The AI Chip Three Kingdoms: The Future of the Hegemony Struggle Between NVIDIA, Intel, and China
As AI technology prepares to reshape the global industrial structure, the development race for AI chips, which lie at its heart, is intensifying like never before. NVIDIA reigns as the absolute king, Intel has joined forces with a historic rival, and Huawei of China is showing astonishing progress, spurred on by strict U.S. regulations. In this article, based on analysis by semiconductor industry expert Dylan Patel, we unravel the current state and future of the AI chip warring states period, where the strategies of each company are intricately intertwined.
1. A Surprising Alliance: The Lightning Partnership Between Eternal Rivals NVIDIA and Intel
In 2025, a shockwave ran through the semiconductor industry. NVIDIA, which holds over 80% of the AI chip market, announced that it would invest $5 billion in its long-time rival Intel to jointly develop data center and PC products. Regarding this news, Patel describes it as "even hilarious," while pointing out the ironic history behind it.
"In the past, Intel was sued for anti-competitive behavior in the chipset market, and NVIDIA received a settlement from Intel. Now that Intel is packaging NVIDIA chiplets to make PC products, it feels like everything has come full circle, which is almost poetic. It might look like Intel is crawling to NVIDIA, but in reality, it could result in the best products on the market."
Once, Intel reigned as the giant of CPUs, and NVIDIA conquered the graphics market with GPUs. Both companies have fiercely clashed in each other's domains. However, with the arrival of the AI era and the shift of computing's protagonist from CPU to GPU, the power dynamic has completely reversed.
The primary goal of this partnership is to counter the ARM camp, which includes Apple's M-series chips. Laptops that perfectly integrate Intel's powerful x86 CPUs and NVIDIA's high-performance GPUs will likely be an attractive option for many users.
This historic partnership will have a serious impact on competitors. It is particularly a nightmare for AMD, which has been fighting Intel in the CPU market and NVIDIA in the GPU market. Patel describes this situation as follows:
"If your two arch-rivals suddenly team up, that is the worst news imaginable."
Furthermore, Patel compares the influence of NVIDIA CEO Jensen Huang on the market to that of legendary investor Warren Buffett, stating, "It is the Buffett effect in the semiconductor industry." A single decision of his has the power to redraw the power map of the entire industry.
2. The Awakening of the Dragon: The Challenge of China's Huawei to Overcome U.S. Regulations
If the partnership between NVIDIA and Intel is a major move in industry restructuring, another massive tectonic shift is occurring in China. At the center of this is Huawei, a telecommunications equipment giant that is also developing its own high-performance AI chip, the "Ascend."
According to Patel, even before being hit by U.S. sanctions in 2020, Huawei already possessed the technical capability to bring 7nm process AI chips to market that were comparable to NVIDIA's. However, due to sanctions by the Trump administration, it was cut off from access to TSMC, which possesses the most advanced manufacturing processes.
Ironically, this adversity accelerated China's path toward semiconductor self-sufficiency. While Huawei focused on manufacturing at the domestic foundry SMIC, it has also sought to survive by any means necessary, such as procuring chips from TSMC through shell companies.
And now, the Chinese government has even banned the import of downgraded NVIDIA chips intended for China, encouraging domestic companies to use domestic chips like those from Huawei. Surprisingly, the next-generation chip roadmap announced by Huawei, which includes splitting inference processing into "Prefill" and "Decode," follows NVIDIA's latest strategy.
2-1. The Biggest Bottleneck is "Manufacturing Capacity"
However, a major wall stands before Huawei. That is the mass production capacity for cutting-edge chips, including custom HBM (High Bandwidth Memory). Even if they can design them, the supply chain to stably manufacture them in large quantities is not keeping up. Patel points out:
"For them (Huawei), the biggest question is production capacity. Specifically, how much custom HBM they can manufacture is the key. It will still take time for China to catch up to the production capacity that South Korea's Samsung and SK Hynix have built up."
Interestingly, recent Chinese trade data shows a surge in imports of "etching equipment," which is essential for stacking HBM. This is evidence that China is working to resolve the manufacturing capacity bottleneck on a national level.
2-2. "Domestic Production" as a Negotiating Card
Patel views China's series of moves not just as a technical development race, but as part of a sophisticated negotiation strategy with the U.S. government. By showing off the capabilities of domestic chips and banning NVIDIA products itself, it is sending the message that "we do not need NVIDIA." The analysis is that they may be trying to induce regulatory easing from the U.S. side through this.
"It's like we're playing checkers while they're playing chess. It's a 10,000 IQ play."
3. The King's Strategy: How NVIDIA Builds Its 'Moat' and Envisions the Future
Even in the face of a formidable challenger like Huawei, NVIDIA's dominance remains unshaken. While CEO Jensen Huang has acknowledged Huawei as a "formidable competitor," he has solidified his company's position. Where does the source of that strength lie?
3-1. Management That Doesn't Fear Risk: The History of Jensen Huang's 'Bets'
NVIDIA's history has been a series of bold, "company-betting" decisions made by Jensen.
Original Xbox Order: Before receiving an official order from Microsoft, he invested the company's entire fortune to begin production.
Crypto Asset Bubble: Amidst uncertain demand, he convinced the supply chain that mining demand was "sustainable gaming demand" and significantly increased production capacity. After the bubble burst, they recorded inventory valuation losses, but they had already reaped far greater profits.
Patel recounts a story he heard from an industry insider about Jensen at the time:
"Literally, he was just like, 'Fuck it, YOLO.'"
This willingness to take risks is a unique strength of a founder, something that professional managers seeking stable quarterly results cannot imitate.
3-2. Overwhelming Execution and Culture
NVIDIA's strength is not just in its boldness. It is supported by overwhelming execution.
First-Time-Right Chip Design: In semiconductor development, it is common to create multiple prototypes (revisions) to fix design flaws, but NVIDIA almost always completes a perfect chip on the very first try. This dramatically shortens development time and accelerates time-to-market.
Rapid Market Response: In the early days of AI, they pulled off a feat by adding "Tensor Cores," specialized for AI processing, just months before the chip design was finalized. Without this decision, another company might have become the hegemon of the AI chip market.
The culture where hardware and software work in close coordination to bring a vision to life at astonishing speed is what forms NVIDIA's unshakeable "moat."
4. A New Phase in the Cloud War: The AI Infrastructure Competition Among Hyperscalers
The battle for hegemony in AI chips is being fought not only by chip manufacturers but also among their largest customers, the cloud providers (hyperscalers).
Oracle's Rise: Database giant Oracle has become the eye of the storm in the AI infrastructure market by signing an unprecedented $300 billion contract with OpenAI. Using their massive balance sheet as a weapon, they are taking the risk of meeting OpenAI's enormous computing demands—a risk even Microsoft hesitated to take—in hopes of reaping huge returns.
Amazon (AWS)'s Comeback: Once the king of the cloud, AWS had temporarily slowed down due to its delayed response to AI infrastructure. However, Patel predicts that with the world's largest data center capacity and its partnership with AI startup Anthropic, AWS's AI-related revenue will return to a rapid growth trajectory.
xAI's Extraordinary Speed: Led by Elon Musk, xAI is building infrastructure at an unconventional speed, such as bringing a massive data center with 100,000 GPUs online in just six months. Their methods are truly extraordinary, such as avoiding strictly regulated states and building power plants across state lines in Mississippi.
5. Conclusion: Surviving the Warring States Period of AI Chips
The competition surrounding AI chips has evolved from simple performance comparisons into an all-out war that tests manufacturing capabilities, supply chains, software ecosystems, and even the vision of executives.
NVIDIA's stronghold still appears unshakable, but the movements of challengers are more active than ever, including Huawei, which is evolving by turning US regulations to its advantage, hyperscalers that move markets with massive capital, and emerging forces building infrastructure at a speed that defies conventional wisdom. The market remains in a state of supply shortages and frenzy, to the point where the method of purchasing GPUs is described as like buying cocaine, calling a few people to ask for prices and inventory."
This dynamic and unpredictable competition is the driving force accelerating the evolution of AI technology. Who will conquer this Warring States period and seize hegemony over the technology of the future? We are witnesses to this historic turning point.
