China’s AI Moment: Manufacturing, Global Values and the (New) End of History

"The competition between China and the US over the global market for large [AI] models is essentially a contest for dominance over global political ideology." – Di Dongsheng (翟东升)

Echoing our previous coverage of Professor Di Dongsheng’s analysis, this speech is characterised by a boosterish “China-first” style that stands out among some of the more pious rhetoric common in Chinese foreign policy publications. To a certain extent, it is refreshingly unusual to come across an international relations scholar in China producing statements such as “the battle for global industrial leadership is not a win-win scenario, but a zero-sum game.”

Professor Di aims to challenge conventional theories of industrial relocation, which describe the process of manufacturing migrating to less economically developed countries as production costs rise. In his telling, it is not just labour costs, but power—particularly maritime and technological power—which has historically determined where the centre of global manufacturing lies. Hence, rather than this centre inevitably shifting to countries with lower labour costs, China’s progress in automation and AI may be able to arrest this process—ensuring China’s pre-eminence in manufacturing for the foreseeable future. Underlying this argument is a conviction that above all else, manufacturing is the foundation of national strength. Presented within a sweeping macro-historical narrative of the past six centuries of industrial relocation, his argument strikingly resembles a theory of “the end of history”—albeit with manufacturing-inflected Chinese characteristics.

Beyond the contest to remain the world’s manufacturing powerhouse, Professor Di further contends that China’s strength in the realm of content creation and social media algorithms—particularly via Tiktok—could allow China to position itself at the centre of global information flows and thereby shape the dominant values and ideology of the 21^st century. However, the scenario he sketches mainly involves the apparently lucrative business of adapting popular Chinese short-video formats to familiar Western settings; whether or not such content can meaningfully convey distinctly Chinese values is not immediately clear. Moreover, his fellow Renmin University professor, Wang Wen, has previously highlighted obstacles to China achieving a level of dominance in global discourse flows comparable to its manufacturing might—namely, informational “protectionism” in the form of the Great Firewall (which Wang has suggested dismantling).

Professor Di makes a case for China’s “centre-left accelerationism” as an alternative to both Western liberalism and the “Dark Enlightenment” of the American tech-right. Yet while technology and pure heft may allow China to maintain its role as the world’s factory even as its population grows wealthier, becoming the legislator of global values could prove a far more delicate balancing act.

— James Farquharson

---

Wang Wen: Why China's Great Firewall Should Be Dismantled

Key Points

1. Traditional explanations of industrial relocation predict that global manufacturing will eventually shift from China to countries with lower labour costs, such as Vietnam and India—a process that has already partly begun.

2. As the Chinese government seeks to raise living standards and strengthen environmental protections—while demographic challenges remain unspoken—China faces significant upward pressure on labour and production costs.

3. However, China’s advances in artificial intelligence and factory automation could offset these effects, enabling it to resist the “gravitational pull” of relocation to lower-cost labour markets .

4. Whether or not China can overturn previous patterns of industrial relocation is key to determining if its rise to global prominence will be ephemeral or long-lasting.

5. This poses a more serious challenge than any threat of containment by the “American Empire”, whose power is already fading.

Work with Us – Project-Based, Flexible Commitment
Requirements: HSK6 or equivalent; native English (≠ near-native); strong writing and translation skills; and a solid background in China-related studies/work. Please send your resume to:

6. China may well achieve this. From a macro-historical perspective, the original industrial relocation from China to Europe during the 15^th century stemmed from Europe’s superior maritime power—not lower labour costs.

7. If China can maintain its hold on global manufacturing through AI and automation, it would mark the long-term return of manufacturing to its natural centre of gravity.

8. Another crucial front in the development of artificial intelligence and online algorithms is the struggle to control the values, ideologies and narratives embedded in global information flows—a contest China must win.

9. The contest over global ideology and values involves three camps:

   1. Left-wing neoliberalism – the declining ideology of the US’s Democratic Party;

   2. Right-wing accelerationism – the elitist “Dark Enlightenment” thinking embraced by Silicon Valley and tech entrepreneurs aligned with Trump, currently on the rise;

   3. Centre-left accelerationism – China’s “people-centred” techno-developmental ideology.

10. The algorithms underpinning Chinese social media platforms and artificial intelligence models will enable China to shape the values and ideologies of their users.

The Author

Name: Di Dongsheng (翟东升)
Year of birth: 1976 (age: 48/49)
Position: Deputy Dean and Professor, School of International Studies, Renmin University of China (RUC) (2017-now); Dean of the Institute of Regional and Country Studies, RUC; Deputy Director and Secretary-General, Centre for Foreign Strategy Research, RUC (2011-now)
Other: Frequent exchanges with officials at China’s National Development and Reform Commission (NDRC), Ministry of Foreign Affairs, Ministry of Industry and Information Technology, International Liaison Department etc.
Research focus: Global political economy of money and finance; US political economy; Chinese foreign policy
Education: BA, MA and PhD Renmin University of China (1994-2004)
Experience abroad (as a visiting scholar or lecturer): Sciences Po Paris, Durham University, Vrije Universiteit Brussel, Georgetown University

CAN ARTIFICIAL INTELLIGENCE BRING ABOUT THE END OF A CENTURIES-LONG HISTORY OF INDUSTRIAL RELOCATION? THIS WILL BE A MAJOR TEST OF ITS CAPABILITIES
Di Dongsheng (翟东升)
Published by Guancha.cn on 28 April 2025
Translated by James Farquharson
(Illustration by OpenAI’s DALL·E 3)

1. How Manufacturing Returned to China (after a 600-Year Hiatus)

The topic of “AI-Empowered Manufacturing” [Note: the topic of the conference at which Di Dongsheng was giving this speech¹] in my mind, relates to global industrial relocation [全球产业转移]. As is widely understood, the narrow sense of “global industrial relocation” refers to the pattern of industrial shifts that have been occurring since the Second World War. Yet from a broader perspective, the relocation of global production and trade centres has been a continuous [feature] throughout the history of industrialisation—one closely linked to shifts in maritime power.

From this perspective, the “global industrial relocation” that occurred after the Second World War is part of a six-hundred-year-long trajectory of shifts in global production centres. Historically, East Asia was originally the core of global industry, with the Song and Yuan dynasties being particularly prominent; the Ming dynasty also held this position, though the Qing dynasty was somewhat weaker in this regard. East Asia became the centre of global industry because its people were industrious and skilled at producing scarce goods such as silk and porcelain.

However, the Ming dynasty altered the course of history when it abandoned its policy of large-scale maritime navigation. [Note: famously, in the decades after the maritime expeditions of Zheng He under the Ming Emperor Yongle, maritime exploration was officially abandoned.] [As the saying goes,] “when one relinquishes heaven-sent opportunities, negative consequences will follow” [“天予不取，反受其咎”]: it was precisely during this period that Europe rose to pre-eminence through maritime dominance.

In the early days, maritime navigation in various countries was what might be termed “small-scale navigation”. For instance, Chinese merchants and fleets operated primarily within the Pacific Ocean, the South China Sea and the Indian Ocean. The Europeans first developed the Mediterranean routes, giving rise to maritime powers such as Genoa and Venice.

Europe then entered the Age of Exploration, during which the centre of maritime power gradually shifted to Portugal and the Iberian Peninsula, where a new global hub of trade and production was established. The Dutch later wrested control from them [抢走了他们的江山], followed by the arrival of the British on the scene. By the end of the 19th century, the global centre of production and trade had shifted to the east coast of the United States. Several decades later, on the eve of the Second World War, it shifted once again—this time to the west coast.

After the Second World War, the global centre of production and trade began shifting towards Japan. However, given Japan’s limited scale at the time, they lacked the capacity to fully accommodate this centre, which soon splintered out to its former colonies, namely the Four Asian Tigers. But after 1992, under the leadership of Comrade Deng Xiaoping, China succeeded in absorbing this production centre—accumulated over five to six centuries and rooted in maritime power—into the Yangtze River Delta and Pearl River Delta regions.

The history of the shifting centre of manufacturing is, in essence, the story of China voluntarily relinquishing its maritime power, while European civilisation seized the opportunities presented by the age of maritime dominance to establish its own centres of production and trade, which shifted from one European country to another. After the Second World War, this centre moved to East Asia, eventually settling in the southeast of China [最后落户中国东南].

In recent decades, [scholars from] Japan proposed the “Flying Geese Paradigm”, [Note: a model by which the production of manufactured goods moves over time from a more economically advanced country—the “lead goose”—to less economically advanced countries—“follower geese”] while our country’s Taiwanese entrepreneurs have put forward the “Smile Curve” theory [Note: a theory which posits the design and marketing of a product as two elevated sides of a value curve, with manufacturing sitting in the less lucrative trough]. Using different perspectives, these both offer explanations of the industrial relocation process. However, overall, there is a broadly accepted assumption: when the prices of local production factors rise sharply, then the centre of industrial production will shift elsewhere.

2. …and How Manufacturing Could Stay in China (with the Help of AI)

Hence, I would like to raise a question. With the growing maturity and development of artificial intelligence, could China actually manage to bring an end to this process of industrial relocation? Since ancient times, the realisation of industrial automation has been a long-cherished dream of humanity—human nature inherently craves greater enjoyment and reduced labour. It is precisely the pursuit of “idleness” that has served as one of the driving forces behind technological progress.

At present, so-called “dark factories” [“黑灯工厂”] have already emerged in regions such as Jiangsu, Zhejiang, Shenzhen and Dongguan. These are factories capable of continuous, twenty-four-hour production, requiring only a handful of technicians to monitor operational metrics from air-conditioned control rooms, with no need to enter the workshop for hands-on tasks. In such a context, is it still the case that future industrial production will follow the historical pattern of industries automatically relocating [to a new site] once labour costs in the original production site reach 5-10 times that of the [new] destination?

We all know that since 2012, some manufacturing that was originally based in China has begun relocating to countries such as Vietnam and India—a trend that has only accelerated since [Trump’s] first trade war [with us]. Naturally, the main entities withdrawing from China are not Chinese-owned enterprises, but rather transnational industrial chains—such as those of Apple and Amazon—as well as capital from Taiwan, South Korea, and Singapore. Therefore, a question of very real significance to us is whether artificial intelligence can bring about the end of the 600-year history of industrial relocation.

I believe this very [issue] is an essential component of the great rejuvenation of the Chinese nation. Rejuvenation itself is not difficult, for we have already returned to [our role as] the global centre; the true challenge of [our] great rejuvenation lies in [the wisdom that] “many fine beginnings are wasted before completion” [靡不有初，鲜克有终]. It has taken us a century to arrive at this moment of resurgence, and the crucial question now is whether this revival will prove to be a momentary blossoming [昙花一现] or a perpetual bloom [常葆青春].

Hence, as a scholar of international relations, I believe that relevant research must address two principal issues in tandem:

1. How to overcome the containment imposed on us by American imperialism;

2. How to resist the “gravitational pull” of industrial relocation [to less developed countries], ensuring that the next generation inherits our achievements—and that the dreams of countries such as India to absorb China’s relocating industries ultimately come to naught.

In my view, the former [challenge] is comparatively simple—the American empire is already showing signs of decline, and its “heavenly mandate” [天命] is approaching its end. The latter, however, presents a far greater challenge. The struggle for control over the world’s industrial centre is not a win-win scenario [并非一个双赢], but a zero-sum game [而是一场零和博弈]: whoever holds the core of manufacturing stands at the centre of the global stage.

If we can rely on artificial intelligence to break the entrenched pattern of “upgrading manufacturing while relocating industries to lower costs”, we will be able bring the [historical] process of industrial relocation to an end.

One [of our priorities] is that we must improve the living standards of our people, which inevitably entails rising costs for various factors of production. Another is our push to establish “green waters and green mountains”, [Note: 青山绿水 qingshan lüshui, referring to pro-environment development goals], [which involves] tightening environmental protection requirements. This year’s government work report has likewise proposed the large-scale establishment of a “people-centred” economy [“投资于人”的经济]—which will necessarily require a greater allocation of resources towards nurturing and supporting talent as well as towards advancing technology. [This shift] will reduce our [effective] fiscal subsidies for global consumers and capital, and in turn, implies higher taxation and increased expenditure on public services.

At a time when costs are rising across the board, we must also resist the natural gravitational pull of industrial relocation. This is the major strategic issue that must be placed front and centre when discussing the significance of artificial intelligence for the manufacturing sector.

3. AI, Algorithms and the Struggle over Global Ideology

The second issue concerns our political security and ideology. I had originally intended to invite a friend to speak [here]—he once served as the head of Midjourney’s China division, the world’s most powerful enterprise in the field of large graphical models (LGMs)—but as his current company is about to go public, he is unable to be present today.

In our conversation, two remarks he made left a particularly deep impression on me: “The foundation of large language models is values; the foundation of graphical artificial intelligence is aesthetics.” To put it simply, when you interact with a large language model and ask it questions about good and evil, it will tell you what is good and what is bad, thereby exerting an influence over your values. When you engage with a graphical large model and input the keyword “beautiful”, it will make choices among elements such as “high nose bridge” or “small eyes”, and its output will, in a subtle and unconscious manner, shape the recipient’s sense of aesthetics. Therefore, once we hold global leadership over [AI] large models, we will, in effect, hold the most sway over global values and aesthetics [掌握了全球价值观与审美的主导权].

But today, I would like to add a third point: the foundation of large video models is narrative and discursive power [叙事和话语权]. With the help of AI, we can now generate images of Middle Eastern, South Asian, or any other ethnic group at the click of a button, yet the story patterns conveyed in these videos remain largely the same. Last night, in conversation with an entrepreneur working in the field of “cultural exports” [文化出海], he confided: “Professor Di, what I am doing right now is promoting all those Chinese short videos formats on overseas platforms—such as ‘Billionaire CEO Falls in Love with Me, Post-Menopause!’ [Note: 千亿霸总爱上绝经的我, a highly popular short-video genre on Chinese social media, typically following a similar narrative logic in which couples find true companionship over class boundaries; the genders are sometimes reversed].”

To this end, he has hired a team of highly paid writers from Europe and America. At the same time, drawing on AI-generated tests that identify the narrative cores most likely to appeal to middle-aged women aged 40 to 50, he instructs these writers to shift the settings of such short videos from Shenzhen to Los Angeles, New York or Paris, mass-producing content for export. Naturally, they can also substitute Chinese “fox spirit” [狐妖] legends with vampires, werewolves and similar themes before reproducing [such content abroad]. This model for exporting short videos not only guarantees steady profits [稳赚不赔] but can even yield windfall gains—effectively harvesting a vast number of Western “leeks”. [Note: 韭菜, or “leeks”, a term roughly equivalent to “suckers”—often refers to stock market losers who generate profit for more experienced traders, but more broadly applicable to any easily monetisable demographic.]

These types of videos, by taking control over the discursive power, narrative structure and the logical framework of such storytelling, effectively place the target audience under a form of “hypnosis”. Over time, viewers become addicted and unable to resist—and this very same logic can be extended to large video models.

In this sense, whoever commands the dominant position in large language models, large graphical models and large video models—and whoever controls the relevant data, algorithms and computing power—will command ideological dominance over the twenty-first century. After all, what we call ideology is the shaping of people’s perception through values, aesthetics and narrative structures. Once mass perception has been shaped, you acquire “legitimacy”—while all voices that differ from yours are branded as “illegitimate”; you become a “god” [神] and others are cast as “demons” [鬼].

So, from this perspective, we can understand why Trump has recently kept contradicting himself [不断打脸] and kept defending TikTok. TikTok is simply too important to him. In Trump’s confrontation with the American political establishment—particularly with the Democratic Party’s “liberal democratic narrative”—only TikTok had the capacity to disrupt the mainstream ideological propaganda machinery, which is overwhelmingly controlled by the Democrats. This is due to the relative independence of TikTok’s algorithm.

Therefore, the competition between China and the United States over the global market for large [AI] models is essentially a contest for dominance over global political ideology. For this reason, the issue of control over artificial intelligence must be elevated to a matter of survival for the Party and the nation.

4. AI’s Revolution in Governance

Artificial intelligence will also bring about a revolutionary impact on national governance. Earlier in our discussion, we touched upon the issue of anti-corruption. In future, combating corruption will become remarkably simple: by implementing real-name systems for all transactions and purchases, we can use big data to screen out individuals with a high probability of engaging in corrupt practices.

At the same time, with the support of big data and artificial intelligence, not only can [our] anti-corruption [processes] be streamlined, but governmental operations may also undergo “flattening” [扁平化]. Civil servants at all levels will, in accordance with their respective responsibilities, be able to use governance-oriented large [artificial intelligence] models to directly grasp the intentions, work plans and guiding directives contained in documents issued by the highest leadership. This will enable the rapid implementation of leadership directives at the grassroots level. In this way, the age-old power contest between imperial authority, ministerial power and the bureaucratic system may well be brought to an end with the advent of artificial intelligence.

5. The Ideology of the US “Tech-right” and the Dark Enlightenment

Finally, I would like to address one point. In his speech, Professor Huang Jing mentioned three individuals: Curtis Yarvin, Peter Thiel and Elon Musk. The reason these figures feel confident enough to advocate the theory of “right-wing accelerationism” [右翼加速主义] is precisely due to the rapid advances in artificial intelligence in recent years. Consequently, right-wing accelerationism has become a key distinction between Trump’s first and second terms in office.

The central conflict during Trump’s first term unfolded between the MAGA faction and the so-called “Deep State” [深层政府] faction hidden within his government, which often resulted in his policies appearing to be a struggle between left and right [左右互搏]. However, in his second term, aside from a few cabinet members associated with the establishment—such as the Secretary of State—the Republican establishment faction within Trump’s cabinet had been almost entirely purged. Precisely because of this, the current Secretary of State still remains relatively obscure to the broader public. One could say that the “tech right” on the US West Coast is striving to take full control of Trump’s second term.

They have invested vast sums in this cause—Elon Musk personally donated $300 million, and together with contributions from the West Coast tech community and the cryptocurrency sector, this enabled Trump’s return to the White House. Since his inauguration, the power struggles within Trump’s cabinet have become increasingly intriguing [耐人寻味].

When discussing mainstream American ideology, I agree with Professor Huang Jing’s reference to the concept of “accelerationism”. However, it’s important to clarify that here it refers specifically to “right-wing accelerationism” or “far-right accelerationism”.

The intellectual origin of this “right-wing accelerationism” is rather dramatic. It stems from Nick Land, a troubled and mentally unstable young philosophy lecturer at the University of Warwick in the 1990s. His ideas captured the attention of Curtis Yarvin, Peter Thiel, Elon Musk and other tech entrepreneurs and capitalists on the US West Coast. On social media platforms such as Facebook and Twitter, they often add a suffix to their usernames indicating they are “accelerationists”. Anyone bearing this suffix is a believer in right-wing accelerationism.

This group is extremely radical and proudly calls itself the “Dark Enlightenment”. They believe that the march of progressivism, driven by left-wing politics over the past century, has not improved humanity but instead pushed us toward extinction, regression and ignorance. This “left-wing acceleration” is seen as pandering to or appeasing the rabble and ignorant masses.

Therefore, they believe that the correct course of historical progress should prioritise the development of the elites: those who are the most intelligent, capable, daring and dominant. Empowered by artificial intelligence and brain–computer interfaces they would become the new “Übermensch” [超人]. The Übermensch ought to have many children, signing multiple prenuptial agreements in order to produce dozens—perhaps more than twenty—descendants. This, [they argue,] would ensure their family maintains perpetual control over capital, technology, intellectual superiority and, most crucially, artificial intelligence. [In this way,] they would rule the world.

This is the “Dark Enlightenment”—the prevailing ideology upheld by one of the most powerful groups in today’s world.

6. China’s “Centre-Left Accelerationism”

By contrast, China also emphasises the importance of accelerating technological progress, but our “accelerationism” leans towards the centre-left [中左翼]. We advocate for capital to develop in a morally-guided way [向善发展]: if capital seeks only to snatch the last coin from people’s savings jars [把老百姓盆盆罐罐里最后一个钢镚都夺走], then it must be admonished [敲打]. We are not opposed to capital; rather, we remind it that its growth must balance profits with ethics [兼顾收益和伦理] and consider political equity alongside economic efficiency [兼顾政治公平和效益], instead of pursuing profit alone.

Precisely for this reason, some describe China’s practice as “left-wing accelerationism”— [one that is] “people-centred,” [advocates] “technology for good”, and [promotes] “technology serving the people”.

We are now entering the mid-21st century. Looking back to the period from 1925 to 1930, the world then witnessed a three-way ideological contest: the declining liberalism represented by Britain; National Socialism (or far-right fascism), as represented by Nazi Germany; and left-wing Marxist-Leninism, as represented by the Soviet Union.

The contest among these three systems of governance and major forces ultimately led liberalism and communism to join forces in defeating far-right National Socialism, culminating in the world war of the 1940s.This conflict largely shaped the global order [that has endured] from the Second World War to the present day.

At this very moment, a new “trinity of ideologies” is brewing globally. First, there is the neoliberalism represented by the US Democratic Party, which has clearly declined. One can sense that the US Democratic Party, a Europeanist [欧洲主义] party, is trembling [瑟瑟发抖], having sensed its abandonment by the common people and the masses.

Then, there are the two forms of accelerationism. One is “right-wing accelerationism”, which centres on the elite. It places its faith in technology and freedom, while rejecting laws, institutions, ethics and traditional historical narratives in the pursuit of radically transforming human history. The other is “centre-left acceleration”— China’s people-centred, care-driven approach to technological development, which I prefer to describe as “people-rooted ideology” [民本主义].

Standing at this new historical juncture, we are now witnessing great changes not seen in a century [百年未有之大变局], unfolding amid a wave of artificial intelligence. Let us together look forward to the great rejuvenation of the Chinese nation and the further development of this profound transformation.

Thank you all!

READ MORE

1. The conference, entitled “AI-Empowered Manufacturing: International Governance Lessons and Industrial Security”, was jointly organised by Renmin University of China and the Chinese Centre for Information Industry Development (CCID), a research centre under the Ministry of Industry and Information Technology (MIIT). The concept was based on a CCID report titled “AI-Empowered New Industrialisation: Paradigm Shifts and Development Pathways” [《人工智能赋能新型工业化：范式变革与发展路径》].