandating at least 50% domestic equipment in fabs is accelerating technical improvement, with projections reaching 31% by 2027.
Entering the HBM War — Challenging the Nvidia Ecosystem
CXMT has begun mass-producing HBM2 and is building HBM production lines in Beijing and Hefei, targeting HBM3 in 2026 and HBM3E in 2027. HBM is a critical component for AI data centers, currently dominated by SK hynix and Samsung. China's entry into HBM competition could reshape the AI semiconductor supply chain and create cracks in Nvidia's AI infrastructure ecosystem. Combined with Huawei's Ascend AI chips, this completes China's indigenous AI hardware stack.
The Dawn of Technology Bifurcation
If this plan progresses successfully, the global semiconductor market will likely split into two ecosystems: an advanced semiconductor alliance led by the US, Netherlands, Japan, South Korea, and Taiwan on one side, and a self-sufficient China-centered ecosystem on the other. Companies will need to maintain dual technical standards, dual supply chains, and dual design platforms simultaneously, leading to higher costs and slower innovation. Consumers will ultimately bear the price.
Positive & Negative Analysis
Positive Aspects
- Acceleration of Technology Self-Reliance
The combination of $70 billion in investment and mandatory 50% domestic equipment policies is rapidly boosting China's semiconductor industry self-sufficiency. Visible results are already emerging — 35% equipment self-sufficiency, SMIC's 7nm mass production, CXMT's HBM development launch. If this trend continues, China could achieve practical semiconductor self-sufficiency within five years.
- Global Supply Chain Diversification Effect
Advanced semiconductor supply is currently over-concentrated on TSMC, exposing the world to risk in a Taiwan Strait crisis. China's strengthening semiconductor capabilities paradoxically contribute to geographic diversification of the global supply chain. More semiconductor production hubs mean better buffering against geopolitical shocks in any single region.
- Innovation Through Competition
Chinese companies' challenge serves as an innovation catalyst for existing semiconductor giants. As Samsung, SK hynix, TSMC, and ASML increase R&D investment to maintain their edge, the overall pace of semiconductor technology advancement accelerates. Competition drives innovation better than monopoly — that's basic economics.
- Alternative AI Hardware Ecosystem Formation
If a Chinese AI hardware stack combining Huawei Ascend, CXMT HBM, and indigenous EDA tools is completed, it creates an alternative to Nvidia's monopoly. Nvidia holding over 80% of the AI chip market isn't healthy, and the emergence of a competing ecosystem could contribute to lower prices and more stable supply.
Concerns
- Structural Questions About Investment Efficiency
China's semiconductor investment history is littered with massive failures like the Wuhan HSMC debacle. Bureaucracy-driven industrial policy risks ignoring market signals and wasting resources on political objectives. The Made in China 2025 target of 70% self-sufficiency landing at roughly 50% reflects this structural problem. There's no guarantee the $70 billion will be efficiently allocated.
- Global Cost Escalation From Technology Bifurcation
If the semiconductor market splits into two ecosystems, global companies face enormous additional costs for dual certification, dual design, and dual supply chain management. As standardization benefits disappear and economies of scale weaken, electronic product price increases become inevitable. The cost of tech hegemony competition ultimately gets passed to consumers.
- The Fundamental Bottleneck of Talent Shortage
Semiconductor manufacturing is extremely talent-intensive. China has the money but lacks sufficient engineers with cutting-edge process experience. Even the Shenzhen EUV prototype required help from former ASML engineers. Geopolitical barriers hinder overseas talent recruitment, and domestic talent development takes time. Experience is the one thing money can't buy.
- Amplification of Geopolitical Risk
As China's semiconductor self-sufficiency grows, US containment intensity is likely to escalate. US lawmakers are already demanding bans on ASML's DUV sales to China. This escalation pattern increases losses for both sides and could, in extreme cases, lead to military tensions. Weaponizing semiconductor supply chains as a security tool is dangerous for all involved nations.
- The Wall of Yield and Quality
Achieving 7nm mass production and producing quality chips at scale are entirely different challenges. Industry assessments suggest SMIC's multi-patterning process has significantly lower yields than Samsung's or TSMC's EUV-based processes. Higher per-unit costs undermine price competitiveness, limiting the practical market penetration of Chinese-made chips.
Outlook
China is unlikely to catch TSMC's cutting-edge processes anytime soon. The gap at 3nm and 2nm doesn't close by throwing money at it. But China's real goal may be mass-producing "good enough" chips domestically. At 7nm, most demand can be met — smartphones, servers, AI inference, industrial semiconductors, automotive chips. Once China can fully self-supply these chips, US export controls effectively become toothless. By 2028-2029, if equipment self-sufficiency crosses 50% and HBM3 mass production stabilizes, the global semiconductor market enters a true era of bifurcation. Long-term, companies face dual standards, dual supply chains, and dual design platforms. The best scenario is both sides maintaining cross-dependency. The worst is a geopolitical shock like a Taiwan Strait crisis paralyzing supply chains instantly.
Sources / References
- China mulls $70 billion domestic chip fabrication injection — Tom's Hardware
- China's semiconductor equipment self-reliance surges past targets — South China Morning Post
- China's 15th Five-Year Plan Recommendations - Key Takeaways — China Briefing
- Where has China's push for chip self-reliance taken it? — The Korea Times
- The Export Control Loophole Fueling China's Chip Production — CNAS
- China's next move: the Five-Year Plan that could reshape semiconductors — Yole Group
- America's chip export controls are working — Asia Times