Newsletter commentary Aug 2025

Fom being “uninvestable” to “unignorable

In August, discussions of a bull market turned into reality. With consumer and healthcare sectors taking a pause, the rally was carried almost entirely by the AI compute chain. This strength is tied both to the visible progress and accessible market space of the computing power chain, as well as to the unique structure and ecosystem of China’s market.

AI is likely to remain the single biggest driver of global economic growth for a long time. While the pace of advancement in large models has slowed, commercialization is making rapid progress in both depth and breadth. Gains in advertising efficiency are already evident, programming productivity has improved, and workflow transformation is underway, and the employment impacts may already be visible. There is already voice in the market calling for “redoing everything with AI.” China's computing power chain is integrating into the global computing power supply chain, the entire domestic computing power industry chain may already be on the verge of a new round of upgrading. New progress has been made across the board, from design and manufacturing to computing power clusters, models, and applications; for investors, the only remaining issue is valuation.

The market transaction volume has reached around 3 trillion yuan in August. It doesn’t take long for the market to shift from a frigid state to a seemingly boiling one. Calculated at a daily trading volume of 2 trillion yuan, over 240 to 250 trading days a year, the annualized transaction volume is nearly 500 trillion yuan. Corresponding to a circulating market capitalization of approximately 40 trillion yuan, the annual turnover rate exceeds 10 times. In contrast, the U.S. market’s S&P 500, with an almost fully circulating market capitalization of about 60 trillion U.S. dollars, sees a daily trading volume of around 300 to 400 billion U.S. dollars, resulting in a turnover rate of roughly 1.5 times.

From the perspective of transaction structure: the circulating market capitalization of the CSI 300 accounts for about 50% of the total, while its transaction volume typically accounts for around 20%. The combined circulating market capitalization of the CSI 1000 and CSI 2000 makes up just over 20%, yet their transaction volume often reaches 45%. On Aug 31th, 2025, the trading volume of NVIDIA in the U.S. stock market was 42.5 billion U.S. dollars, and that of Tesla was 27.3 billion U.S. dollars. Together with Apple, Meta, Microsoft, Google, and Amazon, their total trading volume was approximately 110 billion U.S. dollars, with their transaction share roughly matching their 30% market capitalization share.

Currently, the TTM (Trailing Twelve Months) P/E ratio of the CSI 2000 stands at 159 times, which is presumably a record-high P/E ratio. Around the end of 2020 and the beginning of 2021, the "Ning Index" (a group of high-growth new economy stocks) had a P/E ratio of about 80 times, and the "Mao Index" (a group of leading blue-chip stocks) had a P/E ratio of 50 times. Subsequent adjustments took 3 years to repair these valuations. It remains uncertain how this 159-times P/E ratio will "land" (i.e., adjust to a more reasonable level). After experiencing the shock in early last year, small-cap stock-related indices have reached new highs again, further strengthening some investors’ conviction in such indices. This could be a problem, and it may merely mean standing on the edge of a higher cliff. Although some people believe they can exit first, this will be a difficult challenge for the group as a whole.

It’s possible that the transaction taxes and fees in our market are not too high, but rather too low. In a market with almost no frictional costs (from low taxes/fees), it is prone to over-excitement. A certain linkage between transaction taxes/fees and transaction volume could serve as a "brake mechanism" (to curb excessive speculation). For the overall market, the share of these indices (small-cap focused) is limited, so their impact on the broader market is actually minimal. However, the associated investor behavior indicates that many medium and small investors are involved here.

An extremely active market, coupled with very low transaction taxes and fees, a large number of retail participants, and a growing number of passive and quantitative institutional investment behaviors that lack value assessment—this is a combination that can easily lead to extreme market movements.

I believe that what was deemed “uninvestable” by some investors a year ago is now turning into a market that global investors cannot afford to ignore. Whether such exuberance can be sustained, however, is another question. Shifts in narratives also act as amplifiers, both on the way up and down. So far, there has not yet been a massive shift of savings into equities, but relative to China’s vast pool of savings, the tradable stock market remains small. This structural reality underscores a deeper challenge: how to make the market truly benefit the majority, an extremely difficult but meaningful task.

This month, the Federal Reserve’s stance on rate cuts has shifted. At the start of the year, concerns about U.S. growth were widespread, yet the economy kept surprising to the upside. Soft and hard data were at odds, but more recently, signs of weakness have appeared, particularly in employment. When Trump first took office, his first 100 days delivered limited results, but what followed was an almost unrivaled, bullying style of governance that pushed through many initiatives, whose long-term consequences are still unfolding. Markets initially feared tariffs but later shifted to seeing any trade deal as a positive. Ultimately, however, tariffs plus other conditions are still tariffs, and the real impact on industry is only beginning. Trade data in the coming months will reflect this, with China perhaps less affected thanks to its strong competitiveness. Over time, this will likely show up in RMB exchange rates.

Since the trade war, global equity markets have performed strongly. In the U.S., the rally was first driven by the AI narrative, then by rising expectations of rate cuts. In many other markets, capital flows reversed from the U.S. back home. Yet outside of China, there are few truly new growth drivers, and valuations remain stretched in many places. The risk of stagflation in the U.S. is not low, while in China, deflationary pressures, property weakness, and subdued consumption will take more time to digest. Current market enthusiasm risks underestimating these challenges.

On the booming trend of “On-Demand Flash Delivery” promoted by major food delivery platform, we agree with the view that efficiency without scale is meaningless. If high efficiency comes with excessive pricing, it becomes a burden rather than a moat，for both merchants and couriers. A moat exists only within reasonable pricing; at RMB 0.5 it may be sustainable, but at RMB 1.5 it may collapse. With AI advancing, many businesses may be restructured, with human work reshaped by AI programming and consumption patterns altered by AI agents.

We lean toward seeing the current rally as a reminder: This is a critical market that cannot be ignored in the long run. Its greatest allure lies in its ability to continually generate new industries, companies, and products. Beyond the cycles, there is relentless innovation. Viewing China’s economy and market only through the lens of old industries awaiting recovery may be as futile as “waiting for Godot.” What makes China unique is its capacity to offset the decline of old industries with the rise of new ones. That, ultimately, is the most beautiful aspect of this market.