Will Brute Force Take China to Lead?

There is a technology race between the US and China. The most important areas in this race are hardware, software, and energy. China is developing a new model to counter the US's high-tech and high-efficiency production strategy. This model is called Brute Force.

In the technology race, the US has a significant advantage in hardware and software. US companies have the technological expertise and patents necessary to produce high-tech and high-efficiency chips. The US uses these advantages as leverage against China, restricting Chinese companies access to these technologies. However, China also has its own methods and areas where it would be more advantageous. Despite having limited access to high-efficiency chip production, China attempts to close the gap by producing a large number of less efficient chips. This model constitutes China's "Brute Force" strategy.

Domestic Companies and Strong State Subsidies

Chip foundry company SMIC is one of the companies most affected by US sanctions. Lacking access to ASML's advanced EUV chip machines, SMIC cannot produce high-yield chips like TSMC and Intel. Continuing production with ASML's older DUV machines, SMIC is able to produce 5nm chips using methods like multi-patterning. However, this method simply means processing a chip multiple times, which lengthens the production process, increases costs, and significantly reduces the efficiency of chips produced using this method. Despite these disadvantages, the Chinese government provides SMIC with billions of dollars in strong state subsidies to enable it to continue production. These incentives and subsidies are actually aimed at sustaining a larger ecosystem. Approximately 90% of SMIC's total revenue comes from the Chinese market. Other state-subsidized companies within the ecosystem constitute SMIC's customers.

US-based chip design company Nvidia is the world leader in the production of AI data center chips. Nvidia's market share in this field is over 80%, leaving no room even for its US competitors. The other major market share holders are also US companies, Intel and AMD. Chinese chip design companies need to compete with Nvidia in a field where even US companies are struggling.

Nvidia's sale of its most advanced chips to Chinese companies was banned by US sanctions. However, Nvidia CEO Huang disagrees with the US government on this point. He believes that if Nvidia chips cannot be sold to Chinese companies, those companies may develop competing products. Following the ban on the sale of its most advanced AI chips to China, Nvidia produced lower-performance AI chips specifically for the Chinese market. Now, whether these chips can be sold to China remains a mystery. The Trump administration approved the sale, but the US government demanded a 15% commission. Nvidia accepted the terms, but then the Chinese government refused to buy them under those conditions. Aiming to improve and develop domestic competitors.

Although Huawei's Ascend series AI chips lag far behind Nvidia's most advanced chips, they are not significantly inferior to low-performance Nvidia chips specifically designed for the Chinese market. The Ascend chips developed by Huawei are manufactured domestically by SMIC. Huawei's strategy is to produce its chips in large quantities using subsidies, even with lower performance, and then bring thousands of them together to create GPU clusters that rival Nvidia GPU clusters. This method is known as Brute Force.

In China, besides Huawei, there are other chip design companies like Cambricon and Hygon. Just like in the US where there's Intel and AMD alongside Nvidia, a separate kind of competition is emerging within China. Thanks to state incentives and investments, many new ventures are also emerging. Biren, Enflame, and More Threads are other emerging chip design companies in China. However, due to US sanctions, all these companies cannot work with the world's most advanced chip foundries, such as TSMC, Intel, and Samsung Foundries. They are all forced to continue their operations using SMIC's local solutions.

The limited access of Chinese companies to advanced hardware is pushing them to find new ways. The DeepSeek start-up has proven that competitive AI models can be produced without the need for advanced hardware. China is also succeeding in producing competitive models at a much lower cost by leveraging its educated workforce advantage in creating new language models.

Belt & Road Initiative and Technological Collaborations

China's Belt and Road Initiative is an effort to reduce US influence by strengthening economic and political ties with friendly and allied countries. Through the Belt and Road Initiative, China exports technology to countries in the Global South, such as South America and Africa, that are deprived of US technology. This creates new markets for Chinese companies. China's recent relations with South American countries are a strategy that has seriously disturbed the US. Through the Belt and Road Initiative (BRI), China also aims to profit from the sale of artificial intelligence chips by its companies. While US companies cannot sell advanced AI chips to many countries, China wants to turn this into an advantage.

Huawei not only sells the chips it develops, but also provides software and services. Nvidia's CUDA software ecosystem is only accessible through Nvidia chips. Huawei aims to create an alternative to CUDA with the software libraries and services it provides. By offering software, services, and chips as a package, Huawei aims to dispel doubts about its ecosystem.

While US companies offer the most advanced and efficient chips and the best software infrastructure and libraries, Chinese companies have many disadvantages in both chip and software. However, they are trying to overcome these disadvantages with the services they provide. China aims to develop technological cooperation with friendly and allied countries by extending the Belt and Road Initiative to the digital field.

The Importance of Energy Production

Tesla CEO Elon Musk stated at a technology forum, "Next year we may be producing more chips than we can turn on."

Musk points out that the rate of increase in energy production is very slow, and there is a risk that there will not be enough energy to power the chips being produced. The US has almost stagnated in energy production growth over the last 10 years compared to China. Construction times for new nuclear power plants and natural gas power plants are much longer in the US than in China. China is rapidly building its energy infrastructure, leveraging its construction capability. The high domestic production of solar panels and wind turbines alongside nuclear power allows China to reach very high capacities in a short time.

Although China has surpassed the US in total energy production, it still lags behind the US in per capita energy production. However, while China is expected to increase its electricity production by another 2,800 GW by 2030, this figure is only 110 GW for the US. This means China is closing the gap rapidly, 25 times faster than the US.

Although the US produces the most advanced chips, it faces the risk of not being able to meet the energy needs of its data centers. In contrast, China, despite having chips that consume high energy and operate with low efficiency, can close this gap through increased energy production. Moreover, if China can successfully improve its technology, it can easily surpass the US in AI with the help of its energy advantage.

In an interview with a TV channel, Nvidia CEO Jensen Huang stated that China is "nanoseconds behind us," highlighting how quickly China is advancing in technology.

Education and Workforce

China is rapidly increasing its ability to train a skilled workforce in artificial intelligence, shifting the focus from quantity to quality. In 2019, 29% of the top 20% of AI experts worldwide were Chinese, that figure rose to 47% in 2022. Not all of these researchers are based in China; even a significant portion of AI experts working in the US are Chinese. China is attempting to bring these individuals back to China.

The brain drain from China is also reversing as opportunities and research possibilities increase. While only 11% of top researchers remained in China after graduation in 2019, this figure rose to 28% in 2022. And today it is probably even more.

Chinese universities produce far more graduates than US universities. The rate of workforce retention is increasing. Furthermore, the quality of education at Chinese universities is rapidly improving. Chinese universities publish far more papers and obtain far more patents in the field of artificial intelligence. China is rapidly closing the gap and improving in AI.

Summary: Brute Force or Efficiency?

While the US possesses all the technological patents and equipment for the most advanced chips, China is far from having these elements. However, China is not entirely absent from chip technology and is closing the gap with its domestically developed capabilities. Thanks to its manufacturing power, China is trying to close the gap using a brute force strategy by producing a large number of chips. In addition, the speed it has achieved in energy production is rapidly putting China in an advantageous position. The US, on the other hand, faces the risk of not being able to provide the energy to run its data centers even if it can produce a large number of chips. Efficient modeling solutions like DeepSeek, developed by Chinese companies, make it possible to produce good models even with low-efficiency chips.

While the US has many advantages, China has too, and there are also significant risks facing the US, such as the energy problem. Between the US and China the name of the race is Brute Force or Efficiency?