DeepSeek and China's Chip Awakening: Beyond the Moore's Law Trap

Emerging Technology Expert from Fudan University on How Chinese Chips Finding its own path

Feb 08, 2025

I've been staying out of the DeepSeek discussion for a relatively long time. First, I’m not a tech expert. Second, I haven’t done much research on it. However, I found an interesting interview with DeepSeek founder Liang Wenfeng, conducted by Chinese tech media 36Kr’s account, “暗涌Waves.” (Thanks

Zichen Wang

for sharing it) When asked whether he thinks the US excels at technological innovation while China excels at applications. Liang answered

with economic development, China must gradually become a contributor rather than always being a free rider. During the IT wave of the past thirty years, we barely participated in genuine technological innovation. We've grown accustomed to Moore's Law falling from the sky, lying at home waiting for better hardware and software to emerge every 18 months. Scaling Law is being treated the same way.
In reality, these were created through the tireless efforts of generations in the Western-led technical community. Only because we hadn't participated in this process before, we tended to overlook its existence.

This reminds me of another recent article that highlights the boost of an open and vibrant tech community. The article argues that Moore’s Law is both a gift and a barrier to the Chinese chips.

On one hand, Moore's Law has driven the continuous advancement of advanced manufacturing processes. On the other hand, the intense national focus on advanced process chips can cause businesses and countries to overlook the massive role of the market. Companies and decision-makers tend to use a simple quantitative metric to obscure the complex conditions needed for industrial development, mistakenly believing that simply throwing money at advanced process nodes will naturally lead to success and solve everything.

Decision-makers should focus on the market and domestic demand environment that supports advanced industrial development. Only when enough companies implement high fixed-cost strategies can China create a large enough market and domestic demand, and in turn, a larger market will encourage more companies to implement high fixed-cost strategies. This is the only way to forge a sustainable path for high-tech development. (Contrary to many’s first intuitions, DeepSeek didn’t get funding from the state but from a private quantitative hedge fund company. Don’t forget that hedge funds are relatively unfavored by the Chinese supervising body and went through tightening regulatory crosshairs in 2024) The role of national power is not to fund the companies or to guide the exact direction but to open up market space and create opportunities for innovative companies to implement high fixed-cost strategies.

The article named Breaking Free from the “Moore's Law Trap” - China's Semiconductor Industry's Path to Breakthrough, authored by Li Yin 李寅 and his student Gao Ke. It was first published on

Beijing Cultural Review

Li is an Associate Professor at the School of International Relations and Public Affairs at Fudan University. His research interest focuses on the economics of innovation, industrial policy, and emerging technology governance. His research has been published in leading innovation study journals, including Research Policy, Technovation, and JASIST.

The article reviews the history of China's chip development. During the Cold War, China was isolated from the global division of labor system with a weak economy, making it difficult to nurture a semiconductor market of sufficient scale. In the era of globalization, China's chip industry not only failed to break free from its technological dependence on foreign countries but actually deepened its reliance on overseas markets. As a result, on one hand, domestic emerging electronics industry giants were unwilling to use domestically produced chips, and on the other hand, domestic chip manufacturers were content to serve merely as outsourcing providers for multinational companies' low and mid-end chips and even fewer were willing to use domestically produced chip manufacturing equipment. Ironically, it took the US-China trade war and Trump's attempts to sever supply chains to ZTE and Huawei to catalyze meaningful change in China's semiconductor industry. The US restrictions ultimately strengthened rather than weakened Chinese chip-making capabilities.

Thanks to Professor Li’s authorization, I translated his full article. If you find his perspective helpful, you can also purchase his book China’s Drive for the Technology Frontier on Amazon. This book debunks the myths surrounding the Chinese model with a fresh take on China’s strategies for technological innovation. The central argument is that indigenous innovation is critical in transforming the Chinese high-tech industry. Like any successfully industrialized nation in history, indigenous innovation in China allows industrial enterprises to assimilate knowledge developed elsewhere, utilize science and technology resources and human capabilities accumulated in the country, and eventually approach the technological frontier.

It is worth mentioning that Dr.Li’s proposition that the economic essence of high-tech innovation is a high fixed-cost strategy was cited by ITIF in its report China Is Rapidly Becoming a Leading Innovator in Advanced Industries as evidence that US-China industrial competition is a zero-sum game. However, the author himself rebukes that the proposition emphasizes the significance of market demand for innovation, and for innovators, creating new markets is often more important than seizing market share from incumbents.

Thanks for reading Inside China! This post is public so feel free to share it.

Below is the full article I translated：

Breaking Free From the "Moore's Law Trap"

The Path to Breakthrough for China's Chip Industry

Advanced process chip manufacturing stands at the forefront of global technological and industrial competition. Since the United States enacted the CHIPS and Science Act to heavily subsidize chip manufacturing, Japan, India, Europe, and even Southeast Asian countries and regions have successively introduced chip industry subsidy policies. Developed countries like the US, Japan, and Europe are particularly targeting cutting-edge chip manufacturing below 5nm process nodes. Advanced process technology has become the focus of US efforts to contain China's high-tech industrial sector's further development. Over the past four years, the Biden administration has gradually tightened the technology export controls initiated by Trump, attempting to use the control of "chokepoint" key technologies by the US and its allies to lock China's chip manufacturing at the 14nm node. How China's chip industry can breakthrough these key technology restrictions and independently develop advanced chip manufacturing has become the most important challenge facing the technology and industrial sectors.

This article explores this topic from the perspectives of public policy and industrial history. Advanced process nodes—a dry and ambiguous engineering and marketing term that was only discussed by nerds a decade ago—has become a focus of international competition and a topic of public discussion under the promotion of public policymakers, which itself demonstrates the importance of policy discourse. However, few have reflected on whether it is reasonable to use the concept of advanced process nodes as a basis for industrial policy-making, as done by the US and Europe. While the role of advanced processes in improving chip performance is undeniable, policymakers' obsession with process nodes stems more from their faith in rules like "Moore's Law." By examining the history of global chip industry transformation and China's chip industry development, we can easily see that this blind faith is fundamentally unfounded.

Moore's Law, Product Platforms, and Global Chip Industry Transformation

In 1965, Gordon Moore, one of Intel's founders, summarized an industry pattern in a short article for Electronics magazine: since the invention of integrated circuits in 1958, the number of transistors that could fit on a chip roughly doubled every 18 to 24 months. This meant that approximately every two years, chip performance would double while the cost for the same performance would halve—this became the famous Moore's Law. Since its inception, Moore's Law has accompanied dramatic advances in chip technology and rapid semiconductor industry development, becoming the core high-tech industrial sector of the modern information society, and consequently, Moore's Law became a widely believed iron law.

The validity of Moore's Law has not been without controversy. In fact, Moore's Law is not a scientific theorem based on natural laws but rather an induction from experience—Gordon Moore initially only predicted ten years of development and didn't anticipate that chip technology could maintain rapid progress for so long, leading Moore himself to adjust the law's expression multiple times. Moore's Law itself is very broadly stated and easily reinterpreted in different contexts. For example, analog chip development has never followed Moore's Law. Around the beginning of the 21st century, the cost of shrinking individual transistor sizes to fit more transistors on each chip began rising increasingly, to the point where at multiple technology nodes, the cost increases from manufacturing difficulties began to exceed the benefits of transistor shrinkage, making next-generation transistors more expensive than previous generations. After 2014, Intel, which had long led technological development, took 5 years to complete the upgrade from 14nm to 10nm process, far exceeding the 2-year pattern. Even though Intel's slow progress had its own issues, Jensen Huang, CEO of NVIDIA—currently the most successful chip company—has repeatedly declared Moore's Law dead since 2017.

However, precisely because Moore's Law isn't a true scientific law, these controversies haven't really shaken its status. Moore's Law has had a lasting and profound impact on the entire chip industry because it established a rhythm of competition and created societal expectations (including government, industry, and consumers) for the chip industry's continuous technological advancement. This means that when incumbent companies fail to meet these expectations in some form, not only is their replacement by newcomers seen as natural (without recourse to common monopolistic tools like quality or brand), but catching up later may also be very difficult. For public policymakers, maintaining the technological leadership of flagship chip companies became the primary task in maintaining national cutting-edge industrial competitiveness. From this perspective, it's easy to understand why Washington politicians become extremely alarmed whenever U.S. semiconductor companies' leadership position wavers—the last time was during the Japanese challenge in the 1980s, and now it's TSMC replacing Intel in advanced process leadership. Under Moore's Law's guidance, advanced process nodes have become one of the most important industrial policy objectives.

But is Moore's Law really the most important driving force behind chip industry development? Looking back since the birth of integrated circuits, while rapid technological progress under Moore's Law has been the visible thread in chip industry development, the evolution of computer industry product platforms approximately every 20 years has been another "hidden thread" driving chip demand expansion and the succession of industry leaders. When first invented, integrated circuit chips were mainly used for aviation and military purposes at high costs, only finding their way into civilian markets in the 1960s. To date, there have been three major generations of computer product platforms driving chip demand.

The first generation, from the 1960s to the 1980s, was mainframe and minicomputers. IBM released its mainframe "System/360" system in 1964, introducing transistor circuit design and achieving huge market success. To maximize profits, IBM invested in its own chip manufacturing facilities and became the world's largest chip manufacturer in the 1970s just by producing chips for its own use. During the same period, startups like DEC and Wang Computers produced smaller, cheaper minicomputers using integrated circuit chips, opening up a larger civilian market.

The second generation, from the 1980s to the early 21st century, was personal computers (PCs). Affordable personal computers based on Intel's x86 microprocessors and Microsoft's operating systems triggered an explosion in the home computer market, launching the internet economy boom of the late 20th century. Since the 1990s, as PCs and servers became the largest chip market, Intel recovered from near bankruptcy to become the semiconductor industry leader, pioneering advanced processes for over two decades. Meanwhile, the previous generation's computer industry stars fell: Wang Computers went bankrupt in 1992, DEC was acquired by Compaq in 1998, and even giant IBM exited the PC business, ultimately selling its chip production business to GlobalFoundries in 2015.

The third generation, from Apple's iPhone launch in 2007 to present, is smartphones. The rapid development of the smartphone industry coincided with a new semiconductor industry division model—the rise of fabless design houses and pure-play foundries collaboration. Although TSMC, the first pure-play foundry, was established in 1987, foundries were long synonymous with cheap and low-end under the shadow of traditional chip giants like Intel and Texas Instruments, with an industry saying that "Real men have fabs." However, when Intel refused to supply chips for Apple's smartphones, Apple turned to foundries and launched its self-designed A-series chips in 2010. As smartphones became ubiquitous and the main driver of chip technology development, fabless chip design companies (including independent design companies like Qualcomm and MediaTek, and system companies' design divisions like Apple and HiSilicon) and foundries (including TSMC and Samsung Electronics) became industry leaders. By late 2024, industry-leading foundry TSMC's market value exceeded $1 trillion, while Intel's value was less than 1/10 of that and faced major difficulties in advanced process development.

Looking at the rise and fall of these three generations of product platforms, it's difficult to say that companies leading Moore's Law technologically can maintain perpetual success. Due to strong scale effects in chip manufacturing, companies that succeed in the product market are more likely to invest heavily in R&D and manufacturing, gradually gaining control over advanced processes, while the reverse isn't necessarily true. In fact, besides the main computer track, there are many "side tracks" in chip markets where Moore's Law's effects are weaker and product market forces are stronger. For example, electronic calculators, invented in the early 1960s, were one of the early major markets for integrated circuits. Japanese companies like Sanyo, Canon, and Sharp led Japan's electronics industry takeoff through miniaturization innovation, reducing calculator prices from initial $4,000 to less than $10 by the late 1970s. Another example is IBM's introduction of integrated circuit memory chips replacing core memory in 1970, launching the memory chip market. This market not only nurtured early Intel but also became a major track for later Japanese and Korean semiconductor industries. Furthermore, a notable characteristic of chip market platform transitions is "disruptive innovation," where new product platforms emerge as simpler, cheaper low-end products, eventually replacing mainstream and high-end products through continuous improvement. In this process, incumbent manufacturers' collapse isn't due to lack of technological advantage, but rather being locked into old platforms and unable to effectively respond to new demands. From Intel to TSMC, all rose from niche markets; yet once becoming dominant, like Intel a decade ago, even possessing the most advanced process technology couldn't help enter the mainstream mobile market, ultimately falling into difficulties—a "Moore's Law trap."

China's Winding Road to Advanced Chip Manufacturing

Global chip industry transformations reveal that developing the chip industry requires both following Moore's Law's technological push and market demand pull. Without market demand support, even leading companies struggle to maintain advanced process investments, making it harder for latecomers to catch up. While Chinese leaders have long emphasized chip industry development and catching up with international advanced technology, maintaining strong long-term investment in the chip industry, we must acknowledge that Chinese chip companies still significantly lag behind international frontiers in scale and technical level. China's difficult path in developing its chip industry is closely related to needing to solve both technical and market challenges simultaneously.

China was among the first countries to develop semiconductor chip technology. As early as 1956, the Central Committee called for a "March toward Science," and semiconductor technology was included in the twelve-year perspective plan. Under basically closed conditions, Chinese researchers developed the first silicon planar transistor in late 1963 and the first integrated circuit in 1965—just seven years after Texas Instruments invented the integrated circuit. By the late 1960s, China had independently built a complete integrated circuit industry chain to meet military and research needs. Although China's early semiconductor industry achievements under extremely difficult conditions were remarkable, we must acknowledge that by the pre-reform period, China's semiconductor industry had fallen far behind not only Western countries but also emerging Asian economies.

(1) State Investment Construction Period

After reform and opening up, China's semiconductor industry development experienced two phases divided by the year 2000. In the first phase, China's semiconductor industry development relied on government investment in key enterprises and the introduction of foreign advanced technology, with three projects being the most typical: Wuxi Factory 742, Project 908, and Project 909. Factory 742, short for State-owned Jiangnan Radio Materials Factory, was the largest and most profitable integrated circuit enterprise built in the early reform period. To meet strong domestic TV demand, China imported many TV production lines from Japan in the 1980s, and Factory 742 produced integrated circuits for TVs. From the State Planning Commission's project approval in 1978 to production verification in 1985, Factory 742 used 66 million USD in state investment to import a 3-inch production line from Japan's Toshiba for TV integrated circuits. After starting production, Factory 742 achieved annual integrated circuit production exceeding 30 million units, once reaching nearly 40% of national total production, greatly accelerating TV domestication and popularization. Factory 742 later cooperated with Germany's Siemens to build a 5-inch line and merged with local research institutes (No. 24) to establish Wuxi Huajing Electronics Group. Among semiconductor production lines imported by China in the 1980s, most were poorly utilized due to lack of industrialization experience and technical personnel, making Factory 742 a rare success. Although Factory 742 achieved good economic benefits, its acquired technology lagged far behind foreign countries. By this time, Western countries had entered the large-scale integrated circuit era, and neighboring Japan had reached the international competitive frontier through its 1975-1979 VLSI project.

In August 1990, the State Council approved Project 908, targeting large-scale integrated circuits through state investment in leading enterprise Huajing, building a 6-inch 0.8~1.2 micron integrated circuit production line. After negotiations, Project 908 purchased technology from AT&T Lucent. Lucent not only transferred process technology and trained engineers but also provided advanced communication IC design tools and databases for producing chips used in Lucent's programmable switches. However, Huajing's 6-inch line took over 7 years to complete and immediately faced difficulties upon completion. On one hand, 908's decision-making and negotiation period was lengthy; although it achieved its stated goals, international mainstream technology had advanced to 8-inch lines by then, failing to effectively narrow the technology gap between the domestic semiconductor industry and developed countries. Notably, 908's implementation time wasn't particularly long for its era; the exposed problems were more chronic issues of the existing planned system. On the other hand, Lucent's technology had significant limitations, requiring Huajing to develop secondary innovations based on the imported technology to expand market reach. However, Huajing lacked such capabilities, and with the home appliance market saturated, existing production lines were losing money, leaving no resources for further investment. Finally, Huajing partnered with Taiwan Mosel Electronics founder Chen Zhengyu to establish a joint venture, Huajing-CSMC Semiconductor, becoming mainland China's first pure-play foundry handling overseas processing.

By the mid-1990s, national leaders fully recognized 908's institutional problems and shifted focus to Project 909. Launched in 1995, Project 909 was jointly funded by the central government and Shanghai municipality, aiming to build an 8-inch integrated circuit production line, costing nearly 10 billion yuan—exceeding all previous state semiconductor industry investments combined. Huahong Microelectronics Company was established in 1996 as 909's main entity, with the Minister of Electronic Industry and Shanghai's Vice Mayor serving as key leaders to demonstrate the central government's attention and determination. By then, the U.S.-led Wassenaar Arrangement controlling conventional weapons and high-tech exports made obtaining advanced semiconductor manufacturing equipment and technology from abroad more difficult. Eventually, Japan's NEC agreed to join Project 909, forming a Huahong-NEC joint venture, transferring 8-inch, 0.35~0.5 micron technology. With strong central and Shanghai support, Project 909 built China's first 8-inch line in just 18 months. In February 1999, Huahong-NEC began producing 64M memory chips, reaching international mainstream product levels.

High specifications and investment helped Project 909 overcome initial obstacles, greatly enhancing domestic semiconductor technology and production capacity. However, by 2000, giants like Intel and TSMC had advanced international frontiers to 12-inch, 0.18-micron process levels, leaving Huahong-NEC 2-3 generations behind. This lag had two causes. First, post-Wassenaar Arrangement, Western companies were unwilling to cooperate with China, with the U.S. effectively establishing an "N-2 generation gap" rule for chip technology exports to China, requiring exported chip equipment and technology to lag international frontiers by over 2 generations. Although NEC had substantial investments in China, its memory chip technology wasn't at Moore's Law's forefront. Memory chips had simpler designs and slower process advancement than logic chips, with manufacturers mainly competing through economies of scale. Second, Huahong-NEC initially relied on NEC's technology and market, but the NEC-represented Japanese semiconductor industry was declining. With the internet bubble burst and memory chip prices plummeting, NEC's semiconductor division's difficulties began affecting Huahong-NEC. After 2002, Huahong-NEC introduced a new management team of overseas Chinese and returnees, successfully transforming into a foundry, but subsequently shifted focus to mature process specialty technology, no longer pursuing advanced processes.

(2) Market-oriented and Globalization Period

In 2000, the State Council issued the "Notice on Several Policies to Encourage the Development of Software and Integrated Circuit Industries" (hereafter "Document 18"), marking China's chip industry's entry into a new development phase. Document 18 opened industry access to private and foreign-invested enterprises, allowing all companies established in China to enjoy preferential policies regardless of ownership. This industry policy shift toward market-based incentives drove hundreds of companies to enter China's chip industry in the early 21st century. These emerging enterprises, comprising numerous fabless chip design companies and a few large pure-play foundries and packaging plants, established a vertical division system closely integrated with global supply chains.

SMIC represents the emerging chip manufacturers of this period. In April 2000, Richard Chang, a Taiwanese engineer from Texas Instruments who left Worldwide Semiconductor, founded SMIC. In May, Chang led hundreds of international engineers and managers to Shanghai, broke ground in August, and built a mega chip factory with 100,000 wafer monthly capacity in Shanghai's Zhangjiang within just one year—breaking the world record for fastest factory construction. Understanding the industry's scale-based competition, SMIC rapidly expanded in Beijing, Shanghai, and Tianjin through construction and acquisitions with local government support. SMIC began its first 12-inch line in 2002, and by its 2004 IPO, had become the world's fourth-largest pure-play foundry—the first mainland Chinese company to enter the global top five.

Early SMIC represented a fully globalization-based business model. Chang's startup team mainly comprised overseas Chinese, international experts, and returnees, while investors included state capital from Beijing and Shanghai, U.S. investment banks, Silicon Valley venture capital, multinational companies, and Singapore's Temasek. Leveraging the founder's deep overseas connections, SMIC obtained substantial equipment and technology from the U.S., becoming one of the first five companies approved for the U.S. Commerce Department's 2007 Validated End-User (VEU) indefinite export exemption program. Operationally, early SMIC followed a pure foundry model with "both ends overseas": importing equipment and materials, using foreign customers' chip designs, and exporting manufactured chips. This model's significant advantage was effectively utilizing expensive production lines and training local workers through complete global supply chain integration, gathering substantial demand, and leveraging enormous economies of scale to reduce costs. By 2009, SMIC's capacity represented half of China's integrated circuit capacity, and China's share of world integrated circuit production rose from less than 1% in 2000 to nearly 9%.

While globalization enabled SMIC's early success, it became its business model's Achilles' heel and failed to fundamentally change China's chip industry's catch-up position. With markets overseas, SMIC first faced competitive pressure abroad. Shortly after its establishment, SMIC became embroiled in intellectual property disputes with TSMC. From 2003 to 2009, TSMC filed multiple U.S. lawsuits alleging patent infringement and trade secret theft. These lawsuits ultimately cost SMIC over $450 million in economic and equity compensation for settlement and led to founder Chang's resignation. The founder's departure caused short-term management instability, slowing growth and technological catch-up in subsequent years. During this period, SMIC's process development was slow, with revenue mainly from mature processes above 90nm, actually widening the gap with industry leaders. Only after the National Integrated Circuit Industry Investment Fund's investment in 2015 and the recruitment of process expert Liang Mong Song, achieving 14nm node breakthrough, did SMIC return to rapid growth and large-scale investment, though by then SMIC had completely shifted toward domestic sources in funding structure, core management, and revenue.

As U.S.-China technology competition intensified, China's chip industry globalization era slowly ended. In December 2020, SMIC was placed on the U.S. Entity List; on October 7, 2022, the U.S. completely banned exports of technology and equipment for producing advanced computing chips to China, repeatedly upgrading restrictions over the next two years. Under U.S. sanctions, SMIC faced increasing difficulty obtaining overseas technology and equipment, and couldn't purchase ASML's EUV lithography machines to produce advanced process chips below 10nm. Although SMIC achieved equivalent 7nm "N+1" and "N+2" process nodes using older DUV lithography and double exposure technology, under escalating U.S. restrictions, SMIC's future advanced chip development still requires more fundamental breakthroughs.

Breaking Free from the "Moore's Law Trap"

Reviewing China's winding path in chip industry development, we find two threads leading to the current predicament. The visible technical thread is well-known: China's repeated cycle of technology import—lag behind—re-import in chip industry development, never catching up with Moore's Law's rhythm, and struggling to break foreign dependence on key technologies. In the post-globalization era, as technology imports become increasingly difficult, breaking this cycle becomes more urgent.

Less noticed is the market development thread supporting advanced technology progress. During the Cold War, the East-West division isolated China from the global division of labor, and low economic development levels couldn't nurture sufficient semiconductor market scale to support an independent advanced chip industry—this is understandable. Since the reform and opening up, strong consumer demand for home appliances like TVs, refrigerators, and washing machines has supported the domestic electronics industry, driving upstream integrated circuit demand, though these demands often didn't require advanced processes. As China established an export-oriented economy in the late 1990s, joining the Western-led international division of labor, China's chip industry reorganized according to modular production paradigms. While the chip industry's globalization model brought high investment and growth, it also caused a disconnection between chip investment and domestic demand. For example, around 2007, when SMIC received the U.S. Commerce Department's indefinite export exemption, up to 80% of its revenue came from overseas markets. The globalization-era Chinese chip industry not only failed to break foreign technology dependence but deepened overseas market dependence. Consequently, emerging domestic electronic industry system giants were unwilling to use domestic chips, while domestic chip manufacturers were content being outsourcing options for multinational companies' low-end chips, and nobody wanted to use domestic chip manufacturing equipment.

From both technical and market perspectives, in China's advanced chip industry development, finding demand support might be even more urgent than pure technical catch-up on the production side. In the early 21st century, many think tanks and consulting firms optimistically believed that China's rapidly rising semiconductor consumption would necessarily stimulate local production, driving Chinese semiconductor industry development. In reality, under global supply chain production models, although China became the world's largest semiconductor consumer in 2008, local production severely lagged long-term. Besides much of the statistical consumption actually being for assembly and re-export, there was probably very little effective demand in remaining domestic consumption that could support domestic advanced chip technology development and production. This situation only improved after the U.S.-China trade war erupted when the Trump administration attempted to sever ZTE and Huawei's chip supply chains to destroy them. SMIC's recent annual reports show its mainland China revenue ratio quickly jumped above 50% after 2018, reaching 80% by 2023. However, this transformation may have come too late and incomplete. To date, except for sanctioned Huawei producing Kirin processor chips using advanced domestic processes, there aren't many examples showing domestic manufacturers having sufficient interest in and demand for domestic advanced process chips.

One might ask why chip industry development can't reach world-advanced levels through export processing like countless other Chinese export-oriented industries. This is because competition in high-tech industries like chips essentially revolves around market control. The authors previously pointed out that from an economic perspective, the essence of modern high-tech competition is enterprises converting cost disadvantages into competitive advantages through market control under high fixed-cost strategic choices. The first part of this proposition indicates that high-tech innovation requires companies to dare make large, long-term investments in human, material, and financial resources. Such persistence requires treating "variable costs" like labor as fixed costs to maintain sufficient patience for technological and product success. Conversely, companies can choose not to invest in high fixed costs, merely adjusting "variable costs" through hiring and firing according to market signals, but such companies inevitably cannot create differentiated products and are destined for mediocrity. However, the latter part of the proposition further indicates that if companies want to innovate, they must be able to convert innovation outcomes into revenue and profit through obtaining, controlling, and even creating markets; otherwise, high fixed-cost strategies are impossible. As the source of core competitiveness in the contemporary electronic information industry, cutting-edge chips are the foundation for innovative enterprise giants supporting modern developed countries' industrial systems to implement high fixed-cost strategies, thus inevitably becoming the focus of market control competition. Since integrated circuits entered civilian markets, innovation-leading giants have held enormous market power over chips' end markets. This was true for IBM in the mainframe era and Intel in the PC era. TSMC might be an exception, but its position is an extension of Apple's main customer market power. This is precisely why TSMC is more vulnerable than previous-era giants, easily manipulated by the U.S. government. In such markets, the Chinese export industry's usual strategy of good quality at low prices will inevitably face blockades from giants and their backing governments.

In other words, developing a cutting-edge chip industry is impossible without sufficient market control power to fully absorb innovation's high fixed costs. China's market for developing advanced chip manufacturing can only come from its vast domestic market. Of course, this doesn't mean the government should grant monopolistic power to certain enterprises; in fact, only monopolistic capability gained through market competition represents true market power, and only such enterprises can truly convert the market position into innovation drive—this principle is self-evident in China's current market economy. However, we must clearly recognize that after thirty years of globalization, with multinational companies locking down most domestic advanced chip markets, expecting them to voluntarily surrender domestic market share is unrealistic. At appropriate times, using state power to open market space and create opportunities for innovative enterprises implementing high fixed-cost strategies is also a necessary choice. Developed countries have never hesitated on this point, and neither should we.

Moore's Law becomes a "trap" because it obscures complex conditions needed for industry development with a simple quantitative indicator, misleading government, and enterprise decision-makers to believe that merely investing in advanced process nodes will naturally lead to success. There have been numerous examples of falling into this trap in the past, and there will be more in the future. For China to further develop its advanced chip industry, it must break free from the myth of laws like Moore's Law, shift decision-makers attention away from introducing one or two pieces of equipment, and focus on the market and domestic demand environment supporting advanced industry development. For example, the currently intensifying "involution" leading to domestic demand collapse will severely damage industrial development; regarding this, the government must support enterprises implementing high fixed-cost innovation strategies. Only when enough enterprises implement high fixed-cost strategies can sufficient market and domestic demand be created, and larger markets will drive more enterprises to implement high fixed-cost strategies, ultimately leading to a sustainable path for high-tech development.