ESSAY: Why China No Longer Needs Nvidia — Technological Autonomy as a Doctrine of Power

ESSAY: Why China No Longer Needs Nvidia — Technological Autonomy as a Doctrine of Power

Introduction: The Day Beijing Turned the Page on Western Silicon

A $295-Billion Plan to Rewrite History

In June 2026, China made official what many Western analysts still refused to accept: Beijing no longer needs Nvidia, AMD, or any other American supplier of artificial intelligence chips. The five-year plan of 2 trillion yuan — approximately $295 billion at current exchange rates — to build a national AI data center grid by 2028 is far more than an infrastructure program. It is a declaration of technological independence, formulated in concrete numbers, regulatory mandates, and clearly designated national champions.

The core rule is unambiguous: at least 80% of the AI accelerator chips that will power this colossal infrastructure must come from domestic suppliers. This requirement, if enforced with the bureaucratic rigor that Beijing knows how to deploy when it chooses to, effectively closes off Nvidia and AMD from the largest AI infrastructure project in human history. The remaining 20% can theoretically come from international sources, but the signal is clear: the Chinese market is closing.

The Domestic Ecosystem: Huawei, Alibaba, Biren

In May 2026, Beijing formally approved nine categories of AI chips developed in China for deployment in government and security-sensitive sectors. This approval officially cleared Huawei's Ascend series, Alibaba's Hanguang 800, Biren Technology's BR100 and BR104, and Moore Threads' MTT S80 for public procurement. These names, still largely unfamiliar to the Western public, are the future dominant players in the largest AI market on the planet.

Huawei's Ascend 910B, produced on SMIC's 7nm process, is today China's most advanced domestic AI chip. Its performance remains below that of Nvidia's H100 in raw speed, but the gap is closing at a pace that industry experts describe as alarming for American players. Biren Technology and Moore Threads are advancing in more specialized segments, while Alibaba Cloud integrates its own chips into its cloud service offerings, creating a vertical ecosystem that is difficult to challenge from the outside.

The Doctrine of Technological Autonomy: The Story of a Silent Revolution

Decades of Deliberately Constructed Dependence

To understand the full scope of the 2026 pivot, we need to look back. During the 2000s and 2010s, China deliberately adopted a strategy of calculated dependence on Western technologies. Buying Intel chips, Dell servers, Microsoft software — this strategy allowed rapid economic development while accumulating technological capital through absorption, reverse engineering, and forced partnerships. Dependence was a choice, not a constraint.

Everything changed in 2022 with American export controls. Washington's decision to block advanced chip sales to China — first the A100, then Nvidia's H100 — acted as an electric shock. Within months, Beijing transformed what had been perceived as an inconvenient dependency into a national existential emergency. Budgets allocated to semiconductor R&D exploded. Diaspora talent was recalled. Companies like Huawei, which had been targeted by American sanctions, found themselves elevated to the status of national champions.

2026: The Crystallization of a Strategy

The $295-billion plan did not emerge from nowhere. It is the culmination of a coherent industrial policy built over several years. As early as 2024, China had extended its zero-tariff policy on domestic chips to 100% of tariff lines to encourage local adoption. State subsidies channeled toward SMIC, Huawei, and a constellation of semiconductor startups exceeded $150 billion over the 2020–2026 period, according to various estimates.

What makes the 2026 plan particularly significant is its systemic character. This is not simply about subsidizing companies or training engineers. It is about creating a guaranteed captive market large enough for the Chinese semiconductor industry to reach the economies of scale required for global competitiveness. A $295-billion market, with an 80% quota for domestic suppliers, represents an anchor of demand that few Western nations can match.

The Implications for Western Technological Hegemony

Nvidia, the Big Loser in a Battle That Has Shifted Terrain

Nvidia dominated the AI chip market with a market share of approximately 80% at its peak in 2024–2025. China represented, before export controls, roughly 20 to 25% of the company's revenue. Losing access to the Chinese market entirely is not an immediate catastrophe for Nvidia — global demand is strong enough to compensate in the short term. But losing the market of the largest AI infrastructure deployment in history, the very one that will shape the technological standards of the next decade, is a strategic loss of an entirely different nature.

The technical standards that Beijing will adopt — interfaces, communication protocols, system architectures — will be built around Huawei, Biren, and Alibaba chips, not around Nvidia GPUs. When these technologies spread to the rest of the world through Chinese infrastructure exports — notably across Africa, Southeast Asia, and parts of Latin America — they will carry with them a software and hardware ecosystem incompatible with Western standards. This is what strategists call technological fragmentation, and its geopolitical impact is difficult to overestimate.

Europe Facing a Bifurcation It Did Not Choose

For Europe, the situation is particularly uncomfortable. The continent has no national champion in AI chips — ASML dominates lithography equipment, but there is no European equivalent of Nvidia or Huawei Ascend. At the same time, Europe is caught between a Trumpian America that threatens its allies with 100% tariffs at the slightest commercial disagreement, and a China that is quietly building a credible technological alternative.

The European AI Act and the GDPR regulatory framework have established ambitious governance standards, but they do not produce chips. The European Chips Act initiative, funded with €43 billion — a fraction of what Beijing spends in a single year — is hampered by unavoidable industrial timelines and the difficulties of coordinating 27 member states whose interests do not always align. Meanwhile, China is building a national grid. This is an asymmetry of speed as much as an asymmetry of resources.

The 20% Opening: Strategic Opportunity or Decoy?

The International Margin as Diplomatic Leverage

The 20% left to international suppliers in the Chinese national AI grid is not inconsequential. This margin represents, on a $295-billion market, a potential of approximately $59 billion in international orders over five years. This is enough to maintain commercial relationships with European, Japanese, or even American players operating in segments not covered by export restrictions.

But this opening must be read through a geopolitical lens. Beijing keeps that 20% as a bargaining chip. Reducing or eliminating this share represents immediate economic pressure on Western companies that have built supply chains integrating the Chinese market. It is a soft weapon, an implicit threat: maintain reasonable commercial relations with us, or we shut the tap entirely.

The ASML Equipment Question: The Gordian Knot

The Achilles' heel of the Chinese strategy remains EUV lithography. ASML, the Dutch company that manufactures the machines essential for producing chips smaller than 5nm, is subject to export restrictions that prevent the sale of its most advanced equipment to China. SMIC, China's leading semiconductor manufacturer, is constrained to 7nm or less advanced processes, which limits its performance compared to the 3nm and 2nm that TSMC and Samsung have mastered.

This constraint is real and significant. But it is not insurmountable over the long term. China is investing heavily in developing its own EUV lithography — a technology that ASML took thirty years to perfect, and that Beijing hopes to master in ten. Impossible? Someone once told Huawei it could never build a premium smartphone. Huawei is today the top smartphone manufacturer in China.

The Geopolitical Dimension: Technology and Sovereignty

AI as the New Atomic Bomb

In the great-power competition of the 21st century, artificial intelligence plays the role that the atomic bomb played in Cold War competition: it is the ultimate weapon, the one whose possession or absence determines the balance of terror. Beijing understood this before most Western leaders. Xi Jinping's speech on AI since 2017 — "China must be a global leader in AI by 2030" — was not empty rhetoric. It was a state objective with a budget, a timeline, and designated executors.

The $295-billion plan fits this logic. A national AI data center grid is not merely economic infrastructure — it is an infrastructure of power. Facial recognition systems, mass surveillance, military planning, foreign intelligence analysis, information warfare: all of this requires enormous computing power. By building this grid on domestic chips, Beijing eliminates the strategic vulnerability of a technological dependence that Washington could exploit in the event of conflict.

The Chinese Digital Governance Model as an Export Product

Beyond the chips themselves, what is at stake is the export of a digital governance model. Countries of the Global South that adopt Chinese AI infrastructure — cables, servers, software, platforms — will adopt along with it the norms of surveillance, data control, and digital regulation that Beijing has developed. This is not a minor matter: it is the propagation of a worldview in which the state controls the flow of information, privacy is subordinate to national security, and technology companies are instruments of foreign policy.

The West has fought — rightly — to impose its values in international law, in trade, in human rights. It is losing this battle in the digital domain, not because its values are wrong, but because it has failed to provide a sufficiently attractive, accessible, and financially competitive alternative to nations that are simply looking to develop. China offers cheaper servers, submarine cables, easy financing, and no press conferences about democracy.

The United States Confronting the Paradox of Its Own Restrictions

Export Controls: A Double-Edged Weapon

The American policy of export controls on advanced chips, initiated under Biden and maintained under Trump, was intended to slow China's technological advance. To some extent, it did — by restricting access to H100 and H200 GPUs, Washington created real delays in certain AI research domains. But it also triggered exactly the scenario that the most cautious strategists feared: a massive acceleration of Chinese investment in technological autonomy.

China might never have invested $295 billion in a national grid with an 80% domestic chip quota had Washington not proven to it that technological dependence was a lethal vulnerability. In this sense, American export controls have ironically accelerated the very transition they were meant to prevent.

The Lobbying of American Technology Companies

There is a deep tension within the American technology establishment itself. Companies like Nvidia, Qualcomm, and Intel have argued to Washington that export restrictions hurt their revenues without actually stopping China's rise — China is developing alternative chips while losing access to the best ones. This debate, conducted largely behind closed doors in Washington, reflects a fundamental contradiction: how do you maintain a technological lead while continuing to finance your own competition?

Under the Trump administration, commercial logic and security logic clash constantly. Nvidia's Jensen Huang was among the American CEOs who accompanied Trump during his Beijing summit in May 2026. The image of this convoy of technology executives in the Chinese capital, in the midst of the $295-billion plan's implementation, says something about the complexity of the interests at stake.

The Talent Race: The Human Capital Behind the Chips

Engineers as the Secret Weapon

Behind every chip, there are engineers. And behind the engineers at Huawei Ascend or Biren Technology, there are often degrees from MIT, Stanford, Caltech, or leading European technical universities. China has invested heavily for twenty years in sending students abroad to the best electrical engineering and computer science programs, with the stated objective — rarely challenged head-on — of seeing them return to build the national industry.

This strategy is paying off. Companies like Biren Technology and Moore Threads are founded or led by engineers who previously worked at Nvidia, AMD, or Intel. This is knowledge transfer at a scale that export controls cannot regulate — because you cannot place an embargo on human minds.

The American Response: The CHIPS Act and Its Limits

The American CHIPS Act, passed in 2022, represented an investment of $52 billion in domestic semiconductor manufacturing. A significant sum, but one that pales against the hundreds of billions that China has mobilized through its various funds, subsidies, and industrial plans. TSMC and Samsung are building factories in the United States, but the timelines are long, production costs are higher, and skilled workers are harder to find than in Asia.

The hard truth that the American political establishment struggles to accept is that the CHIPS Act alone cannot guarantee American supremacy in semiconductors. It can create minimum resilience — ensuring the United States does not depend entirely on TSMC in the event of a crisis in Taiwan. But catching up with the productivity and scale of the Asian ecosystem — and countering China's rise — requires a national effort of a scope that American politics, divided and short-term in its thinking, struggles to sustain.

Washington's Allies: Taiwan, South Korea, Japan

TSMC: The Island on Which the Western World Rests

TSMC — Taiwan Semiconductor Manufacturing Company — is the most strategically important company in the world that most people have never heard of. With a market share exceeding 90% in advanced chips below 5nm, the Taiwanese company is the hub of the entire global digital economy. Apple's iPhones, Nvidia GPUs, AMD processors, the chips that power the data centers of Amazon, Microsoft, and Google: all of them come out of TSMC's factories in Taiwan.

This concentration is an existential vulnerability. If Beijing decided to act militarily against Taiwan — a scenario that a growing number of analysts consider possible within the 2030 horizon — the global economy would be plunged into an unprecedented semiconductor supply crisis. This is precisely why the United States pushed TSMC to build factories in Arizona, and why Europe is seeking to attract its own suppliers. But the urgency of these efforts says something about the fragility of the existing model.

South Korea and Japan: Caught Between Two Fires

South Korea, with Samsung and SK Hynix, and Japan, with its renaissance in semiconductor equipment, are leading strategic allies in the technology competition. But they are also economies deeply interconnected with China — the first or second trading partner of each of them. The pressure that Washington exerts to push them toward stricter export restrictions on China creates tensions that diplomacy must manage carefully.

Xi Jinping's summit with new Bangladeshi prime minister Tarique Rahman in June 2026, as part of a series of visits by more than a dozen world leaders to Beijing that year, illustrates the systematic charm offensive that Beijing conducts toward middle powers. China offers alternatives: cheaper AI infrastructure, more flexible financing terms, and discretion about domestic governance choices.

The Real Stakes: The Technology Standards of Tomorrow

Whoever Defines the Standards Defines the World

In technology competition, standards are as important as the products themselves. Communication protocols, hardware-software interfaces, data formats, system architectures: these constitute the invisible infrastructure on which the digital economy rests. Those who set these standards exercise considerable economic and political power over all who adopt them.

China's $295-billion national AI grid will create de facto standards. Developers who build applications on this infrastructure will use the frameworks, APIs, and tools provided by Huawei, Alibaba, and their partners. These standards will then be exported with Chinese infrastructure to third countries. This is the same phenomenon observed with Android: whoever controls the platform controls the ecosystem.

China's Drive in International Standardization Bodies

China is not content to create national standards. It systematically pushes its norms into international bodies — ISO, ITU, IEEE. The number of Chinese proposals in these bodies has grown exponentially over the past ten years. The United States and Europe participate in these discussions, but with less aggressiveness and coordination than Beijing.

For the West, the answer cannot be a defensive withdrawal into maintaining two incompatible technological silos — one Western, one Chinese — and hoping that the world's nations choose "the right one." Most will choose economic practicality. That means the West must offer a real alternative, not merely a rhetoric of values.

The Fragmentation of the Internet: Toward Two Parallel Networks

The "Splinternet": From Hypothesis to Reality

Five years ago, the idea of a fundamentally fragmented internet split into two distinct spheres — one Western, one Chinese — was still largely theoretical. In 2026, it is becoming an operational reality. China has its own ecosystem of platforms (WeChat, Baidu, Douyin, Alibaba), its own security protocols, its own cloud infrastructure. The national AI infrastructure under construction will only deepen this separation.

What the technical literature calls the "splinternet" is not merely a technical phenomenon. It is a geopolitical rupture. Companies operating in both spheres — and there are many — will have to maintain two distinct technological ecosystems, with the duplication costs and compliance risks that implies. For third countries, the choice of infrastructure increasingly dictates the geopolitical camp in which one finds oneself.

Submarine Cables and the Geopolitics of Physical Infrastructure

The technology competition is not fought only in chips and software. It is also fought in the concrete and steel of submarine cables, land stations, and data centers. China has invested massively in deploying physical infrastructure across Asia, Africa, and Latin America through the Belt and Road Initiative. These cables, towers, and data centers are not neutral — they create dependencies and chokepoints.

Europe and the United States belatedly recognized the stakes of this physical infrastructure. Projects like Amazon's Project Connect, Google's investments in submarine cables, or the EU's connectivity initiatives seek to offer alternatives. But they often start with a head start deficit and fewer resources than China.

The Western Response: Between Urgency and Paralysis

A Transatlantic Consensus That Is Slow to Form

Faced with China's rising technological power, the Western response suffers from a chronic lack of coordination. The United States and Europe share an analysis of the problem — China represents an unprecedented technological and geopolitical challenge — but diverge on the tools and the method. Washington, under Trump, favors tariffs, export controls, and bilateral pressure. Brussels prefers regulation, standards, and multilateralism.

These two approaches are not necessarily incompatible, but they require coordination that transatlantic trade frictions — over digital taxes, steel tariffs, industrial subsidies — make increasingly difficult. While the West disputes the terms of its own trade agreement, Beijing is building accomplished facts worth $295 billion per project.

What the West Must Do — and How Much Time It Has Left

The West's strategic response must be organized around three pillars. First, a massive and coordinated investment in semiconductor research, with clear objectives, dedicated funds, and a bureaucracy capable of disbursing them rapidly — not in five years after impact assessments. Second, a technology immigration policy that attracts and retains the world's best talent, including engineers trained at the world's best universities, regardless of their country of origin. Third, a proactive export strategy for Western digital infrastructure to third countries — not just speeches about values, but concrete, financially competitive offers.

Time is pressing. The window in which the West can maintain a significant technological lead over China narrows with every passing year. The $295-billion plan will not be the last. Beijing thinks in decades; Washington thinks in electoral cycles. That may be the real technological gap to bridge.

The Military Dimensions of Technological Autonomy

AI in Service of Beijing's War Machine

The stakes of AI technological autonomy are not only economic. They are profoundly military. Command and control systems, automated reconnaissance, cyber operations, and autonomous weapons systems all require massive computing power. China's national AI grid, with its $295 billion, is not merely civilian infrastructure — it is military capability disguised as industrial policy.

The Chinese military — the PLA — is deeply integrated into the national technology ecosystem. Companies like Huawei have documented ties to the PLA, which is precisely one of the reasons Washington placed Huawei on its blacklist. But banning Huawei in the United States and Europe does not change the fact that Huawei is building the backbone of Chinese AI infrastructure, and that this infrastructure will have military applications.

Ukraine as a Laboratory for Western Military AI

While China builds its AI infrastructure with domestic chips, Ukraine is testing in real conditions the military applications of AI powered by Western chips. Object recognition systems for drones, logistics planning algorithms, intelligence processing tools: the Ukrainian conflict has become an unprecedented testing ground for military AI technologies.

This battlefield reality has a direct implication for the technology competition with China: the West is accumulating combat data, operational experience, and lessons that can only be obtained in the reality of a conflict. But capitalizing on these lessons requires a speed of integration and adaptation that Western military bureaucracies have traditionally struggled to maintain. This is a competitive advantage that can be squandered for lack of structures capable of converting it rapidly.

The Future of Dependence: Toward a Two-Speed Technological World

The Partial Decoupling Scenario

The most likely scenario is not a total decoupling between Western and Chinese technological ecosystems — economic interdependencies are too deep to allow a clean separation — but a partial and asymmetric decoupling. In this scenario, China achieves substantial autonomy in AI chips destined for its national and military priorities, while maintaining technology exchanges in less sensitive domains.

Western companies continue to operate in China in permitted segments, while Chinese companies continue to sell to global markets in domains where they are competitive. This is not the technological Cold War that some predict — it is something more complex, more nuanced, and potentially more difficult to manage, precisely because the lines of demarcation are not clear.

Who Wins the Long-Term AI Competition?

The honest answer is: we do not yet know. China has considerable advantages — financial mass, political will, captive domestic market, and the capacity to mobilize resources at a speed that liberal democracies cannot match. The West has its own assets — academic freedom that fosters disruptive innovation, the depth of the venture capital ecosystem, military and economic alliances that create critical mass, and the world's best universities.

The competition is real, it is intense, and its results will be decisive for the geopolitics of the coming decades. What is certain is that remaining a spectator — thinking that market forces will solve the problem on their own, or that sanctions will slow Beijing indefinitely — is a major strategic error. The $295-billion plan is not waiting for the West to decide whether it wants to be in the race.

The Stakes of 5G and Artificial Intelligence in the Sino-American Rivalry

5G as Strategic Power Infrastructure

The semiconductor war is not fought only in TSMC's factories or in American export regulations. It is also fought in the deployment of 5G networks, those invisible infrastructures that condition the speed of processing military, industrial, and economic data tomorrow. Huawei, despite the sanctions, today controls a significant share of 5G infrastructure in more than forty developing nations. This is not a market accident. It is a deliberate strategy of entrenchment.

The Biden administration attempted to counter this offensive with the Clean Network program, but results fell short of ambitions. The Trump administration adopted a more fragmented approach, oscillating between tariff pressure and bilateral negotiation. Meanwhile, Beijing continues to deploy. The map of global 5G networks increasingly resembles a map of geopolitical influence.

Artificial Intelligence as a Power Multiplier

The other dimension of this rivalry is artificial intelligence. Beijing explicitly announced its ambition to become the global AI leader by 2030. Chinese investments in this sector have reached levels comparable to American investments, with one major structural difference: access to data. A market of 1.4 billion consumers without European-style privacy regulations constitutes a considerable competitive advantage for model training.

Language models like DeepSeek demonstrated in 2025 that China was no longer technologically behind on AI. It is advancing with its own chips, its own architectures, its own data. The West woke up to an unpleasant surprise: GPU export restrictions had not stopped the race. They may even have accelerated it by forcing Beijing to innovate differently.

Conclusion: The Time for Decision, Not Diagnosis

The Essay as a Call to Action

China does not need Nvidia. This is no longer a projection, a fear, or a clickbait headline — it is a reality materializing through a concrete, quantified, budgeted plan, executed with the rigor of a state that has decided technological autonomy is a matter of national survival. The $295-billion plan, the 80% domestic chip quota, the rise of Huawei Ascend, Alibaba Hanguang, and Biren Technology: these are the bricks of an edifice that is changing global geopolitics.

For the West, the question is no longer whether this reality will crystallize — it is crystallizing. The question is whether we will endure it or respond to it with the determination and resources it requires. Diagnoses are plentiful. Reports are numerous, conferences proliferate, experts converge. What is missing is the translation of these diagnoses into concrete, funded, and coherent actions over the span of an industrial cycle — not an electoral one.

Hegemony Is Lost in Silence

History shows that hegemonies are not always lost in the roar of war or revolution. They are lost silently, through the accumulation of delays, inactions, and deferred decisions. The moment a nation realizes it has lost its technological advantage rarely coincides with the moment it could still have prevented it. China understood this. It is building its advantage while the West debates the nature of that advantage.

Beijing's national AI grid, powered by Chinese chips, will trace the contours of the digital world of the next decade. That world will be more or less open, more or less free, more or less aligned with Western values depending on the response that Europe and the United States can provide. This is a civilizational choice. And it is being made today.

By Maxime Marquette, columnist

Columnist's transparency note

Who I am and my acknowledged biases

I am Maxime Marquette, a columnist and analyst specializing in geopolitical and technological power dynamics. I am pro-West in the sense that I believe liberal democracies represent the best available model of governance. I am concerned by China's rise not out of hostility toward the Chinese people, but because I believe Beijing's authoritarian model represents a threat to the rules-based world order. These biases inform my analysis — I try to be transparent about them.

On semiconductor technology, I am not an engineer — I am an analyst. My readings of the technical capabilities of Huawei Ascend versus Nvidia H100 rely on specialized sources that I attempt to cross-reference. I may be wrong on certain technical details — I would readily acknowledge it if that were the case.

What I do not know

I do not know with certainty how quickly China will close the technology gap with the West in advanced chips. Projections vary considerably among experts — from "never" to "within five years." I also do not know whether the $295-billion plan will be executed on time or will encounter the bureaucratic delays typical of large state projects. What I do know is that the direction is clear and that the trajectory merits serious attention.

My method: I read the available primary sources (official Chinese plans, specialized firm reports, academic analyses), cross-reference them with reporting from Western and non-Western media, and attempt to formulate a reading that is honest about uncertainty and useful for decision-making. I can be wrong. I strive never to invent.

Sources

Primary sources

China Plans $295B AI Grid, Locking Out Nvidia — ABHS Research — June 23, 2026

China implements historic zero tariffs for all African nations with diplomatic ties — Chinese Government (Xinhua) — May 1, 2026

Xi Jinping has hosted more than a dozen leaders this year, as 'middle powers' look beyond the US — The Guardian — June 26, 2026

Secondary sources

China, centre stage — Lowy Institute — June 1, 2026

Multipolar world: What Xi and Putin announced after Beijing summit — Al Jazeera — May 20, 2026

European and Chinese exports kept growing despite the 2025 Trump trade shock — Bruegel — February 2, 2026

China's Zero-Tariff Policy Holds Strategic Promise in Africa as AGOA Falters — Horn Institute — 2026