China AI chip startup was acquired by American giants?

Recently, Shen Jian Technology was acquired by Xilinx, the leader of American FPGA. According to industry estimates, the acquisition amount may be around $300 million. After the news came out, many netizens seemed indignant. Some netizens said that China's AI chip troika: CAMBRIAN, Horizon and Shen Jian technologies were thus abolished by the United States, and they thought that the government did nothing.

Perhaps artificial intelligence is the frontier, and the importance of artificial intelligence chips is actually overestimated by netizens. For China, Shen Jian Science and Technology is in a state of "more things than less things", and the completion of this acquisition is a win-win move for Shen Jian Science and Technology and Xilinx.

There are some similarities between Shen Jian technology and CAMBRIAN. The founders are all technology giants from the Chinese Academy of Sciences, and their scientific research achievements have won many awards abroad, which belongs to "applying what they have learned is business". However, there are also differences between them, that is, the CAMBRIAN chose to take the road of dedicated chips, while Shen Jian Science and Technology chose FPGA based on Xilinx as its own artificial intelligence chip.

Specifically, the CAMBRIAN brought the designed circuit diagram to TSMC for streaming transmission, while Shen Jian Science and Technology ran its optimized neural network pruning technology on Xilinx FPGA, and then used these FPGAs to run artificial intelligence applications.

In other words, Shen Jian technology has a strong dependence on Xilinx's FPGA. On the other hand, Xilinx also needs Shen Jian technology to help it develop the artificial intelligence application market. Therefore, the two companies are "winning and losing together", and no one can do without anyone. In this case, from a commercial point of view, it is normal for Xilinx to complete the acquisition of Shen Jian Technology.

At present, domestic AI chip companies have mushroomed, and more than 40 companies have completed various financing. Most of them often announce how to hang NVIDIA on their chips, but the specific application is not obvious at all, and their hematopoietic capacity is also unstable. Since Xilinx is willing to invest in Shen Jian technology and continue to burn money so that FPGA can better explore the artificial intelligence market, that's good. After all, for China, Shen Jian's technology belongs to a state of more than one thing, which will neither threaten national security nor affect foreign monopoly. In this case, business is business. After all, there is no need for the government to intervene in this small matter.

At present, a large number of domestic start-ups are catching up with artificial intelligence, and even Internet giants such as Ali and Baidu have joined in. These artificial intelligence chip companies have won the favor of capital and gained high valuation, and the whole industry has shown a thriving scene.

Although it seems that China's artificial intelligence chip industry is in full bloom, domestic artificial intelligence chips rarely reach consumers in specific applications. In addition to the CAMBRIAN licensing IP to Huawei, which realized tens of millions of applications, other domestic artificial intelligence chip companies performed very well in marketing and application.

There are too many artificial intelligence chip companies, and one problem is redundant construction and scattered strength. First, the whole society's investment in artificial intelligence chips is scattered; Secondly, the power of constructing industrial ecological chain is dispersed; Finally, the potential for large-scale popularization and application is weakened, and it is easy to become that everyone eats a little, but no one has enough to eat.

I believe that what domestic artificial intelligence chip companies really need to do is not how to hang NVIDIA on the PPT of the conference, but how to combine it with specific applications, use the chips and sell them. What China lacks is not companies that design artificial intelligence chips, but companies that apply these domestic artificial intelligence chips. Only when everyone uses domestic AI chips, will they have a chance to eat into NVIDIA's market share and not be carved up by foreign giants.

Source: Observer Network