China’s three cornerstones of AI have made breakthroughs: the performance of domestic chips is close to NVIDIA H800, and the computing cost is reduced by 20%; DeepSeekV3 runs ten times faster than GPT-4, and the training cost is only 5.5 million US dollars; the energy efficiency of industrial quality inspection has increased by a thousand times, and the response speed of medical CT has increased by 40%.
This comprehensive surpassing from algorithm breakthroughs, computing power innovation to application landing is reshaping the global artificial intelligence competition landscape.
The domestic autonomy rate has reached full quota, and the “three cornerstones” in the field of AI have been mastered by China. This time, the United States can be said to have lost.
Under the wave of scientific and technological progress, the competition in the field of artificial intelligence (AI) among countries around the world is becoming increasingly fierce, especially the technology leaders China and the United States, which are launching an all-round competition in the three core areas of algorithms, computing power and applications.
On March 24, China received exciting news: Alibaba’s research team announced that the chips used for its large-scale AI model training have been fully localized, with performance approaching that of NVIDIA’s H800 chip, while the unit computing power cost has been reduced by 20%.
AI robot in China
This progress not only ended the United States’ dominance in the field of chip technology, but also reshaped the landscape of global artificial intelligence competition.
The cornerstones of AI technology cover three aspects: algorithms, computing power, and practical applications. In the past, China was relatively weak in algorithms and computing power, but in the application field, relying on huge data resources, China maintained a certain advantage.
Despite this, China’s technological leaps in the above three fields in recent years have won wide attention from all over the world.
China ranks among the world’s top in industrialization, with the most comprehensive industrial system and huge industrial data resources.
This gives China unique conditions in the field of AI applications, such as industrial quality inspection, traffic control, and medical diagnosis.
In the field of industrial quality inspection, the brain-like computing technology of domestic storage and computing integrated chips has achieved a thousand-fold increase in the energy efficiency ratio of image recognition; in the medical field, the response speed of domestic CT imaging systems has increased by 40%, while energy consumption has also decreased by 35%.
These data not only demonstrate China’s strong technology in application practice, but also build a more stable and virtuous cycle foundation for its technology ecosystem.
Ling-Plus performance
However, in contrast, the United States has difficulty providing sufficient real data to support AI technology due to the hollowing out of the industry.
For a long time, the United States has been dominating the field of algorithms thanks to the support of companies such as OpenAI.
However, the emergence of China’s DeepSeek model has completely overturned this pattern.
The operating speed of DeepSeekV3 is ten times that of OpenAI GPT-4, and its model performance has also increased by 2 to 3 times, while the training cost has been greatly reduced to US$5.5 million.
This is in sharp contrast to OpenAI’s single training cost of hundreds of thousands of dollars.
Chinese AI for humanoid robots
If commercial gross profit is taken into account, the actual cost of DeepSeek may be only 20 to 50 times the cost of GPT-4.
With its significant advantages in cost and efficiency, DeepSeek has eclipsed up to 80% of AI algorithm solutions in the US market, thereby reshaping the competitive landscape of the global algorithm industry.
The cornerstone of AI computing power lies in hardware equipment, especially high-performance chips designed for model training.
In the US market, Nvidia H800 has long been the dominant player, with its excellent performance and huge market share difficult to shake.
Despite this, Chinese chip companies are actively attacking this leading position with technological innovation.
On March 24, Alibaba announced that its large-scale AI model training has fully adopted domestic chips, and its performance is almost comparable to Nvidia’s H800.
In addition, the unit computing cost of these chips has been reduced by 20%, which once again strengthens its economic advantage.
Comparison between Ling-Plus-Base model and other representative models
This achievement is not only a technological breakthrough in computing power, but also a profound change in the global chip industry.
In the past, China relied heavily on the United States for high-end chip technology, with companies such as Intel and Nvidia dominating the market.
Nevertheless, the trade dispute between China and the United States and the US’s technological containment measures against China have unexpectedly pushed Chinese companies to accelerate their pace of independent innovation.
Comparison between Ling-Plus and other representative models
With technological innovation, companies such as SMIC and Huawei HiSilicon have advanced from 14nm to 7nm process technology, steadily narrowing the technological gap with the United States.
The significant progress made in domestic chips has had a significant impact on the global artificial intelligence chip market.
The price of Nvidia’s H800 chip experienced a sharp drop in 2023, from the original 300,000 yuan/piece to 60,000 yuan. At the same time, its market share in China has also fallen from a high of 92% to 47%.
This market dynamic not only reflects the booming development of Chinese technology, but also exposes the vulnerability of the pricing strategy of American chip manufacturers.
At the same time, China’s contribution rate to the results of top artificial intelligence conferences is also increasing.
By 2025, it is expected that the contribution rate of Chinese teams in fields such as low-power training and heterogeneous computing will climb to 38%, and in terms of quantum computing chips, its patent applications will account for 27% of the global total.
These data show that China has not only achieved a leading position in technology, but also comprehensively challenged the long-standing dominance of the United States in the two major fields of academia and market.
The comprehensive leap in the three pillars of China’s AI has laid a solid foundation for the rise of domestic chips.
This technological innovation not only enhances China’s competitiveness in the world, but also opens a new page for rebuilding AI technology benchmarks and market norms.
The United States may use technical barriers, industrial chain constraints and other measures to hinder China’s technological progress.
At the same time, China urgently needs to deepen its investment in technology research and development and optimize the industrial environment to fundamentally consolidate its existing advantages.
It is foreseeable that the technological competition between China and the United States will be protracted and continue to have a profound impact on the global economic and technological layout.
China’s domestic chips have made comprehensive progress, prompting China to achieve a comprehensive leap from application advantages, algorithm frontiers to computing power in the field of artificial intelligence.
This marks that China’s technology has stepped onto an important development ladder, and it also indicates that the global AI technology field will usher in a profound change.
China’s resilience and potential have been clearly demonstrated on the stage of technological innovation and market competition.
However, in this technological competition, who will laugh last remains to be revealed in the years to come.
There is no doubt that China is gradually showing an increasingly important position in the global competition in the era of artificial intelligence.
AllAIApplications.com focuses on all AI applications, AI tools, global AI ideas and AI product advertising services. We are committed to AI technology research, AI idea exploration, using AI for business, AI for education, AI for cross-border e-commerce and other projects. If you have this idea, please contact us to discuss it together.
If you need to place AI ads on our website, you can also contact us at any time.
