What is China’s stance on AI Supply chain? Isn’t the US working to diversify dependence on China? Who needs who?

Why the ban exists

An AI chip, in simple terms, is a processor built specifically to run the massive calculations that training and operating AI models requires. Think of it as the engine under the hood of every AI system, from a chatbot to a weapons-targeting algorithm. The most powerful of these, Nvidia's H100 and its successors, can do in hours what would take ordinary computers weeks.

The US government prevents Nvidia from selling some of its more advanced AI chips to China in order to prevent the Chinese military and surveillance agencies from accessing cutting-edge processing power. The logic is straightforward: whoever controls the most powerful chips controls the pace of AI development, and AI increasingly underpins everything from battlefield decision-making to facial recognition networks.

The restrictions, first announced in August 2022, targeted China's semiconductor and supercomputing industries and specifically impact Nvidia's A100 and H100 integrated circuits, as well as any systems incorporating them. The export controls are complex and address a variety of parameters, including the total processing performance of a chip, its "performance density," its interconnect bandwidth, and its memory bandwidth. In other words, it is not just one chip that is banned: any future chip that hits certain performance thresholds is automatically controlled.

These restrictions were initially imposed by President Joe Biden and were then expanded by President Donald Trump in 2025, though Trump has since reversed some of them. As of May 2026, Nvidia can sell its H200 chip and AMD can sell its MI308 chip to China, but those are not the companies' most powerful products, and the deal is tangled. The US Department of Commerce cleared approximately ten Chinese firms, including Alibaba, Tencent and ByteDance, to purchase Nvidia's H200 chip with a limit of 75,000 units per customer, but no chip deliveries have been made, as the deal remains in legal limbo due to the ongoing US-China tech rivalry and Beijing's new supply chain regulations.

Chinese AI firms have reportedly accessed advanced Nvidia computing power overseas, and lawmakers are now weighing closing that cloud-access gap. Cloud access means a Chinese company can rent time on a powerful chip sitting in a data center in Singapore or the Netherlands, without ever importing the chip itself. That is the loophole the controls are still chasing.

Does China need these chips?

Absolutely, and the evidence is visible in what Chinese companies did the moment access was cut off. The launch of China's DeepSeek model took the world by surprise, and it was powerful, but it ran on old Nvidia compute technology. That tells you something important: even China's most celebrated AI breakthrough relied on chips it was no longer supposed to have access to.

The gap is measurable. According to DeepSeek's own testing, Beyond Huawei's best current chip, the Ascend 910C, achieves roughly 60% of Nvidia H100 inference performance. Inference is the step where a trained model generates answers for users. For training, the heaviest compute task of all, the Ascend 910C delivers roughly one-third the BF16 throughput of Nvidia's B200, which is the flagship chip Nvidia sells in the US market.

What alternatives does China have?

Not none, but not enough yet. Chinese semiconductor firms delivered 1.65 million AI GPUs in 2025 out of a total of 4 million units, claiming 41% of the local AI server market, with Huawei alone shipping roughly 812,000 AI chips.

Huawei is the centerpiece of the domestic push. Huawei's upcoming Ascend 960 AI chips, to be commercially available in 2027, are being positioned as China's answer to Nvidia. A further upgrade, the Ascend 970, is already planned for 2028, with each new generation promising to double the computing power of its predecessor. But Huawei's AI chips are still trailing Nvidia's top offerings.

Beyond Huawei, at the high end, national champions are deploying massive clusters for foundational model training, a domain once exclusively held by Nvidia: Baidu is leveraging a 30,000-chip cluster of its own Kunlun processors to train its ERNIE large language models, while DeepSeek has publicly opted for Huawei's Ascend chips for some training tasks.

The manufacturing constraint is the hardest wall to climb. Wafer fab capacity remains a big constraint, with all certified Chinese chipmakers competing for limited production slots at SMIC, whose most advanced stable node is its N+2 process, roughly equivalent to 7nm, while Nvidia manufactures on TSMC's 4nm process. That gap in manufacturing precision translates directly into a gap in chip performance and energy efficiency that software alone cannot bridge.

Morgan Stanley estimates China's total AI chip market could reach $67 billion by 2030, with domestic supply covering roughly 76% of demand. That would be a dramatic shift, but it is still a projection, and it still leaves a meaningful gap at the frontier.

Nvidia Taiwan was founded in California and its legal home remains there. It does manufacture most of its chips through TSMC, Taiwan Semiconductor Manufacturing Company, which is based in Taiwan, and it is building a major second headquarters in Taipei. But being designed or assembled in Taiwan does not make Nvidia a Taiwanese company, and it certainly does not put it outside the reach of US law.

Taiwan is not China

This is the more important piece. Taiwan governs about 23 million people, issues its own passports, has its own professional military, and conducts an independent foreign policy. Since the unofficial end of the Chinese civil war in 1949, China has considered Taiwan to be a rebel region that must be reunited with the mainland, and asserts what is known as the "One-China policy." But Taiwan operates as its own government regardless of what Beijing claims. Washington does not formally recognize Taiwan as a sovereign state, but it keeps robust unofficial ties, sells Taiwan weapons, and treats it as a separate entity from the People's Republic of China. The two are politically and legally distinct.

The ban follows the chip, not the border

US export controls are not a geographic fence around China's land border. They are rules that govern what any company subject to US law can sell or ship, to whom, anywhere on earth. Nvidia is a US company. Its chips are subject to US export law whether they are sitting in California, Taiwan, Japan, or Singapore. A Chinese buyer cannot get around the ban simply by purchasing through a Taiwanese middleman.

And people have tried exactly that. Taiwan charged nine individuals in a scheme to smuggle Super Micro servers containing Nvidia AI chips to China, with servers shipped to Taiwanese subsidiaries ostensibly for domestic use, then diverted to China, both directly and transshipped through Japan, Hong Kong, or Indonesia. One case involved a March US criminal indictment in which prosecutors alleged a wider $2.5 billion smuggling ring moved Super Micro servers through a US-Taiwan-Thailand-Hong Kong-China network. These are not loopholes; they are crimes, and the US is prosecuting them aggressively. Congress approved additional funding to the Bureau of Industry and Security to combat illicit chip exports, with provisions for export control officers specifically stationed in Taiwan and Finland, and the Trump administration is requesting enough to double the workforce devoted to export controls.

Why the question feels logical

It feels logical because Taiwan really is the center of the global chip supply chain. Nvidia has pledged to invest $500 billion in US infrastructure, but the company carries deep dependence on Taiwan's supply chain and scientific institutions. TSMC manufactures the majority of Nvidia's advanced chip designs. So the chips physically exist in Taiwan at various points in their journey. But physical presence is not the same as legal freedom to sell. The moment a chip is routed to a Chinese buyer without a US license, the transaction violates American law, regardless of which country it crossed through to get there.

That is the whole game China has been forced to play: either build chips domestically, or find elaborate workarounds that carry serious criminal risk. The export controls are imperfect, but they are not as porous as they might appear from the outside.

What Beijing says and what analysts actually think

China's leadership has repeatedly declared that reunification with mainland China is both a historical necessity and a non-negotiable objective, calling the Taiwan question a "red line" at the core of its vital national interests. President Xi Jinping stated in his New Year message that China's reunification is "an unstoppable historical process," while not excluding the use of force.

But rhetoric and reality are two different things. US intelligence assessments judge that Beijing will likely continue "setting the conditions for eventual unification" short of conflict, partly because Chinese officials themselves recognize that an amphibious invasion of Taiwan "would be extremely challenging and carry a high risk of failure, especially in the event of US intervention." China publicly insists unification is required to achieve "national rejuvenation" by 2049, which suggests the timeline is long, not immediate.

What has shifted is the confidence on the Chinese side. The Chinese policy community is increasingly convinced that an effort to assert control over Taiwan will happen, and the fundamental driver of that new assessment is the perception that President Donald Trump has little interest in defending Taiwan militarily. The most likely near-term scenario sees all three sides, the US, China and Taiwan, continuing to maintain the current status quo across the Taiwan Strait, with none undertaking a fundamental policy change.

How Taiwan and China actually differ

The two share a cultural inheritance but have grown into genuinely distinct societies over the past 75 years. Taiwan and China share Confucian values, Mandarin, traditional festivals, and classical cuisine roots, but Taiwan spent 130 years on a divergent path: Japanese colonization, four decades of martial law, and a transition to liberal democracy in the 1990s produced a distinctly Taiwanese identity.

Politically, they are opposites. China's government is centralized and authoritarian, while Taiwan's is decentralized and democratic. Taiwan is a self-governing democracy with its own military, currency, passport, government, and elected president. Taiwan is experiencing a growing sense of its own identity over time, with its democratic government and constitution shaping the political perceptions of younger generations who are distancing themselves from Chinese rule.

Economically, Taiwan emphasizes free-market principles and private enterprise, while China's economy retains significant state control and intervention.

Why this matters for chips

This is where everything in this conversation converges. TSMC, the company that physically manufactures most of the world's most advanced chips, including the ones Nvidia designs, is based in Taiwan. If China were to take control of Taiwan by force, it would instantly control the facility that produces the chips the entire global AI industry depends on. That is not a side note; it is arguably the highest-stakes single chokepoint in the global technology supply chain.

The logic underlying the status quo prediction is that any change would have significant implications for regional stability, inviting potentially catastrophic risks for all three parties. Xi understands that a military campaign to "reunify" Taiwan would completely change the geostrategic landscape in East Asia and beyond. That calculus, mutual catastrophe, is currently the most reliable thing keeping the peace. Not goodwill. Not diplomacy. The sheer cost of getting it wrong.

The clearest way to understand Taiwan is to walk it forward in time.

The original owners

Austronesian indigenous peoples inhabited Taiwan for roughly 12,000 to 15,000 years before significant outside contact. They are the undisputed first people of the island. Indigenous peoples still make up about 2.5% of the total population, with 16 officially recognized ethnic groups, many of whom reside in the mountainous areas of central and eastern Taiwan.

The colonial relay

Dutch Formosa (1624–1662) in southwestern Taiwan and Spanish Formosa (1626–1642) in northern Taiwan were the first colonial endeavors that exercised control over some parts of the island. The Dutch brought in Han Chinese laborers to work the land. A Ming loyalist general known as Koxinga then expelled the Dutch and established a government in the south. Koxinga established the Kingdom of Tungning on the island, using it as a base to launch attacks against the Qing Dynasty.

By 1683, the Qing Dynasty defeated Koxinga's grandson, integrating Taiwan into the Qing Empire. Despite the Manchu Qing designating western Taiwan as administratively a prefecture, Taiwan continued to be perceived as a remote and often "rebellious" island, instead of an integral part of the Chinese inner lands. This matters because Beijing's claim rests heavily on the Qing period, but the Qing were not ethnically Han Chinese either. They were Manchu, a people from northeast Asia who conquered China itself. Claiming Taiwan on the basis of Qing rule is a bit like Spain claiming the Philippines because they both happened to be Spanish colonies.

Following defeat in the First Sino-Japanese War, the Qing government signed the Treaty of Shimonoseki, ceding Taiwan to Japan. Japan held it for fifty years. After Japan surrendered at the end of World War II, it handed Taiwan and Penghu to the Republic of China. In 1949 the ROC government relocated to Taiwan while fighting a civil war with the Chinese Communist Party. About 1.2 million people from the mainland came with it. The Communist Party won the mainland but never controlled Taiwan, not for a single day.

How it compares to other disputed territories

Three comparisons come up most often: Hong Kong, Kosovo, and Crimea. Each one is instructive precisely because it is different.

Hong Kong is the closest parallel in terms of Beijing's framing, but it is fundamentally unlike Taiwan. Britain held Hong Kong under a lease that expired, and both sides agreed in 1984 to the handover terms. Taiwan was never handed to the People's Republic of China under any treaty, by any power, at any time.

Kosovo separated from Serbia and declared independence in 2008 after a NATO military intervention. The formal and legal status places Taiwan in a group of territories with disputed sovereignty, along with Kosovo, Western Sahara, and Transnistria. The surface similarity is real: both are self-governing entities whose independence is contested by a larger neighbor. The key difference is that a large majority of Kosovars mobilized in support of independence from Serbia and were recognized by the US and much of the international community fairly quickly. Taiwan has governed itself for over seven decades, built a thriving democracy and economy, yet still lacks formal recognition from most countries, largely because of Beijing's diplomatic pressure.

Crimea is the comparison Beijing most dislikes, because there is concern in the West that if Russia's war against Ukraine is successful, it will encourage Beijing to resolve its dispute with Taipei in a similar manner. But the situations are legally distinct: Crimea being part of Ukraine had been confirmed in more than 400 treaties agreed between Moscow and Kyiv. No such treaty has ever placed Taiwan under the People's Republic of China.

What this means for the chips

The history makes the chip stakes clearer. Taiwan did not just accidentally end up as the world's most important semiconductor hub. Decades of deliberate industrial policy, plus a highly educated and competitive population, built TSMC and the ecosystem around it on an island that has always had to earn its own survival. If Beijing ever took control, it would instantly control the manufacturing of the chips that run most of the world's advanced technology. That is why Washington treats the Taiwan question not as a territorial dispute but as a technology and security emergency dressed in historical clothing.

Your Africa comparison is sharp. The African parallel is the principle of uti possidetis juris, a Latin phrase meaning roughly "as you possess." When former colonies gained independence, this principle was used to uphold colonial-era borders. It was a pragmatic, imperfect fix: it froze arbitrary lines in place to prevent every border becoming a war. The Taiwan question is different in that it predates any colonial-era freeze, but the underlying problem is identical: lines drawn by outside powers, contested by the people living on either side of them.

Who actually governs this

There is no world government with the authority to simply decide. What exists is a layered set of institutions and norms that work only when the parties agree to use them.

The International Court of Justice (ICJ), the principal judicial organ of the United Nations, plays a significant role in the resolution of territorial disputes. United States involved in a dispute can voluntarily submit their case to it. The operative word is voluntarily. The ICJ cannot review a dispute between states unless each state involved has recognized its jurisdiction. China and Russia have never officially considered applying the ICJ mechanism in their international affairs. So for the Taiwan question specifically, the court most qualified to rule is one Beijing has no interest in consulting.

Beyond Huawei the ICJ, another avenue is arbitration, where disputing parties agree to submit their case to an impartial third party or tribunal. The Permanent Court of Arbitration (PCA) in The Hague is a notable institution that facilitates this. The Philippines actually used the PCA against China over the South China Sea, and the tribunal held that China's "nine-dash line" had no legal basis. China ignored the ruling entirely. A verdict without enforcement is just a piece of paper.

The UN Charter itself lists the recognized means of peaceful resolution: "negotiation, enquiry, mediation, conciliation, arbitration, judicial settlement, resort to regional agencies or arrangements, or other peaceful means." That is the menu. None of the items on it are compulsory.

Where it has actually worked

There are real success stories, and they share one thing: both sides chose to settle rather than fight.

In Libya v. Chad (1994), the ICJ resolved a sovereignty dispute by applying a bilateral treaty, and in El Salvador/Honduras (1992), the court used both historical records and state practice to settle the border. The territorial dispute between Peru and Ecuador was resolved through the 1998 Brasilia Presidential Act, which followed years of diplomatic negotiations. The Northern Ireland conflict and the 1998 Good Friday Agreement, facilitated by a mediation team including George Mitchell of the US, Harri Holkeri of Finland, and John de Chastelain of Canada, won the respect of both sides and resulted in a historic compromise. It was the first time the two governments, along with parties from across the divide, agreed on a new political framework.

There is a consensus in the scholarship that negotiation is the most frequently used method in territorial disputes and that mediation is the most popular form of third-party settlement, but legal methods such as arbitration and adjudication have a better success record.

Why Taiwan is harder than most

What all the successful cases have in common is that neither side was a nuclear-armed great power treating the outcome as central to national identity. Taiwan hits every difficulty at once: Beijing frames reunification as the completion of a century of national revival, the United States has made clear it will intervene, and there is no neutral power large enough to credibly mediate between them. The Good Friday model worked because the US could play honest broker. No equivalent broker exists for China and the US.

The honest answer is that the international system has the tools to resolve this peacefully, and they work when smaller powers use them. When the parties are great powers who have decided the stakes are existential, the tools sit on the shelf, and the outcome depends entirely on whether leaders on both sides decide that backing down costs them less than a war.