“AI Prophet” holdings explode Taiyo Yuden's stock price skyrocketed! Situational awareness breaks the lever, yet unravels one of the key bottlenecks in AI infrastructure—MLCC

The Zhitong Finance App learned that according to a Japanese market disclosure document, Leopold Aschenbrenner, the youngest hedge fund leader with the title of “AI Investment Prophet,” frantically increased his position on MLCC supergiant Taiyo Yuden (Taiyo Yuden) from Japan last month, holding over 16% of the company's shares at the peak. According to reports, after disclosing information showing that the hedge fund Situational Awareness, which Aschenbrenner is at the helm and managed throughout the process, once held up to 16.6% of the shares in this Japanese electronic component manufacturer, Taiyo Yuden's stock price surged 7.5% during the Tokyo trading session, highlighting that the fundamental demand for AI computing power was not falsified due to capital bursts, and that capital was further allocated to the MLCC value chain.

These disclosure documents revealed this investment, which was previously unknown to the outside world. The hedge fund is led by former OpenAI researcher Leopold Aschenbrenner. However, there is still huge uncertainty about whether the fund still holds shares in Taiyo Yuden.

Taiyo Yuden's stock price surged on August 12. The more reasonable explanation was not “Situational Awareness bought back today,” but the market suddenly discovered: ** A fund with extreme knowledge of the AI industry chain used Taiyutai TV as a highly concentrated AI core position, and the huge supply previously forced sell-off may have been absorbed by Citadel, which took over most of Situational Awareness's stock positions. Situational Awareness is likely to sell shares to Citadel at a loss, and as a result, the market has formed an expectation: if they can buy Taiyo Yuden shares at a price close to Citadel's takeover position, they should be able to continue to make profits.

Leopold Aschenbrenner's brilliance comes from directly transforming a set of highly penetrating technical judgments into capital market positions: he graduated from Columbia University in first place at age 19, joined the OpenAI “Super Alignment” team, and published 165 pages of “Situational Awareness” in 2024, suggesting that AGI will force the world to purchase data center power equipment, liquid cooling, data center CPUs, DRAM/NAND/HBM, data center optical communication and optical interaction at an exponential pace AI data center delivery process with a complete chain including high-performance Ethernet network infrastructure, transformers, gas turbines, and data center energy storage systems.

The fund of the same name was then established to attract well-known capital such as Jane Street and the founder of Stripe, and its management scale exceeded 20 billion US dollars within two years; by the end of June 2026, the fund's net return during the year had reached an astonishing 439%, and the return since inception was over 1000%, making it the most legendary concentrated bettor in this round of AI bull market, and was even praised by some retail investors as an “AI investment prophet” and “the version answer under the AI superinvestment boom”. However, as the wave of stock sell-offs closely linked to the global stock market and AI computing power infrastructure intensified in July, the agency's record loss of 67% and the collapse of high-leverage AI betting positions in July forced the fund to sell most of its open market holdings to hedge fund giant Citadel, headed by Wall Street billionaire Ken Griffin, and completely remove all leverage.

How sacred is solar induction?

Taiyo Electric produces multilayer ceramic capacitors (MLCCs). These products are widely used in smart phones, electric vehicles, high-performance industrial equipment, and AI server clusters. After news of Situational Awareness positions came to light, Taiyo Yuden once became the best-performing constituent stock in the Nikkei 225 Index in a single day.

However, the relevant reporting documents also show that the hedge fund has been reducing its holdings. According to the agency's review of disclosure documents, Situational Awareness sold 185,200 shares on July 28, then 1.7 million shares on July 30, and sold 14.6 million larger shares through OTC transactions on August 3.

This made the ownership of the remaining 4.4% of the shares unclear. MarketWatch's team of analysts said that this portion of the shares may still be held by Situational Awareness, transferred to Citadel, or eventually held by another entity.

This point in time is particularly noteworthy, mainly because Citadel acquired Situational Awareness's open market equity portfolio on July 30, which enabled the hedge fund to meet additional margin requirements.

These latest disclosures also highlight that although Taiyo Yuden's market popularity is relatively low in global AI computing power topics, it is still attractive to investors. The company's price-earnings ratio is approximately 46.8 times, according to FactSet data.

MLCC can be described as an “AI computing power industry bottleneck” link that is easily underestimated in the unprecedented expansion of AI computing power infrastructure demand. The power consumption and transient current of GPUs, TPUs, and custom ASICs continue to increase, so that AI server power supply systems must stabilize core voltage, absorb sudden load changes, and suppress ripple and noise within a very short time scale. Therefore, it is necessary to configure a large number of high-capacity, low-voltage, and miniaturized MLCCs near accelerators, voltage regulation modules (VRMs), and packages/motherboards as decoupling and local energy storage devices.

As power density increases, the power supply path must also move closer to the chip, or even to backplane/vertical power supply and built-in capacitors on the board. Taiyo Yuden officially stated that high-current power supply, high-density installation, and reduced circuit loss for AI servers are driving the increase in the number and performance requirements of MLCCs, and has commercialized high-capacity embedded MLCCs for AI servers. According to a recent TrendForce research report, the volume of Google TPU, AWS Trainium, Meta MTIA, and next-generation Nvidia VERA-RUBIN GPU platforms has caused a sharp rise in demand for high-end MLCCs. Murata, Samsung Electro-Mechanics, and Taiyo Electric's BB Ratio (that is, the “order shipment ratio” used to measure the strength of supply and demand) rose to 1.30, 1.31, and 1.25 respectively at the end of June, the highest since the industry entered the pandemic One of the tight order cycles.

MLCC has risen from a common consumer electronics component to one of the biggest winners in the AI infrastructure boom

In the context of the global AI investment boom since this year, the MLCC investment boom was once more intense than the AI-related semiconductor sector surrounding leaders in the chip industry chain such as Nvidia, AMD, Micron, and SK Hynix.

On June 26, Taiyo Yuden had already surged 438% during the year, and Murata rose 268%. Even Taiyo Yuden clearly surpassed Nvidia supplier Ibiden's increase of about 292% over the same period, making it one of the fiercest AI second-order derivatives transactions in the Nikkei Index. However, what really supports this fervor is not just thematic imagination: Taiyo Yuden's capacitor orders in the latest quarter increased 41% month-on-month, and the capacitor business BB Ratio directly rose to 1.72, compared to 1.31 in the previous quarter; the company also raised the operating profit guideline for the fiscal year ending March 2027 to 45 billion yen, an increase of about 125% year over year. In other words, the core of the market's crazy trading in the past is actually a very typical AI supply and demand mismatch model — demand for high-spec MLCCs required for AI servers is growing much faster than the expansion rate of high-end production capacity, resulting in common upward elasticity of orders, product portfolios, ASP, and profit margins.

A hedge fund with an extreme understanding of the AI industry chain once used Taiyao TV as a highly concentrated AI core position, while the huge supply of forced sell-offs previously may have been absorbed by Citadel or other institutions. Situational Awareness, a subsidiary of Aschenbrenner, had previously made impressive gains in the AI stock bull market, but used about 3-4 times leverage in open stock trading; after a sharp retracement of AI assets in July, it triggered margin pressure, and was eventually forced to sell most of its public stock portfolio to Citadel.

The real important signal for investors who focus on the topic of AI computing power is not to “follow the AI prophet to buy solar electricity,” but rather that this liquidation completely exposed MLCC, an AI hidden supply chain, to global capital. Only the latest Japanese regulatory documents let the market know that its previous holdings in Taiyo Yuden had reached a maximum of 16.6%. Furthermore, the market is increasingly convinced — if you can buy Taiyo Yuden shares at a price close to Citadel's takeover position, you should be able to continue to make profits.

Situational awareness's failure was first and foremost a failure in leverage and liquidity management, rather than the AI server MLCC demand logic being falsified: Taiyo Yuden's stock price had plummeted by more than half from its July high, but orders accelerated, BB ratio rose to 1.72, and profit guidelines were revised during the same period. This kind of “stock price deleveraging and continued strengthening of industry fundamentals” is critical.

According to Wall Street giants who are optimistic about the topic of AI computing power, the AI bull market is far from over, but is moving from the “AI chip purchase frenzy” to the second stage of “large-scale construction of AI factories” — that is, the next round of excess alpha earnings will no longer only belong to the list of the strongest leaders in the AI GPU/AI ASIC field, but will also spread systematically to data center high-performance CPUs, DRAM/NAND/HBM storage, AI PCBs, liquid cooling systems, data center optical interconnect systems, ABF carrier boards/glass substrates, MLCCs, electronic cloth, and extensive foundry processes” An “AI factory” level full-stack AI computing power infrastructure layer.