Zhitong App learned that Guosheng Securities released a research report this week on NVIDIA's &#34;report card&#34; for the last quarter, driving up the stock prices of American AI companies. The core reason behind the stock price surge is the performance, with industry leaders taking the majority of profits in various segments. &#34;Cognitive difference&#34; is more important than &#34;expectation difference.&#34; The current market is driven by fundamentals combined with industry trends, and considering the industry landscape and development stage, it continues to favor leading companies in the optical communication sector.
At the same time, addressing some current market doubts, the institution raised insights on NVIDIA's optical module, explaining the reasons for the continued optimism in the AI and computing power industry chain from three dimensions: the restructuring of the AI industry chain valuation system, marginal changes in computing power demand, and comments on the optical module's abnormal movements.
<h2>Guosheng Securities' main points are as follows:</h2>
On February 22, 2024, the highly anticipated NVIDIA released its FY2024Q4 financial report, achieving a quarterly revenue of $22.1 billion, a 265% year-on-year increase, a 22% increase quarter-on-quarter, surpassing the expected $20.41 billion. The company achieved a net profit of $12.29 billion in the quarter, a 769% year-on-year increase, and a 33% quarter-on-quarter increase. The revenue for the first quarter of the 2025 fiscal year is expected to be $24 billion, with the market expecting $21.9 billion. The company stated that generative AI is becoming a tipping point. Following the financial report, the company reversed its previous downward trend and hit a historical high in pre-market trading that evening (with a market value exceeding $2 trillion). AI-related stocks such as AMD, SMCI, AVGO, Fabrinet, and the Nasdaq and Philadelphia Semiconductor Index all saw varying degrees of increase. The financial report and market response have profound implications for the domestic optical communication/optical module sector.
Currently, computing power infrastructure is driving AI development, so the computing power industry chain can raise the valuation ceiling, and high PE ratios are expected to be complemented by &#34;explosive&#34; EPS growth. Looking at the valuations of NVIDIA and Supermicro Computers, after the AI outbreak, the market's tolerance for their valuations is continuously increasing, with sustained performance explosions being a necessary condition. At the same time, the increase in valuation reflects the market's high recognition of the company's prospects and the AI outlook, especially as we enter 2024, with overseas tech giants accelerating their AI-related deployments, and the release of OpenAI's Sora showing the potential of large-scale world models. As the most important provider of AI tools, NVIDIA has become the &#34;locomotive&#34; of AI, with its continuously iterated products improving performance, reducing computing costs, and accelerating the arrival of the AGI era.
The demand for inference is just beginning, and the cycle of &#34;investment - profit, increased investment&#34; may just be starting. NVIDIA stated at the performance meeting that inference revenue in FY2024 already accounts for 40% of the data center business, which is $19 billion. This means that various AI competitors have already invested hundreds of billions of dollars to maintain their respective AI applications running.
In the future, we can expect to see not only larger models with higher modal dimensions, such as Huang Renxun's vision of &#34;adding sound to videos with large models,&#34; but also the hope of seeing advertisers like Meta continuously running high-growth advertising businesses using AI, and consumer internet companies like Amazon leveraging AI shopping assistants to further increase GMV.
In the AI era, the importance of communication is constantly increasing, and optical modules are entering a period of favorable development. With the continuous development of AI, the importance of communication in training/inference is constantly rising, the speed of product iteration is significantly accelerating to complement GPU computing iterations, and the value is continuously increasing, with industry leaders almost dominating the market. The market has a fairly comprehensive understanding of the importance of optical communication in the AI era. From an industry perspective, optical modules are entering a period of favorable development, having experienced the baptism of the previous round of cloud computing. Domestic core optical module companies are gradually becoming global leaders, and the advantages of the domestic optical communication industry chain are quite evident in many aspects.
Computing Power—
Optical Communication: Zhongji Xuchuang, Xinyisheng, Tianfu Communication, Taichen Optical, Tengjing Technology, Guangku Technology, Guangxun Technology, Dekeli, Liantek Technology, Huagong Technology, Yuanjie Technology, Cambridge Technology, Mingpu Optoelectronics. Computing power optimization/scheduling/leasing: Hengwei Technology, Sitech, Zhongke Jincai, Dongfang Materials, Borui Data, Zhongbei Communication, Zhongke Shuguang, China Mobile, China Unicom, China Telecom.
Computing Power Equipment: ZTE, Ziguang Shares, Ruijie Networks, Shengke Communication, Feiling Kes, Industrial Fulian, Hudian Shares, Hanwuji.
Liquid Cooling: Inveck, Shenling Environment, Gaolan Shares, Jialitu.
Edge Computing Platform: Megvii Intelligence, Guanghetong, Moveon Communication.
Satellite Communication: China Satcom, China Satellite, Zhenyou Technology, Hualichuangtong, Elecchip, Haige Communication.
Data Elements—
Telecom Operators: China Telecom, China Mobile, China Unicom.
Data Visualization: Haohan Depth, Hengwei Technology, Zhongxin Saic.
BOSS System: AsiaInfo, Tianyuan Dike, Eastern National Information.
Risk Warning: AI development falls short of expectations, computing power demand falls short of expectations, market competition risks.

NVDA

002670

Guosheng Securities has released a research report, continuing to be optimistic about the leading companies in various segments of optical communication. NVIDIA's earnings report has driven up AI stocks in the United States, and the market's tolerance for its valuation is constantly increasing. The computing power industry chain can enhance the valuation ceiling, and high PE ratios are expected to be filled in by an "explosive" EPS growth.

Guosheng Securities: Insights from NVIDIA's Earnings Report on Optoelectronic Modules