As the sun rises in the east and sets in the west, Chinese stocks are experiencing a remarkable surge, with specific sectors like artificial intelligence (AI) applications, deep learning tools such as DeepSeek, and various domestic cloud services all poised for explosive growth.

The most buzz-worthy topic during the recent holidays has been DeepSeek and the seismic shift it brings to the global landscape, primarily characterized by a significant reduction in training costs. The implications of such a decline are monumental. For example, if gasoline costs 1000 yuan per liter, the demand for cars would surely diminish, leading people to opt for bicycles instead. When the price of fuel drops to just a few yuan, however, the entire automotive industry would be revitalized—from manufacturing to parts supply, and even extending into transportation-related services.

The significant decrease in key industry costs benefits the entire sector, paving the way for rapid and extensive expansion. This notion cannot be overstated! Just as decreased mobile phone prices enabled the rise of giants like Apple and Xiaomi, and reduced internet charges helped create the titans of the mobile internet era like ByteDance and Tencent, even data service providers are thriving. Financial prosperity is not only a possibility; it’s a reality.

Therefore, the emergence of DeepSeek acts as a turbocharger for the development of the artificial intelligence industry, presenting new opportunities for global tech and industrial transformation! More nations can engage in this quest, emboldening research sectors to boldly explore cutting-edge technologies and hasten the pace of technical iterations. In the realm of industrial applications, small to medium-sized enterprises can now embrace AI at lower costs, catalyzing the smart transformation of sectors such as manufacturing, healthcare, and education. The competitive landscape is in flux as technological equality brings about a reconstruction of competitive dimensions, allowing innovative firms to seize the moment. The market dynamics are ready to be reshaped by these cost reductions.

The response from the United States has been fervent. Previously, the U.S. reigned supreme in the artificial intelligence domain, holding a lofty position atop the competitive hierarchy. However, the unanticipated rise of DeepSeek has jolted the American tech landscape awake, prompting a recalibration of the comparative distance in artificial intelligence capabilities, which has fueled a surge of intense rivalry. Consequently, DeepSeek's breakthrough is likely to inspire domestic alternatives and prompt a reassessment of the value of local tech companies. It remains to be seen if the U.S. will leverage this situation to impose stricter sanctions, a potential risk for which future considerations must be made; the contest for "AI sovereignty" poses an intriguing question: who will emerge victorious, and who will ultimately guide the next era of innovation?

Strategically looking towards 2025, we see a defining year for the concentrated explosion of AI applications. This entails more than mere software; it represents a comprehensive integration of AI into various fields—manufacturing, education, and private lives alike! This surge will likely trigger a comprehensive restructuring across all sectors amenable to being empowered by AI. Enterprises failing to adapt may perish, while those that adeptly address industry pain points and strive for advancement will thrive, unlocking exponential opportunities!

In this context, echoes of the phrase attributed to a classic poem resonate: "On the side of the sunken boat, thousands of sails pass by; before the sick tree, myriad woods bloom in spring!" If a bull market is indeed brewing, it will be a rejuvenated coalition of countless industries breathing new life, bolstered by artificial intelligence!

With the continuous advancements in AI technology, the establishment of AI infrastructure has emerged as one of the key drivers of the global tech industry. The innovative strides made by DeepSeek not only promote the adoption of AI modeling but also present fresh opportunities and challenges throughout the AI infrastructure supply chain.

The impact of AI infrastructure on various segments can be illustrated as follows: GPU demand exhibits a neutral stance.

As the core hardware for AI training and inference, GPU market demand remains robust. Citi analysts suggest that although DeepSeek's technology might reduce the need for extensive training, computational requirements during the inference stage are expected to rise, leading to a generally stable GPU market.

Retimers present a neutral to negative outlook.

Primarily utilized for high-speed data transmission, Retimers are particularly crucial during AI training. As AI computation shifts from training to inference, demand for Retimers may decline as inference computations are less intensive and the need for rapid data transfer diminishes.

Optical Modules (Intra Server/DC) reflect a neutral to negative perspective.

Similar to Retimers, demand for optical modules is anticipated to be high during the AI training stage. However, with the increasing emphasis on inference, the market demand for optical modules might witness some downturn, particularly concerning interconnections within data centers.

Server OEM (Original Equipment Manufacturers) maintains a neutral stance.

While the demand for training servers might decline, there will be an uptick in demand for inference servers, keeping the server OEM market relatively neutral overall.

ASICs (Application-Specific Integrated Circuits) present a neutral to positive outlook.

ASICs particularly excel during the AI inference stage. With a rising demand for inference, the market share of ASICs is expected to grow. Despite the likelihood of decreased demand during the training phase impacting the ASIC market to some extent, in the long run, the increasing demand for inference will offset the negative implications.

DCI (Data Center Interconnect) has a positive outlook.

DCI is relatively insulated from the nuances between training and inference, as its demand is less contingent on the specific details surrounding AI models and workloads. The growth in inference stages will consequently provide new opportunities for DCI.

Switches present a neutral to positive outlook.

As core components of data center networks, the market demand for switches is closely tied to the proliferation of AI computing. Though a reduction in training requirements might exert temporary effects on the switch market, the growth phase of inference is predicted to boost the need for higher network bandwidth. Hence, overall demand for switches will trend toward a neutral to positive trajectory.

Connectors are categorized as neutral to positive.

In a similar vein to switches, connectors are viewed as largely indifferent to training or inference phases, with their market demand being closely associated with the overall growth of AI infrastructure. With increasing proliferation of AI computing, the demand for connectors is expected to remain stable, particularly propelled by advancements in distributed computing, which offers promising new avenues for growth.

Storage holds a neutral to positive standpoint.

While the impacts of AI on storage OEMs are yet to crystallize, the influence of AI is rarely noted as urgent at this stage. Nevertheless, with the acceleration of inference processes, the demand for aspects such as data management, movement, and security is poised to drive growth within the storage market.

PCs and smart devices show a neutral to positive perspective.

With the optimization of AI models, computing abilities for AI on local devices will improve, leading to an uptick in demand for updates on PCs and smart devices, particularly those equipped with enhanced computing capabilities, fostering a neutral to positive market outlook.

The advent of DeepSeek will yield varied effects across different segments of the domestic AI industry chain, thereby impacting corresponding stock performances. Below, we delineate segments that may face both favorable and unfavorable consequences, along with related domestic stocks: