The top of Japan's market capitalization ranking has changed again! From "storage as king" to "banking as king," funds return to the theme of cash flow amid the AI correction.

Japan's banking super giant, Mitsubishi UFJ Financial Group Inc, has soared to the top of the Japanese stock market's highest market value rankings. This is the first time that a bank stock has reached the top of the market value rankings since the formation of the so-called three major bank groups. This year, foreign capital has generally viewed the Japanese stock market as the "second battlefield of the artificial intelligence computing power infrastructure industry." NAND storage giant Kaili Knight briefly topped the Japanese market value rankings as a result, with the majority of companies in the top fifteen market value rankings focusing more on AI computing power infrastructure. However, the recent AI computing power theme has experienced a sharp and volatile pullback due to overcrowded positions and excessive leverage, causing Kaili Knight's market value to drop from the top spot to the third. According to institutional data, Mitsubishi UFJ, Japan's largest commercial bank, saw its stock price rise by 2.3% on Monday to a new high of 3,541 yen per share, reaching a market value of 42 trillion yen (approximately 259 billion US dollars). This surpasses the market value of Toyota Motor Corporation at about 41 trillion yen and NAND storage chip giant Kaili Knight at approximately 36.7 trillion yen. Another large commercial bank, Sumitomo Mitsui Financial Group, also saw a substantial increase in its stock price by 1.6%, also reaching a new record high. The rise of Mitsubishi UFJ Financial Group to the top of the Japanese market value rankings essentially marks a shift in the flow of funds in the Japanese stock market from "artificial intelligence computing power scarcity and high beta momentum" to "normalization of interest rates and profit certainty." On July 13, Mitsubishi UFJ's market value reached approximately 42 trillion yen, while Kaili Knight dropped to about 36.72 trillion yen after a single-day decline of 12.86%. With the Bank of Japan ending negative interest rates and raising the policy interest rate to 1%, loan rates are rising faster than deposit costs, leading to an expansion of banks' net interest margins. Mitsubishi UFJ's net profit for the 2025 fiscal year reached 2.4272 trillion yen, a record for the third consecutive year, with the 2026 fiscal year target further raised to 2.7 trillion yen, with approximately 170 billion yen expected to come from the rise in yen interest rates. Since the Bank of Japan ended its negative interest rate policy in March 2024, Japanese bank stocks have been on the rise primarily because large commercial banks can charge higher interest rates on loans. This contrasts sharply with the struggles of the banking industry in the past few decades. After the burst of the bubble economy in the early 1990s, banks were burdened with massive bad debts, and the aggressive monetary easing policy implemented to combat deflation made it difficult for them to raise loan rates. Since the first rate hike in decades in 2024, the Bank of Japan has raised interest rates four times, raising the policy interest rate to 1%. Rising interest rates allow banks to profit from the widening gap between loan rates and deposit rates, enhancing their profitability. Mitsubishi UFJ's analyst team's latest data shows that for every 0.25 percentage point increase in the benchmark interest rate, its annual net interest income will increase by 180 billion yen. The shift from Kaili Knight to the top spot reflects the transition of the AI computing power industry chain from "unlimited demand" to "verification of expected rate of change." Its stock price dropped by about 40% from the annual high of 112,700 yen on June 22 to 67,100 yen on July 13. On the day of closing on July 13, a number of AI computing power leaders in the Korean storage chip and other industries experienced a severe correction due to escalating geopolitical risks, rising oil prices, leveraged trade liquidation, and market doubts about the sustainability of high growth in AI profitability, causing the Korean stock market to plunge again and trigger circuit breakers. Kaili Knight's previous valuation included multiple optimistic assumptions such as NAND price increases, long-term supply shortages, explosive demand for AI inference, and the company's leading NAND flash technology, so even though the basic fundamentals of AI computing power infrastructure have not reversed, any excessive leverage or overcrowded long positions, slowdown in AI capital expenditure growth, adjustments in storage pricing or pace of AI application realization will result in a sharp compression of valuation. It is worth noting that the exchange of market values between the two companies does not mean "banks replacing technology" or "the end of the storage super cycle driven by AI computing power," but is instead a typical factor rotation: funds are shifting from momentum semiconductor stocks with overcrowded positions and high valuation dependence on future growth to financial stocks with higher cash flow visibility, benefiting from nominal interest rate increases, and able to strengthen shareholder returns. Kaili Knight trades the storage demand curve brought by AI inference in the coming years, while Mitsubishi UFJ trades the current net interest margin repair and cash profit growth already happening today; when market risk appetite decreases and discount rates rise, the latter naturally has an advantage. Kaili Knight briefly topped the Japanese market value rankings before due to the diffusion of AI demand from training to large-scale inference, causing storage bottlenecks to extend from high-bandwidth storage to high-capacity NAND. Enhanced retrieval generation, long-term context, vector databases, inference cache, and massive generative content all require larger non-volatile storage capacity than traditional enterprise computing; at the same time, storage chip manufacturers such as Samsung Electronics, SK Hynix, and Micron have previously prioritized investment in high-bandwidth storage and dynamic random access memory storage, leading to a relative lack of investment in NAND, creating a "suddenly accelerated demand, supply unable to respond in time" super cycle. Kaili Knight's share price had risen by more than seven times earlier this year, exceeding a market value of over 250 billion US dollars, as the market re-evaluated it from a cyclical consumer electronics storage manufacturer to a core supplier of AI inference infrastructure.