News linked to both this project and an event.
According to on-chain analyst Ai’s monitoring, address 0xf29…41244 has increased its TWAP short order for Changxin Memory (CXMT) to 310,000 tokens. If all are filled, they will be worth $3.78 million. Currently, he has opened a short position of 254,000 CXMT tokens at $6.8123 with 1x leverage, valued at $1.7 million. Together with the 310,000 tokens in the TWAP order, the total value will approach the current top position for Changxin Memory on Hyperliquid.
According to Hyperinsight monitoring, the CXMT mark price on Hyperliquid is quoted at $6.7274, down approximately 6.97% from the previous day, reaching a low of $6.62, with a 24-hour trading volume of approximately $6.837 million. As CXMT retreats, the largest short seller, a whale with an address starting with 0xf292, continues to maintain large-scale limit orders to add short positions on rebounds at higher levels.
Odaily reports, according to on-chain analyst Ai Yi's monitoring, an address has opened a $2.155 million TWAP short position on ChangXin Memory Technologies. If fully executed, this would become the second-largest position of Hyperliquid CXMT. The TWAP order size is 300,000 CXMT, with an opening average price of $7.1825. Currently, 1.14% of the order has been filled, and the account margin is $15 million.
Odaily "White Hair Stock God" Serenity posted on platform X, stating that due to the recent market downturn, his investment portfolio experienced a maximum drawdown of 49.4% this month. However, he still maintains his view on the long-term trend of the AI industry chain.Serenity revealed that his investment portfolio is mainly concentrated in key segments of the AI industry chain, including: semiconductor upstream, memory chips, photonics, humanoid robotics, and AI infrastructure-related companies. Because these areas typically have higher beta attributes, he previously used leveraged investments but has reduced the leverage level after the current round of market decline.Facing market skepticism towards AI-related assets, Serenity stated that recently a large number of investors have begun to believe: "AI is a bubble," "memory chips and the Korean KOSPI market are a bubble," "photonics is a bubble,""humanoid robots will not succeed," and "Neocloud (new AI cloud service providers) will eventually be replaced by hyperscale cloud vendors like Meta." Meanwhile, some retail investors and trading bots have even started advocating for "liquidating everything, the market will not recover."Serenity said he still believes these investment themes are supported by structural revenue growth and technological change. He experienced similar drawdowns in the past when global tariff risks impacted the market, and the market eventually rebounded. His investment horizon is long-term, allowing him to withstand higher volatility, and he will not change his long-term judgment based on short-term price fluctuations. Sharing this drawdown data is also to maintain transparency, allowing the market to see the real risks behind high-volatility growth investments.Serenity added: "If my prediction is that the revenue inflection point will come in the second half of 2027, and it is only 2026 now, then a decline of just a few weeks or months doesn't prove that the investment thesis has failed."
according to on-chain analyst Ai Yi's monitoring, a trader who went short on ChangXin Memory Technologies (CXMT) on Hyperliquid saw their position liquidated within 28 minutes, resulting in a loss of $44,000 from a position of approximately $220,000. Ai Yi noted that the liquidity of the CXMT contract on Hyperliquid is currently relatively low, leading to more volatile price movements. Data shows that a wallet with total assets exceeding $75 million is building a long position in CXMT.
On-chain analyst Ai Yi posted on platform X, stating that on the Hyperliquid order book, someone has placed a buy order for 6.824 million US dollars worth of ChangXin Memory Technologies (CXMT) at a price of 6 US dollars; if calculated at the current price of 6 US dollars, the market value of ChangXin Memory Technologies would be approximately 2.72 trillion yuan, with a total of 66.88 billion shares, surpassing the current market value leader Industrial and Commercial Bank of China (ICBC), which stands at 2.66 trillion yuan.
According to on-chain analyst Ai Yi’s monitoring, a significant price gap has emerged between SK hynix’s domestic Korean stock and its U.S. ADR. SKHX is priced at approximately $1,303.85, while SKHY stands at around $170.61, with the price difference reaching up to about 30.9%. This reflects strong demand from U.S. capital markets for leading AI chip and HBM (High Bandwidth Memory) assets. Analysis suggests that the conversion between the two is unidirectional—ADR shares can only be canceled and exchanged for ordinary Korean shares, and no one would do so when a premium exists. This explains why the price gap persists and cannot be arbitraged away.
According to CCTV Finance, the US stock memory chip sector encountered a collective correction after hitting a high in late June, with industry leaders such as SanDisk, Micron Technology, Seagate Technology, and Western Digital seeing stock price declines of over 20% in the past few weeks. The triggering factor was Meta selling computing power, which sparked market concerns about a surplus in computing power, while the core variable lies in whether the technical gaps between various AI large models will continue to narrow. Industry insiders point out that the memory chip industry has historically exhibited significant cyclicality—during booms, manufacturers collectively expand production leading to plummeting prices and industry-wide losses, followed by collective contraction in capital expenditure. Meanwhile, the industry's business model is undergoing profound changes, with cloud vendors and AI data centers increasingly signing 3-to-5-year long-term supply agreements with original manufacturers, including price ranges and minimum purchase volumes, to ensure the stability of key supply chains.
According to Odaily, "1011 Insider Whale" agent Garrett Jin pointed out in a post that there has been a clear change in market structure this week, with funds within the AI industry chain being reallocated.Change 1: Signs of a cyclical peak in Memory chipsHe stated that Micron's stock price faced resistance and fell back around the $1250 level. Despite earnings results exceeding expectations, the stock price is still declining on increasing volume, displaying typical top-forming characteristics of "weakening after good news is priced in."Concurrently, capital is rapidly flowing out of the memory chip sector. DRAM-related ETFs are experiencing declines on heavy volume, and SK Hynix and Samsung Electronics in the South Korean market are also weakening. Data shows that foreign investors have withdrawn over 100 trillion Korean Won (approximately $650 billion) from the South Korean stock market in the past two months.Change 2: Funds rotating towards AI HyperscalersHe noted that the real direction for absorbing this capital is not small and mid-cap AI concept stocks, but rather the core cloud computing giants represented by Google, Microsoft, and Amazon.Last Friday, when the chip sector came under pressure, GOOG and MSFT had already stabilized on increased volume, and this week META has further strengthened this trend by rallying on high volume.Garrett Jin believes the logic behind this capital migration is the "token optimization trend." As more simple tasks are handled by low-cost models, value will gradually concentrate on the token-based billing cloud services and orchestration layers, rather than the foundational model layer. This also forms the core moat for hyperscale cloud providers. The current strategy should focus on catching up opportunities in hyperscale cloud names.
: In response to the community debate over "Micron VS Nvidia," Jukan, an analyst at Citrini Research, posted on platform X, stating, "MU (Micron) may not be Nvidia, but its future importance could surpass Nvidia. Inference is now directly tied to revenue, but improvements in inference performance cannot be achieved simply by adding more Nvidia GPUs. In inference, GPUs are often underutilized and idle due to memory bottlenecks. For inference, increasing memory provides higher value. The return on investment for inference ultimately depends on memory, not GPU. Therefore, why are people still fixated on obtaining Micron through Nvidia's framework? One must think more comprehensively. Inference is memory."
According to public information, the Roundhill Memory ETF (DRAM) has included GigaDevice in its holdings, with a weight of 2.91%. Jukan, an analyst at Citrini Research, stated that GigaDevice is regarded as the parent company or an affiliated entity of China's DRAM manufacturer CXMT. The Roundhill Memory ETF primarily focuses on the global memory chip supply chain, with core holdings including major global memory chip companies such as Samsung Electronics, Micron Technology, and SK hynix.
Odaily News, Bloomberg Senior ETF Analyst Eric Balchunas stated on the X platform that the pure memory chip exchange-traded fund DRAM (Roundhill Memory ETF, launched by Roundhill) has seen its assets under management (AUM) surpass $6.5 billion just 36 days after its launch, setting a new all-time record for ETFs and exceeding the previous record of 43 days held by BlackRock's spot Bitcoin ETF, IBIT.Eric Balchunas noted that DRAM surged 13% in a single day last Friday while attracting approximately $1 billion in inflows, a key driver behind its rapid asset expansion. He described this performance as "stunning" and stated that regardless of future trends, it is one of the most "perfectly timed" ETF launches on record.