Editor's note: Recently, as well-known investors and institutions have begun to liquidate core holdings like Nvidia, discussions around the "AI bubble" have intensified. From the "AI frenzy" at the beginning of the year to today's concerns about an "AI bubble," what exactly has happened in the market?
【Anchor】Hello everyone, welcome to DDN Business Insider. I am Yunfei Zhang. Recently, as well-known investors and institutions have begun to liquidate core holdings like Nvidia, discussions around the "AI bubble" have intensified. From the "AI frenzy" at the beginning of the year to today's concerns about an "AI bubble," what exactly has happened in the market? Today, we have invited Wang Xinjie, Chief Investment Strategist from Standard Chartered China's Wealth Solutions Department, and Ms. Guo Hanbing, researcher from the Institute of Finance at the Chinese Academy of Social Sciences, to provide their insights and analysis. Welcome, both of you.
First, I'd like to ask, when did the discussions about the AI bubble begin? Was there a clear turning point in attitude from hype to concerns about a bubble? Mr. Wang, could you share your thoughts?
Currently, the global capital market discussion on the AI bubble essentially began in the third quarter of 2025 and has continued into the early fourth quarter. A significant turning point occurred in October when several top institutions initiated concentrated selloffs of Nvidia. Additionally, OpenAI realized it needed substantial financing, raising concerns in the market about the lengthening investment cycles leading to declines in investment returns and profit margins. By November, the market shifted its focus from whether AI constituted a bubble to what stage the AI bubble is at. We believe that there are evident bubbles in certain areas of the AI-related sector, but it is still uncertain whether the entire AI industry is experiencing a bubble.
【Anchor】Considering recent market data, how do you see the relationship between valuation adjustments, institutional actions, and the debate over the AI bubble?
These three elements are not isolated; they have an interactive logic. We can examine this from several perspectives: market sentiment, capital behavior, and fundamental expectations. What emerges is that once there is a revision of expectations and capital withdrawal, the discourse will intensify, leading to a recurring discussion that further stimulates the next round of capital exit. We have observed that the overall AI investment cycle has shifted from a phase of fervent belief to a more rational investment approach since October and November this year. The recent selling by institutions reflects this shift from enthusiasm to reason. Most institutions still believe that the long-term demand for AI remains strong; however, the increased demand for computing power has led to higher capital expenditures for data center establishments, extending the capital return cycle. This reevaluation is based on long-term capital return rates and capital expenditure efficiency, rather than a bearish stance on the AI concept, which is a significant distinction.
【Anchor】Some views suggest that the massive capital investments by tech giants in AI have led to a circular capital flow with idle investments, where capital is recycled internally without generating adequate real income to justify these investments. Do you think this situation currently exists? If it does, how should we mitigate such risks? Ms. Guo, could you respond?
First, it is clear that the internal capital circulation among these tech giants, which is detached from real-world applications, not only exists but has also formed a relatively classic operational chain. It's not just a simple exchange of money; rather, it is supported by a sophisticated business structure. This cycle is no longer an isolated case; it has evolved into a business model that is an open secret in the industry. While it appears that all parties are profiting, the core issue remains unresolved: AI products have not generated sufficient real income and are instead supported by contracts that exchange investment for orders. This situation is somewhat akin to a house built on sand.
To mitigate this risk, several perspectives should be considered. For investors, it is crucial not to focus solely on the big news from tech giants but to pay attention to two key indicators: first, whether the actual revenue of AI companies can cover their contractual expenditures; and second, to examine the actual utilization rates of hardware. For example, if an AI company's GPUs are sitting idle, the capital chain will quickly break. Real-use scenarios, such as applying AI to industrial quality inspection and medical imaging, which directly help clients reduce costs and increase efficiency, are what truly matter.
For regulatory authorities, there may be a slight delay; we have not yet seen a requirement for these giants to disclose the actual risks associated with these internal circular agreements, such as the debts and revenues within Special Purpose Vehicles (SPVs) and related transactions. This serves as a reminder for investors not to be deceived by superficial revenue and market capitalization.
【Anchor】Yes, Mr. Wang, do you believe that circular financing exists?
Circular financing itself is not a bubble. To define a bubble, one must see that future revenue growth cannot cover existing investment costs and that there is no substantial technological progress. If you increase relevant investments but fail to gain a supposed market position or competitive advantage, then valuation will become detached.
【Anchor】Now, let's turn our attention to the Hong Kong stock market. Currently, compared to the U.S. stock market, the valuations of AI and technology sectors in Hong Kong are still at historically low levels. Does this suggest a rare investment opportunity for long-term investors? Do you think the development of the AI industry in mainland China can sufficiently drive a rebound in the Hong Kong tech sector? Ms. Guo, could you respond?
In my personal opinion, what's truly worth investing in are targets that have industrial logic, rather than those purely driven by market sentiment. There are three areas worth focusing on. First is the upstream hardware technology sector, such as chip packaging and high-end sensors, which are the foundation of the AI industry and are strongly linked to the mainland's industrial chain. They benefit from dual support from both policy incentives and industrial demand. The second direction is AI + vertical scenarios, where leading companies in sectors like fintech and industrial intelligence have already demonstrated real revenue. Their performance certainly will be stronger, leading to quicker valuation premiums.
Another area is services, particularly AI + cross-border service platforms. Given that Hong Kong serves as a hub connecting the mainland and global markets, these companies can capitalize on the incremental benefits of AI going international, aligning with the demand for global capital reallocation. Currently, the sentiment around mainland industry and policy transmission is very positive. For instance, the "14th Five-Year Plan" positions AI as a core component of new productivity, and this policy confidence will provide tangible support, such as R&D subsidies and government procurement, reshaping market expectations.
Currently, the linkage between A-shares and Hong Kong's tech sector is still very significant, and the development of the mainland AI industry will directly bring warmth to Hong Kong stocks. Overall, companies with high commercialization efficiency of R&D and a continuously increasing share of AI business will likely be the first to emerge from the emotional trough, and these are worth paying attention to.
【Anchor】Looking beyond the short-term market fluctuations, from a 5 to 10-year perspective, Ms. Guo, how do you think AI will reshape the global economy and various industries? What key qualities do you believe the companies that will emerge victorious in this transformation should possess?
From a 5 to 10-year perspective, AI is not simply a tool for efficiency in the current applications we use; it will fundamentally change and rewrite the rules of the global economy and industries. Regarding industry transformation, there may no longer be purely manufacturing or service industries; instead, they will all evolve into a form of AI + scenario integration. For example, smart production lines in factories will optimize workflows automatically. In China, we already have factories operating without human intervention, and AI-assisted diagnosis in hospitals is emerging. Additionally, everyday apps can adjust in real-time based on users' habits; some technologies are already becoming prominent.
Data will become the most valuable asset. Whoever makes the best use of data will gain a competitive edge in the industry. The global division of labor and job types will change. Businesses that previously relied on cost reduction and resource accumulation will now leverage AI technology to shape their influence. This is a visible future, especially in a vast market like mainland China, where industries, healthcare, and consumer sectors provide ample soil and scenarios for AI implementation—this is a core advantage for Chinese enterprises.
At the company level, the key companies that will succeed in this AI wave must have several critical qualities. First, they should provide genuine technology without fluff, meaning they are not just riding the trend but are actually capable of solving core algorithms and underlying technologies that others cannot easily replicate. Second, they need to ensure practical application rather than just superficial concepts. Regardless of how impressive AI is, and how strong the technical indicators might be, it must solve real problems—such as reducing costs and increasing efficiency in factories or saving users' time—only then will companies and individuals be willing to pay for these technologies.
Lastly, firms must be capable of handling leverage and cycles. The rate of innovation in AI technology is incredibly fast; what is popular today could become obsolete tomorrow. Companies must resist leverage risks and economic cycles, while remaining agile, to be able to respond quickly and make rapid adjustments to retain top talent.
【Anchor】OK, thank you to all the guests. That's all for this episode. Remember to follow us on YouTube or download our APP. I'm Yunfei Zhang, thanks for watching, and see you next time.
Anchor: Laura Cheung | Edited: Kelly Yang, Laura Cheung, Rachel Liu | Translate: Kato Ip | Proofread: Chris Liu
Related News:
International gold price maintains strong momentum: UBS raises target to US$4,500
Comment