In 2024, China launched the third instalment of its National Integrated Circuit Industry Investment Fund, a $48 billion investment vehicle to turbocharge the nation’s semiconductor industry, even as it continues to lag behind industry leaders. The country has decided to double down on what appears to be an economically irrational strategy. While the chipmaker TSMC produces advanced chips with 90% yield rates, China’s SMIC struggles with just 30% yield rates. As a result, the economics are brutal: chips produced by SMIC costs between 40% to 50% more while delivering a third of the volume compared to chips manufactured by TSMC. But this isn’t about quarterly profits—it’s about technological sovereignty in the age of artificial intelligence (AI). After the US weaponized China’s semiconductor dependencies, China faces an existential decision. It can either accept permanent technological subordination to the US and its allies, or pay any price to get into the game of semiconductor manufacturing. China has chosen to pay and play, betting that the cost of exclusion and subordination far exceeds the exuberant cost of participating in the most critical technology race of the next decade.

In 2024, China invested around $48 billion in semiconductor manufacturing through the third phase of its National Integrated Circuit Industry Investment Fund, colloquially known as the “Big Fund,” pouring resources into an industry where it faces seemingly insurmountable technological and economic disadvantages. The yield rates of the advanced chips manufactured by Chinese chipmakers like SMIC are poor, and the costs of manufacturing are astronomical. By traditional business metrics, this investment appears futile and borderline irrational.

Yet, China continues to double down, treating the development of semiconductor manufacturing as a matter of national survival rather than a business plan fixated on unit-economics. This is a calculated bet that these series of investments will elevate China to become the one of only few players in semiconductor manufacturing, arguably the most critical technology at this moment given its role in powering the current revolution and advancements in artificial intelligence (AI).  

The scale of this commitment is staggering. China’s spending on wafer fabrication equipment alone jumped from $11 billion in 2018 to nearly $30 billion in 2023, a 170% increase that dwarfs most countries’ entire technology budgets. The domestic semiconductor industry has achieved an unprecedented 30.6% annual growth rate, with total sales reaching $39.8 billion in 2020. More striking still, China’s self-sufficiency rate has rocketed from around 5% in 2018 to an estimated 12% in 2023.

These staggering growth rates reflect a fundamental recalibration of national priorities. When President Xi Jinping declared in 2014 that semiconductors were a “core technology” that must be produced domestically, he wasn’t making an economic argument. Rather, Xi was articulating his plan of aligning the country’s resources to achieving technological sovereignty, no matter the costs. 

The Strategic Stakes: Why Semiconductors Matter

Semiconductors are nicknamed the “oil of the 21st century” for good reasons. They underpin virtually every technology that powers the modern, from smartphones and autonomous driving vehicles to missile guidance systems and artificial intelligence. More critically, advanced semiconductors below 7 nanometers (nm) have become the bedrock to the current AI revolution, powering the datacenters that run large language models (LLMs) such as ChatGPT, Gemini, and DeepSeek.

This technological centrality creates unprecedented strategic vulnerability. Unlike oil, which can be sourced from multiple suppliers and stored in strategic reserves, advanced semiconductor manufacturing is constantly evolving and is concentrated in a handful of companies and countries. Currently, the global supply chain for semiconductor manufacturing resembles a complex web of interdependencies, with each stage dominated by different players across different regions.

The design stage is heavily concentrated in the United States, accounting for roughly 47% of semiconductor design process in 2023. American firms like Nvidia, AMD, Intel, and Broadcom have established commanding positions in different chip categories. Nvidia’s dominance in graphics processing units (GPUs) has proven especially crucial, as these chips perform the parallel computations essential for AI training and inference in data centers worldwide.

Manufacturing equipment tells a similar story of concentration. The Netherlands holds a monopoly on the most advanced manufacturing equipment with ASML’s lithography machines, which use extreme ultraviolet (EUV) light to etch microscopic patterns onto silicon wafers. Applied Materials, an American firm, dominates deposition, etching, and ion implantation equipment. Japan’s Tokyo Electron and other specialized companies control critical materials and processes.

Taiwan dominates advanced manufacturing through Taiwan Semiconductor Manufacturing Company (TSMC), which by some estimates produces over 90% of the world’s most advanced chips. Outside of TSMC, South Korea’s Samsung Electronics is one of the few alternatives for cutting-edge manufacturing.

China, despite being the world’s largest semiconductor consumer, accounting for more than 50% of global demand, has been relegated primarily to assembly, testing, and packaging. These backend operations, while important, represent the lowest value-added segments of the semiconductor supply chain. They require less sophisticated equipment, generate lower margins, and create fewer opportunities for technological advancement.

This concentration creates what strategists call “choke points”—single points of failure that can shut down entire industries. For China, this reality transformed from theoretical concern to existential threat when the United States began weaponizing these dependencies to hinder China’s technological development at the dawn of the AI race.

The Chokehold: How America Exploited Semiconductor Dependencies

The turning point for China’s national strategy on technology came in 2018-2019 when the Trump administration began systematically restricting China’s access to advanced semiconductor technologies. This represented a fundamental shift in American strategy from economic competition to technological containment, and the weapon of choice was ASML’s extreme ultraviolet (EUV) lithography machines, the only equipment capable of manufacturing chips below 7 nanometers at scale.

The physics of semiconductor manufacturing make EUV machines irreplaceable. To create transistors at 7-nanometer scales and below, manufacturers must use light with wavelengths shorter than traditional deep ultraviolet (DUV) systems. EUV systems generate light at 13.5 nanometers, roughly 14 times shorter than DUV, allowing them to create much smaller features with fewer processing steps. ASML’s monopoly on EUV technology stems from decades of development and over $4 billion in annual R&D investment, creating insurmountable barriers to entry.

Since ASML holds a de facto monopoly on advanced lithography machines, blocking Chinese access to these machines effectively capped China’s technological advancement by removing the manufacturing infrastructure necessary for future development. Without access to EUV technology, Chinese manufacturers would be permanently relegated to older, less efficient production methods.

The restrictions tightened progressively. In October 2022, the Biden administration implemented what many consider the most comprehensive technology export controls in modern history, with the “foreign direct product rule” (FDPR) extending US export controls to any product containing American technology or software. This gave Washington leverage over ASML’s older deep ultraviolet (DUV) machines as well. This rule meant that even equipment manufactured entirely outside the United States could be subject to American export controls if it contained any American components or intellectual property.

The Netherlands, under sustained American pressure, began requiring export licenses for advanced DUV systems to China in 2023. Japan implemented similar restrictions on critical materials and equipment. These weren’t independent decisions and reflected coordinated American diplomacy aimed at creating a comprehensive technology embargo.

The scope of these restrictions extended beyond manufacturing equipment. US also targeted Chinese access to electronic design automation (EDA) software, where American companies hold a 95% market share. EDA software is essential for designing modern semiconductors, handling everything from circuit layout to simulation and optimization. Without access to cutting-edge EDA tools, Chinese companies would struggle to design competitive chips even if they could manufacture them.

Further, the restrictions included American talent and expertise. US citizens and permanent residents were prohibited from working for Chinese semiconductor companies without licenses, cutting off another potential pathway for technology transfer. This “brain drain” prevention measure reflected growing recognition that semiconductor development depends as much on human capital as physical equipment.

This series of policies represented a strategic containment designed to preserve American technological supremacy. Critically, the timing was crucial: advanced semiconductors had become the foundational technology for AI development, and whoever controlled AI capabilities would likely dominate the next phase of economic and military competition.

For China, the message was clear: continued dependence on foreign semiconductor technologies meant permanent technological subordination. The country that had lifted 800 million people out of poverty through manufacturing prowess suddenly found itself locked out of the most critical manufacturing process of the AI era.

China’s Response: Massive Investment Despite Steep Odds

Faced with technological encirclement, China chose escalation over accommodation. The country’s semiconductor investment strategy, formalized in the “Made in China 2025” plan, set an ambitious target of 70% self-sufficiency by 2025. But the real acceleration came during the Trump administration’s trade war, when Chinese companies like Huawei were effectively cut off from global semiconductor supply chains.

The Chinese approach defied conventional wisdom about semiconductor development. Rather than focusing on niche applications or accepting technological dependence, China committed to developing capabilities across the entire semiconductor value chain. This meant simultaneous investment in design tools, manufacturing equipment, materials science, and production facilities, an approach that multiplied both costs and complexity.

The numbers reflect China’s extraordinary commitment. Beyond the headline figures of equipment spending, China established multiple “Big Funds” for semiconductor development, with the first fund raising $19 billion in 2014 and the second raising $29 billion in 2019. Most recently, the third fund raised $48 billion. Provincial and local governments added their own funding, creating a complex web of financial support. The Economist Intelligence Unit estimates that China’s total state-led investment in chips has likely exceeded $150 billion in total commitments since 2014.

This investment enabled rapid progress in mature semiconductor technologies. Chinese companies now dominate the production of chips at 28 nanometers and above, turning these “mature nodes” into commodities. Companies like SMIC, Hua Hong Semiconductor, and others have achieved competitive cost structures and reliable production at these technology nodes, serving automotive, industrial, and consumer electronics markets.

The crowning achievement came in 2023 when Huawei, partnering with SMIC, successfully produced a 7-nanometer chip for the Mate 60 Pro smartphone. This breakthrough shocked industry analysts who had expected US sanctions to cripple China’s advanced semiconductor capabilities. The achievement was particularly significant because it demonstrated Chinese capabilities in system-on-chip design, advanced packaging, and high-volume manufacturing.

The technical accomplishment was remarkable. SMIC achieved 7-nanometer production using DUV equipment exclusively, requiring innovative multi-patterning techniques and process optimization. The company reportedly uses 34 lithography steps compared to the 9 steps typically required with EUV equipment, which is a testament to both Chinese engineering ingenuity and the lengths to which the country will go to overcome technological barriers.

Yet this success story masks serious underlying challenges. China’s 7 nm chips rely heavily on foreign technology, manufactured using older DUV machines from ASML and equipment from US companies Applied Materials and Lam Research, which were purchased before the October 2022 restrictions took full effect. The country remains vulnerable in critical areas beyond lithography, including advanced materials, specialized chemicals, and precision manufacturing equipment.

The design challenge is equally daunting. While Chinese companies have made progress in some areas of chip design, they remain heavily dependent on American EDA software and intellectual property. Companies like Synopsys and Cadence have spent decades developing sophisticated tools for advanced chip design, creating software ecosystems that are difficult to replicate. Chinese EDA companies exist but are generally considered “usable,” but are far from competitive for cutting-edge designs.

Perhaps most critically, China’s semiconductor progress depends on a shrinking window of available technology. As American restrictions tighten and allied countries implement their own controls, the technological foundation for China’s semiconductor industry faces erosion. Equipment breaks down, software licenses expire, and technical support disappears, creating a race between Chinese indigenous development and technological obsolescence.

The Economics Problem: Why the Numbers Don’t Add Up

From a purely economic perspective, China’s semiconductor push looks increasingly unsustainable. The fundamental problem is yield, the percentage of functional chips produced from each wafer. SMIC’s 7 nm process achieves roughly 30% yield, far below the ideal industry benchmark of 90% for mobile chips fabrication. This deficiency stems from using DUV instead of EUV machines, requiring those 34 lithography steps instead of 9, dramatically increasing error rates and processing time.

The mathematics of semiconductor economics are unforgiving. With each additional processing step, the probability of defects increases exponentially. A wafer that passes through 34 lithography steps faces far higher defect rates than one passing through 9 steps. Combined with the inherent challenges of working at atomic scales, these extra steps create a yield penalty that makes economic production extremely difficult.

The unit economics are brutal: each 7 nm chip SMIC produces costs 40-50% more than TSMC’s equivalent, while delivering only a third of the yield. These cost disadvantages compound throughout the supply chain. Higher production costs mean higher chip prices, which reduce demand and market competitiveness. Lower yields mean higher fixed costs per functional chip, creating a vicious cycle of economic disadvantage.

The issues of costs extend into the construction of fabrication facilities (Fab) as well, in which a single advanced Fab can cost $15-20 billion to build and equip. Operating costs run into billions annually, with TSMC having spent between $32- $36 billion in capital expenditures in 2023 alone, with additional billions in R&D. These costs must be amortized across millions of chips, making yield rates absolutely critical to economic viability.

China’s response has been massive government subsidization on a scale that distorts traditional market analysis. Huawei reportedly received $948 million in direct funding in 2022, though the actual support also includes land grants, construction subsidies, tax breaks, and indirect benefits likely exceeds this figure by orders of magnitude. The entire Chinese semiconductor ecosystem operates on government life support, with state-owned enterprises, development funds, and local governments providing financing that private investors would never risk.

The riskiness and immense difficulty can be illustrated by the struggles of Intel, an established US semiconductor firm that designs and manufactures chips. Despite being a semiconductor pioneer with decades of manufacturing experience, Intel’s foundry division responsible for chip manufacturing reported $7 billion in operating losses in 2023. The company has struggled to achieve competitive yields in advanced nodes, leading to considerable market share losses. If Intel, with its technological heritage and established customer base, struggles with semiconductor manufacturing economics, China’s position appears even more precarious.

The talent challenge adds another layer of complexity. Semiconductor manufacturing requires highly specialized expertise that takes years to develop. Engineers must understand quantum mechanics, materials science, chemistry, and precision manufacturing, forming a combination of skills that exists in limited supply globally. China’s restrictions on American talent, combined with brain drain as Chinese engineers emigrate from the country to seek better opportunities abroad, create human capital constraints that money alone cannot solve.

Moreover, equipment maintenance presents an ongoing challenge. Advanced semiconductor manufacturing equipment requires constant calibration, replacement parts, and technical support. As American and allied restrictions tighten, Chinese manufacturers face growing difficulties maintaining their existing equipment base. When critical equipment breaks down, repair may become impossible, potentially idling entire production lines.

The Strategic Logic: Optionality Over Profitability

China’s persistence makes perfect sense when viewed through a strategic rather than economic lens. After all, the country isn’t trying to build a profitable semiconductor industry in the traditional sense, at least not immediately. Rather, it is buying optionality in the most critical technology of the coming decades. This represents a fundamental shift from commercial to strategic logic, where success is measured not in quarterly profits but in technological sovereignty and future competitive position.

This logic becomes clearer when examining similar decisions by US technology companies. Meta spent $14.3 billion acquiring a 49% stake in Scale AI in 2025, despite uncertain returns on AI infrastructure investments, mainly to poach the company’s AI talents to build its “Superintelligence” Team. Google, Microsoft, and Amazon are collectively spending over $200 billion annually on AI infrastructure, with no guarantee of profitability. These investments are justified not by current returns but by future positioning in what these companies see as the next computing paradigm.

Mark Zuckerberg captured this reasoning perfectly: “The downside of being behind is that you’re out of position for the most important technology for the next 10 to 15 years.” Google’s CEO Sundar Pichai echoed this sentiment: “When we go through a curve like this, the risk of underinvesting is dramatically greater than the risk of overinvesting.”

China’s semiconductor investment follows identical logic. The country is purchasing a ticket to participate in the race for the next phase of technological development in semiconductors and AI, rather than optimizing for near-term profitability. Without domestic semiconductor capabilities, China risks being permanently relegated to technological dependence, unable to fully participate in AI development, autonomous systems, advanced manufacturing, or military modernization.

The strategic calculus is straightforward: the cost of exclusion far exceeds the cost of participation. If China accepts permanent technological dependence, it surrenders control over its economic future. Every major technological transition, from mainframes to personal computers to mobile devices to cloud computing, has created new winners and losers. The countries and companies that missed these transitions found themselves permanently disadvantaged for the next revolution. 

The AI revolution represents the most significant technological transition since the internet. Advanced semiconductors are the foundational technology enabling this transition. Countries without indigenous semiconductor capabilities may find themselves locked out of AI development, advanced manufacturing, and other critical technologies. For China, this represents an existential threat to its technological sovereignty and economic development. 

The Compound Effect: Why Early Investment Matters

Semiconductor development exhibits strong path dependence and increasing returns, by which early advantages compound over time through accumulated expertise, specialized talent, and ecosystem development. For instance, Taiwan’s dominance in advanced manufacturing stems from decades of sustained investment, technology transfer, and ecosystem development that created self-reinforcing advantages.

The semiconductor industry rewards scale and specialization in ways that create natural monopolies. TSMC’s current dominance reflects the cumulative advantages of serving the world’s leading chip designers through which each new customer provides learning opportunities, revenue for R&D investment, and validation for process improvements. These advantages create barriers to entry that grow higher over time.

China understands this dynamic intimately. The country’s manufacturing prowess in industries like steel, solar panels, and electric vehicles emerged through similar patterns of initial investment, technology absorption, and gradual capability building. In each case, China accepted initial economic losses in exchange for learning and long-term competitive positioning. The strategy worked because China combined massive scale, patient capital, and systemic coordination in ways that private companies and smaller countries could not match.

By entering the semiconductor game now, despite unfavorable economics, China ensures multiple future opportunities to compete. Each generation of chip development leads to new learnings and creates new possibilities for breakthrough innovations or geopolitical realignments that could shift competitive dynamics. The company that sits out early development phases may find itself permanently excluded from future opportunities.

Huawei’s approach exemplifies this ecosystem strategy. Rather than attempting complete vertical integration, a model that has largely failed in the semiconductor industry, the company is creating a network of specialized firms that share resources and expertise. This includes equipment manufacturers like SiCarrier and TankeBlue, foundries like SMIC and newer facilities, design houses like HiSilicon, and materials companies throughout the supply chain.

The three manufacturing sites in Shenzhen’s Guanlan district represent this collaborative approach in action. With financial backing from the Shenzhen government and technical coordination from Huawei, these facilities are pushing the boundaries of what’s possible with existing equipment and expertise. Currently, Huawei operates one site producing 7 nm smartphone and AI chips, while SiCarrier and other partners operate complementary facilities.

This ecosystem approach allows China to distribute risks while concentrating resources. Rather than betting everything on a single company or technology approach, the country is developing multiple pathways to technological capability. If one approach fails, others may succeed. If technical obstacles prove insurmountable in one area, progress in related areas may create new opportunities.

The approach also creates redundancy and resilience. Traditional semiconductor companies rely on global supply chains that create vulnerabilities to sanctions and trade restrictions. China’s ecosystem approach aims to create domestic alternatives at each stage of the supply chain, reducing dependence on foreign suppliers and creating strategic depth.

The Broader Implications: Technological Sovereignty in the 21st Century

China’s semiconductor investment represents a broader shift toward technological sovereignty as a core national security priority. The country learned from its semiconductor vulnerabilities that dependence on foreign technology creates permanent strategic weakness. This lesson extends beyond semiconductors to AI, renewable energy, biotechnology, quantum computing, and other critical technologies.

The concept of technological sovereignty reflects a fundamental change in how nations think about economic development and national security. Traditional models of comparative advantage assumed that countries could specialize in their areas of strength while importing capabilities from others. The semiconductor sanctions demonstrated that this assumption breaks down when technologies become geopolitically contested.

China’s response has been to identify critical technologies where foreign dependence creates strategic vulnerability and invest heavily in domestic alternatives. This approach accepts short-term economic inefficiency in exchange for long-term strategic autonomy. The policy represents a form of technological nationalism that prioritizes independence over optimization.

The Chinese approach mirrors historical patterns of technological catch-up, but with important differences. South Korea’s semiconductor industry, Japan’s electronics sector, and Taiwan’s chip manufacturing all began with massive government investment, technology transfer, and initial economic losses. Each eventually achieved global competitiveness through sustained commitment and strategic patience.

However, China’s path will likely be longer and more difficult due to active technological containment by established players. Previous technological catch-up efforts occurred in more permissive international environments where technology transfer was easier and sanctions were less comprehensive. China must develop indigenous capabilities while facing unprecedented restrictions on technology access.

Conversely, the country’s advantages include vast domestic markets, substantial financial resources, and large pools of technical talent. China graduates more engineers annually than the rest of the world combined, creating human capital that can be directed toward strategic priorities. The domestic market for semiconductors provides both funding for development and demand for domestic production.

More importantly, China’s political system enables sustained investment in strategic priorities despite short-term economic losses. While democratic countries often struggle to maintain consistent industrial policies across electoral cycles, China’s centralized system can commit resources to long-term objectives. This institutional advantage may prove decisive in technology races that require sustained investment over decades.

The Logic of Strategic Necessity

China’s semiconductor investment defies conventional economic logic because it’s not primarily an economic decision. It represents a strategic imperative driven by the recognition that technological dependence equals permanent subordination in the 21st century. The country is essentially purchasing insurance against technological exclusion, regardless of the immediate cost.

This investment should be understood as a geopolitical strategy with profound implications for global technology competition. Like US companies investing in AI infrastructure despite uncertain returns, China is paying for the right to participate in shaping the technological future. The alternative, accepting permanent technological dependence, would be far more costly in the long run.

The semiconductor industry will likely fragment into competing technology ecosystems, with Chinese and American supply chains operating in parallel. This fragmentation will increase costs and reduce efficiency in the short term while creating redundancy and resilience in the long term. Both sides will pay economic costs for strategic independence.

The ultimate question isn’t whether China’s semiconductor investment makes immediate economic sense, it clearly doesn’t. The question is whether the country can sustain these investments long enough to achieve technological breakthrough and strategic autonomy. Given the stakes involved and China’s demonstrated commitment, the country appears willing to find out.

The broader implications extend beyond semiconductors to other critical technologies. As technological competition intensifies, both China and the United States will likely increase investments in strategic technologies regardless of economic returns. This represents a fundamental shift from economic optimization to strategic positioning that will reshape global technology development for decades to come.

In an era where technological capabilities determine national power, China’s semiconductor bet represents the price of admission to the most important game in town. Win or lose, the country has chosen to play rather than accept permanent technological subordination. The consequences of this choice will reverberate throughout the global economy and technology landscape for decades to come. 

Image credits: YOT LTD

Andy Liao
Andy is a graduate of the University of Toronto and a researcher of Chinese electric vehicles. He previously consulted for the World Bank on fintech in emerging markets and worked with the IFC analyzing over-indebtedness among women in Sri Lanka. In his free time, Andy enjoys reading and is currently training for a half marathon.