Annual Report Season "Stars" Make Their Debut: AI Computing Power Industry Chain Leads in Positive Outlook, with Upstream Core Component Companies Achieving High Growth

CNR Beijing, March 13 — (Reporter Wang Ying) According to the Voice of China Central Radio and Television’s “World Finance” program, the A-share market’s 2025 annual report disclosures are entering a peak period. As of March 10, among nearly 3,000 listed companies that have disclosed earnings forecasts or quick reports, a clear trend has emerged — the AI computing power industry chain has a forecasted positive rate of over 50%, far exceeding the overall market level, making it a shining “star” during the annual report season. What forces are driving this sector’s rise? Can the high growth logic of upstream core component companies continue?

Recently, the “AI computing power” sector in the A-share market has been very lively. Increasingly positive earnings forecasts have pushed the AI computing power industry chain into the spotlight. Fields like optical modules, PCBs, and storage chips, which were not well known to ordinary investors in the past, have now become the most eye-catching “front runners” during the earnings season.

Is this just short-term capital speculation, or a substantial industry cycle reversal? Market analyst Gui Haoming believes: “Many companies have achieved high growth, mainly reflected in increased sales, some have turned losses into profits, and some have seen significant performance improvements. Overall, their situation is better than other industries.”

However, within the industry chain, the degree of performance realization shows clear differentiation. A review shows that upstream core component companies have much higher performance certainty than downstream application companies. In this regard, Yang Delong, chief economist at Qianhai Kaiyuan Fund, analyzes: “In the past two years, the demand for AI computing power has led to differentiation within the industry chain. For the same computing power, upstream core component companies have much higher performance certainty than downstream application companies. The main reason is that current AI development still relies heavily on demand for upstream components, and it may take more time for this to gradually translate into applications.”

As AI large models are trained more deeply, market attention to chip equipment and manufacturing processes is increasing. In this global wave of AI chip capacity expansion, how much of the pie can domestic companies get? Gui Haoming points out: “In the entire industry chain, what stands out most are companies preparing to invest in the AI industry, including those investing in computing power. Additionally, companies producing equipment are also gaining considerable attention. Both upstream and downstream companies are expected to profit. Relative to others, more effort is needed to develop chip manufacturing. As the demand for AI industry computing power continues to grow, chip design may grow faster than the industry average.”

For ordinary investors, identifying “true growth” amid a complex market remains a challenge. Gui Haoming suggests that investment should return to fundamentals and focus on core indicators.

“The first step is to find genuine industry leaders. Many companies are involved, but which are truly strong, and which are just riding the wave? Second, focus on companies with high market share and advanced technology. Additionally, companies that are making big moves in AI computing power deployment, with strong capital and management capabilities, are worth paying attention to. From an investment perspective, we should adhere to fundamentals, so it’s best to base decisions on the performance of listed companies for rational investing,” Gui Haoming said.

Yang Delong believes that, besides paying attention to net profit growth, order status and whether a company has entered the supply chain of major global manufacturers are also key “barometers” for selecting quality targets.

“Besides net profit, investors should also monitor order situations. Some companies may not yet show profits, but if they have many orders, it indicates good future growth potential and true growth attributes as AI computing power leaders. Also, check whether these companies can secure orders from major firms like NVIDIA or Tesla, and whether they can enter the supply chains of these industry giants, which could lead to future performance growth. By considering upstream core component performance and competitive landscape, investors can choose leading companies as quality targets,” Yang Delong said.

Opportunities often come with risks. In light of signs of overheating in some concept stocks, Yang Delong especially reminds investors: “Some companies are just riding the concept without substantial performance support, and are not industry leaders. There is a risk of bubble bursting. Therefore, investing in tech stocks should adhere to value investing principles—select industry leaders with R&D capabilities and technological breakthroughs, and avoid companies that hype concepts or themes.”