The U.S. State Department announced on the 12th that a multilateral cooperation framework on semiconductors and critical minerals, led by the United States with participation from multiple allies, has been established and the declaration document was signed in Washington.

Old Hu first gives a basic opinion: This collaboration mechanism, which is being promoted loudly, is equivalent to a "band of misfits" at this stage.

Washington gave the new cooperation framework an impressive name, "Pax Silica," which means "peace" and "silicon" in Latin, directly targeting the chip and AI core supply chain. The initial signatories of the framework are the United States, Japan, South Korea, Australia, Singapore, and Israel, with other participants including the Netherlands.

Washington's loud promotion of this cooperation framework is both a move to pressure China and conduct psychological warfare, as well as to encourage allies to stick together, seeking a way to address America's most pressing problem of rare earth shortages, reducing what they call "coercive dependence," and ensuring AI technology is self-sufficient and controllable from the root.

The cooperation framework has outlined a preliminary division of labor: Australia will be responsible for mineral extraction, Japan will provide separation and recycling technologies, South Korea will focus on semiconductor and AI component manufacturing, and the United States will coordinate funding, rules, and the market.

Politico reported that "this initiative highlights the Trump administration's high concern over China's near-monopoly position in rare earths and its dominance in other parts of the global supply chain, viewing it as a major threat." Politico also said that "the declaration reflects American concerns about China's massive investments in artificial intelligence and quantum computing, which could give it an advantage in economic competition in the 21st century."

In recent years, the United States has led multiple rounds of formal and informal alliances in semiconductors, with one of the core goals being to curb the development of China's chip and AI industries and build a "China-free" supply chain loop.

In fact, the United States has strong capabilities in building alliance frameworks and establishing cooperation mechanisms, but after Trump returned to the White House, he showed no interest in multilateral mechanisms. He often viewed multilateral mechanisms as empty forums, useless, and believed that such mechanisms provided opportunities for other members to take advantage of the United States. Therefore, Washington now prefers to engage in short-term, profit-driven activities, likes to exploit resources quickly and make quick money. For example, Washington has now lost interest in the G7, is not enthusiastic about NATO, has basically abandoned the USMCA, and the AUKUS has cooled down.

However, Washington's current rush to establish the new "Pax Silica" cooperation mechanism is an exception in the broader trend of decline in U.S. multilateralism.

Washington did this primarily because establishing a real key mineral supply chain related to rare earths is too difficult for the United States, and even with its allies combined, it is still challenging. Because rare earth resources themselves are scarce, and more importantly, the separation technology of heavy rare earths cannot be achieved just by design. China succeeded because our industrial chain is unparalleled globally, with many metallurgical and chemical industries highly integrated with the rare earth industry, and the large-scale output of various key minerals is the result of the scale effect of industrial cooperation, not something that can be done by setting up a few production lines alone.

Now, the heavy industry of the U.S., Japan, and South Korea combined is still not as strong as China's, and they also suffer from a lack of electricity, which determines their fundamental weakness in rare earth, especially heavy rare earth industries. For example, Japan may not find it difficult to develop separation and purification technology, but it is difficult to match the scale of resource support. It would take nearly ten or eight years for the U.S. and its small allies to figure out these things together and form effective cross-border combinations.

Certainly, in the long run, any problem can be solved if we are determined. However, industrial competition linked to geopolitics has a time frame.

The speed at which countries like the U.S. and Japan can reduce their reliance on Chinese rare earths must be significantly faster than the speed at which China can reduce its reliance on their chip technology, otherwise it will not produce the desired geopolitical significance.

But now, what is very bad for them is that the Trump administration announced on Monday that it would allow the export of relatively advanced NVIDIA H200 chips to China, while the Chinese response was very cold. Bloomberg reported that the White House's AI affairs chief, Sax, said that China has already figured out the strategy of allowing it to purchase NVIDIA H200 chips, so it refused the AI chip and instead supported domestic semiconductors. Sax said, "China rejected our chips. I think the reason is that China wants to achieve independence in the semiconductor industry."

Bloomberg also reported that the U.S. decided to allow exports of H200 chips to China based on an assessment that Huawei, NVIDIA's main competitor in China, provides AI systems whose performance is comparable to NVIDIA's chips, including its Cloud Matrix 384 platform. The report stated that this platform connects hundreds of processors to compensate for the shortcomings of individual chip performance.

Another Bloomberg report said that China is considering launching an incentive plan worth up to $70 billion to support the local chip manufacturing industry.

Meanwhile, the newly formed "Pax Silica" by the U.S. is currently in the "band of misfits" phase. They have money, but they are basically at a loss about how to proceed. Australia mainly produces light rare earths, with less than 5% heavy rare earths, and the cost of mining and refining is estimated to be 30%-50% higher than China's. Japan's separation technology has high costs, and the environmental evaluation and capacity ramp-up cycle are long, making it difficult to form effective alternative capacity in the short term.

The establishment of this cooperation mechanism is more like the U.S. big brother gathering his little brothers to drink together, everyone drinking a few more glasses, patting their chests. But how exactly to do it and whether it can be done, no one is sure. The benefit is that they all feel more confident.

They also have a special hope that when China sees the spectacle of them drinking together, it will feel very uneasy, its determination to develop independent chip technology will waver, and its strict management of rare earth exports will also waver. Deception is the way of war; they hope to scare China into a "empty city" strategy version of the U.S. and its allies.

China should still say the same thing: do our own things, not be disturbed by complex external information, focus on breaking through the high-end semiconductor barrier, and our progress is being accumulated step by step.

The anxiety and panic of the other side are spreading from core manufacturers like NVIDIA to the political arena. In the crucial race between China's chip technology self-reliance and the U.S. and its allies' rare earth supply self-reliance, we have clearly been running faster. Let's continue to maintain this speed difference.

Original: toutiao.com/article/1851403442247747/

Statement: The article represents the views of the author.