A new stock king has arrived in the A-share market. Domestic memory chip leader ChangXin Memory Technologies (also known as CXMT) made its debut on the STAR Market, rewriting a series of records on its first trading day. Its opening market capitalization exceeded 3.3 trillion yuan, making it the first A-share tech stock to break the 3 trillion yuan market cap mark at the open. It claimed the top spot by market value on the STAR Market and surpassed the "universal bank" to become the largest company by market capitalization in the entire A-share market. Within the first hour of trading, turnover exceeded 100 billion yuan, with the full-day volume setting a new A-share record. The trading was so intense that many investors reported lagging issues when placing or canceling orders. Based on the opening price, one lucky subscriber could make a profit of over 20,000 yuan on a single lot.

It has been a long time since a new stock captured the market's entire attention. However, more than the lucky winners, the market is buzzing about another story. In 2021, Country Garden's venture capital arm participated in a Series C funding round for CXMT, investing 900 million yuan for a 1.56% stake. Looking back today, this investment seemed like a textbook example of perfect foresight. However, due to various reasons, all of this equity was subsequently exited. If calculated based on CXMT's closing market value of 3.28 trillion yuan, that stake would now be worth over 511 billion yuan, meaning Country Garden missed out on a potential profit of 491 billion yuan. To put this staggering figure into perspective, the total market capitalization of Country Garden's Hong Kong-listed shares is about 82 billion yuan. This means the missed profit from this single investment is more than six times the entire current value of Country Garden.

Where to start

This is the harshest reality of the capital market. Often, the challenge isn't buying the right stock, but holding onto it; it's not about failing to understand the opportunity, but a lack of patience to wait for it to mature. Back in 2021, no one could have predicted that a domestic memory chip company, still in its investment phase, would grow into a 3-trillion-yuan tech giant today. Similarly, no one foresaw that the once-glorious property giant would find itself in its current predicament just a few years later. When the era shifts, the landscape of wealth is inevitably redrawn.

Why just 10 ASX 200 shares?

The significance of CXMT's listing extends far beyond a single company entering the public market. According to its prospectus, CXMT has become China's largest, most technologically advanced, and most complete DRAM company in terms of R&D, design, and manufacturing integration. Its production capacity ranks first in China and fourth globally. As of the fourth quarter of 2025, its global market share by DRAM sales has reached 7.67%. More importantly, domestic memory chip production is finally beginning to compete on a global stage. For decades, the global DRAM market has been dominated by three companies: Samsung, SK Hynix, and Micron. This landscape is now being disrupted. Consequently, market expectations for CXMT are rising rapidly. Nomura Securities, in its first coverage of CXMT, gave a target price of 116 yuan, corresponding to a market value of approximately 7.76 trillion yuan. This valuation is more than double that of Micron.

Nomura cited several reasons for this premium valuation. First, "global memory supply is unlikely to be loose in the coming years," as the capital expenditures of the three major memory manufacturers have shifted significantly towards HBM, structurally suppressing new supply for general-purpose DRAM. Second, CXMT's market share growth could far exceed market expectations, with an implied long-term market share potentially reaching 25% to 30% or even higher. Third, Chinese cloud computing companies and smartphone manufacturers are continuously increasing their procurement of domestic DRAM. Combined with the explosive demand from AI servers, which require nearly 80 times the DRAM per server compared to a smartphone, CXMT is poised to enjoy a higher valuation premium. Nomura projects that the company's sales and net profit attributable to the parent company will grow by 63% and 74% respectively, with 2026 potentially just the beginning. This suggests that the market is not just betting on CXMT, but on the entire domestic memory chip supply chain.

Meanwhile, the Korean market continues to experience major volatility. Korean companies have just announced a massive semiconductor cooperation plan worth 950 billion USD. Samsung Electronics has signed a 200 billion USD cooperation agreement with Broadcom for the next five years, while SK Group plans to advance a total of 750 billion USD in high-end memory cooperation with global tech companies like Nvidia. At the same time, the National Pension Service of Korea has begun to rebalance its portfolio. After a cumulative net sell-off of nearly 8.7 trillion Korean won in Korean stocks over the past six months, it turned to net buying in July, with a focus on increasing its holdings of SK Hynix. It net purchased 425.8 billion Korean won of SK Hynix shares in July alone, ranking first for the second consecutive month. Today, the Korea Composite Stock Price Index closed up 0.97%, with SK Hynix closing up 3.24% after briefly falling nearly 3% during the day, and Samsung Electronics up 1.8%.

This reflects a new reality. The market is no longer concerned with how many cooperation agreements are announced or how large the orders are. The focus has shifted to a single question: when will these orders actually translate into realized profits? The real pressure is now falling on global tech giants. Recently, prominent short seller Michael Burry has increased his bearish bets on AI stocks. His latest portfolio disclosure shows he has added to his short positions in Micron Technology, Nvidia, and the Philadelphia Semiconductor Index ETF. He also initiated a new short position in Caterpillar and maintained his short positions in Tesla and Palantir. Burry believes that amidst the current frenzy of AI infrastructure investment in US stocks, a significant portion of capital expenditure does not correspond to real end-user demand but instead circulates through opaque financing structures, creating a systemic risk of overvaluation. While AI revenue is indeed growing, costs are growing even faster.

A media outlet compiled a set of data based on LSEG consensus estimates. The five largest global cloud computing companies—Microsoft, Alphabet, Amazon, Meta Platforms, Inc., and Oracle—are expected to see their combined capital expenditure exceed their free cash flow for the first time by 2027. The operating cash flow of these five companies is projected to increase by 340 billion USD compared to 2025, but their capital expenditure is expected to increase by 534 billion USD. In other words, for every additional 1 USD in operating cash flow generated, they would need to invest 1.57 USD in capital expenditure. If this trend continues, within two years, these five of the world's most profitable tech companies may see their operating cash flow unable to cover their capital expenditure, forcing them to rely on external financing to continue their expansion. This is the core paradox of the AI era. Everyone knows AI represents the future, so no one dares to stop. When one company increases its spending, competitors must follow suit; when one builds a data center, another must build a larger one; when one purchases more GPUs, others must continue to place larger orders. No one is willing to be the first to apply the brakes, as stopping could mean being left behind for years. But when everyone pushes the accelerator at the same time, it leads to ever-increasing capital expenditure and a shrinking safety margin for cash flow. This is precisely why the renowned short seller Burry dares to continue betting against US tech stocks, going against the prevailing market sentiment.