硬件 · 3810

Stanley Druckenmiller buys Bitdeer and Hyperliquid Strategies shares for $87.8 million

According to news, Duquesne Family Office founder Stanley Druckenmiller bought 4.1 million shares of high-performance computing company Bitdeer Technologies Group (BTDR) in the second quarter, with a position value of over $64.7 million, with an average purchase price of $12.26. The company produces cryptocurrency mining hardware and operates data centers in the US and beyond. Additionally, Druckenmiller bought 2.9 million shares of HYPE digital asset treasury company Hyperliquid Strategies (PURR), holding positions worth $23.1 million, and indirectly gained HYPE exposure. Hyperliquid Strategies aims to provide US and institutional investors with HYPE token-related investment channels. Druckenmiller's operation is similar to Jane Street and Citadel's increase in BTDR over the same period. Jane Street currently holds shares worth more than $112 million in BTDR. BlackRock, State Street, and Citadel also increased their PURR holdings in the second quarter; HYPE previously rose to record highs due to related compliance developments.

12h ago

Tencent's chip leader Gao Jianlin left his job and started a business to target the RISC-V high-performance AI CPU circuit

According to MaxForAI, according to MaxForAI, the core head of Tencent's chip research and development, has recently left Tencent and started a business. He plans to develop high-performance CPUs based on the RISC-V architecture around high-performance AI servers and agentic AI (intelligent AI). Gao Jianlin is regarded as one of the early core promoters of Tencent's self-developed chip system. According to data, he formed an FPGA hardware team within Tencent in 2013, began setting up AI chip research and development in 2018, established the Penglai Laboratory in 2020, and promoted Tencent's development of various AI chips and data center deployment. This startup focuses on CPUs rather than the currently competitive AI GPU market. According to the report, Gao Jianlin believes that with the rapid development of Agentic AI, the AI inference process will involve model call, tool execution, search, database interaction, and large-scale task scheduling, and the CPU will assume a more important scheduling and control role in the AI system. According to reports, Gao Jianlin was involved in RISC-V related research and development during his time at Tencent, and participated in various technical directions such as chip architecture, verification, and back-end. Its new company plans to build high-performance server CPUs based on the open instruction set RISC-V to enter the AI infrastructure market. Currently, the name of Gao Jianlin's new company, financing conditions, and specific product launch dates have not been disclosed. The market is concerned about whether it will become another emerging force in the field of AI chips in China targeting server CPUs and smart body infrastructure.

12h ago

Anthropic hires Google chip business executives to speed up the deployment of self-developed chips

According to news, Anthropic has hired Amir Salek (Amir Salek), one of the founders of Google's custom chip project under Alphabet (GOOG.O). As this artificial intelligence lab paves the way to enter the field of self-developed semiconductors, Salek joined as part of its hardware layout. Anthropic said on Friday that Salek will join the AI company's computing team. Salek was previously responsible for Google's Tensor Processor (TPU) business until he left office in 2022, and led the delivery of the first seven generations of TPU chips. Anthropic currently purchases chips from multiple vendors, including Nvidia (NVDA.O), Google, and Amazon (AMZN.O). However, the company has recently released signals that it wants to establish its own internal chip business. San Francisco-based Anthropic has begun recruiting people for this project and posting relevant positions.

17h ago

BounceBit Chain updates vulnerability attack progress: will permanently shut down the chain and migrate to BNB Chain

Comparatively, cross-chain revenue protocol BanceBit issued a security incident announcement stating that its blockchain network was attacked by a protocol-level vulnerability attack between 8:02 UTC on August 19 and 01:54 UTC on August 20. The attackers used authorization flaws in the Evmos underlying architecture to transfer BB tokens from 9 main network accounts without the authorization of the account owners. According to the announcement, the attackers transferred a total of approximately 286.5 million BBs through 14 transactions. The impact of the incident was limited to BanceBit Chain itself, and did not involve private key leaks, signature forgery, wallet, hardware devices, or exchange account security issues. BanceBit CeDeFi Strategy, Promo Vaults, Prime, and RWA products were not affected. BounceBit stated that the vulnerability stemmed from a protocol native module authorization verification flaw in the Evmos architecture. When calling the relevant module through a smart contract, the attackers bypass security checks that should verify the authorization relationship of the fund source account, making it possible to specify any account as the source of funds. After the incident, BounceBit Chain stopped generating blocks at block height 20,702,857, then the team decided not to upgrade the chain, but to permanently shut down Bouncbit Chain and re-issue BB as an BEP-20 token based on BNB Chain. BounceBit stated that the new BB token supply will be based on an on-chain snapshot before the first abnormal transfer (block height 20,697,260), and the 286,543,148 BBs transferred by the attackers will not be included in the new token balance. Users do not need to submit an application or migrate their wallets, and the official plan is to automatically distribute the new BB to the corresponding BNB Chain addresses. Regarding the BB in the pledge, BounceBit said it will be restored as soon as the snapshot is in time, and there is no need for coin holders to perform unbundling or redemption operations. Currently, BounceBit has submitted requests for suspension and assistance to relevant exchanges, and reminds users to be wary of scams and not to click on any BB migration or receipt links that have not been officially confirmed. The team said that the new BEP-20 BB contract address and reissue progress will be announced later. This article is sponsored by GENG, Build Your Fortune on GENG (https://geng.one)

1d agoburnking

US Stock Value Investing Is Heading Into Another Trap

Source: Shenchao TechFlow Original title: (Opinion: Value investing in US stocks is not equal to fundamental investment) When “fundamentals are dead” becomes a consensus, investors who blindly organize giants will eventually experience astonishing capital destruction. Guide: When the market shouted “fundamentals are dead” and the capital frenzy formed a group of tech giants, the author used an astronomy discovery to unravel the logical loopholes behind this narrative. Starting from the composition of valuation multiples, this article reminds investors to distinguish between the true quality of an enterprise and the premium that the market is willing to pay. It is particularly cautionary about long-term allocation in the crypto and technology sector. I promise this introduction won't be as long as the last one on the weather. But please give me 90 seconds. More than 100 years ago, a woman named Henrietta Levitt was doing the tedious job of measuring the brightness of thousands of stars on photographic negatives (the way they were imaged before film appeared). She noticed one characteristic of a class of pulsating stars: the slower they pulsate, the brighter they themselves are. ¹ This might just seem a little interesting today, like “OK, that's pretty cool.” But at the time, astronomers couldn't tell the difference between a dark star very close to Earth and a very bright star far away. For them, the two left the same stain on the photographic film. Visual brightness is a messy mix of these two variables: how bright the thing itself is, and how far away it is from us. Henrietta's work decouples these two things: if you can observe the rate of pulsation, you can know its true luminosity; if you know its true luminosity, you can reverse the distance based on how dark it looks. Astronomers call it “standard candlelight.” A few years later, a man named Edwin Hubble discovered one of these pulsating stars, applied Levitt's math, and discovered what he had always thought was a cloud of gas within our galaxy; in fact, it was an entire independent galaxy, one million light years away. So in simple terms, the observable universe has grown about a trillion times larger, just because one person has figured out how to tell the difference between what things look like and what they actually look like. That in itself is obviously pretty cool. But another interesting thing is that around the same time period, two other astronomers each independently drew a scatterplot. One axis was actual luminosity, and the other axis was temperature. They discovered that stars are not randomly distributed in this space, but rather clustered into different families. The meaning behind this is: stars with the exact same visual brightness may and do belong to a completely different family, have a completely different past, and most importantly, have a completely different future... So what is written in the star? Over the past few years, there has been much discussion about markets, narratives, capital, company building, and financial nihilism. This feeling seems to have reached a feverish climax as the tech and financial world begins to face a very different future than a few decades ago. What is particularly clear is that separating progress from asset prices has become more noisy and in many ways more repulsive. But as an investor who makes a living by buying assets that (hopefully) outperform, a simple framework is: forward returns are roughly equal to growth in fundamentals multiplied by changes in valuation multiples (and multiplied by the dividends you've collected along the way). In this case, the valuation multiplier can very cleanly correspond to the smudges on the photographic film. It's an observable data point, but it entangles two things that the market can't directly see: how good the company actually is, and how far (or how long) its future cash flow is now. I think most of the money that can be made comes from investors who are most capable of unraveling these two variables earlier than others (or “perception of differences”), and we will continue to see astonishing capital ruin for investors who treat their stains as stars. Value investing is not equal to fundamental investing. I think there is a misunderstood view: fundamental investing has historically dominated the creation of excess returns. Most of these legends come from the Graham, Buffett, and Tiger Foundation lineage, as well as numerous narratives built around this group of people. It is believed that by some point in the 2000s, this approach was no longer effective, and anyone who invested in this way was overwhelmed by momentum, trends, and “direct buying tech giants.” The conclusion was (and still is?) It's “fundamentals are dead.” ² The modern version of “fundamentals don't matter” itself isn't stupid. It's rooted in a lot of ideas that many of us on the Compound team have written before. The biggest companies get the most mechanical purchases, and the software industry has a winner-take-all economic law. AI means that giants can transform scale into moats faster than challengers, and there are also reasons why the market's microstructure embeds momentum more deeply into our market infrastructure. These are all real...

1d ago深潮TechFlow#US stocks

DingTalk launches AI office app QwenNote, hardware QwenNote A2 revealed

Comparative news, according to the “Reading Best” report, DingTalk is promoting a new AI office application QwenNote (Listen to a Thousand Questions). The application is positioned as an AI portable assistant. It integrates real-time voice transcription, summary, and translation through a combination of software and hardware, and is deeply integrated with AI Agent to embed Agent capabilities into voice input and promote a shift from simple recording to automated execution. The application supports real-time transcription and Chinese-English bilingual recognition and language switching, can generate structured meeting minutes, outlines, and to-do tasks, and has built-in AI Q&A and shortcuts based on memorized materials. QwenNote provides a voice memo function. You need to scan the code to connect to the recording device. Press and hold the button on the back of the device to record inspiration. After the recording is completed, it automatically files, generates a title and brief summary, and marks the time. The product also has an incognito protection mode. After opening, the original audio will be physically deleted, and only the transcribed text will be kept to suit confidential scenarios. The supporting hardware QwenNote A2 has been introduced in the app. It also belongs to the Thousand Question Listening hardware ecosystem. The ecosystem also includes DingTalk A1, DingTalk A1 Pro, Cleer H1, etc., and users can scan the code to complete the binding. According to the report, DingTalk hopes to use the integration of software and hardware to complete offline voice collection portals to form a closed loop of live recording, real-time bilingual transcription, AI minutes Q&A, and DingTalk organization collaboration. The listening material can be synchronized to DingTalk AI to listen to and support personal private isolation. The software side continues to embed large models in scenarios such as documents, meetings, and IM, while the hardware side expands the Thousand Question Listening product line. More official release information has not been widely publicly retrieved for related hardware.

2d ago

Global equity and debt double kill: 30-year US bond yields soared to 5.33%, and the AI industry chain suffered a severe setback

Comparing news, the global financial market has experienced sharp fluctuations. The European, American, Japanese, and South Korean markets have experienced a double slump in stocks and bonds, and the sharp rise in long-term US bond yields has become the focus of market attention. In the Asian market, the sharp decline in the Korean stock market triggered a trading mechanism. The KOSPI index closed down 5.8%, SK Hynix fell nearly 10%, Samsung Electronics fell more than 8%; and the Nikkei 225 index closed down 3.16%. More than 5,000 A-shares fell, leading the decline in AI industry chain sectors such as semiconductors, computing power hardware, PCBs, memory, and CPO. In terms of US stocks, the three major indices fell for the third consecutive trading day, and the Philadelphia Semiconductor Index fell nearly 5% in a single day. Shares of AI-related companies such as Micron Technology, Western Digital, SanDisk, Marvell, AMD, Intel, Coherent, and Credo experienced a sharp correction in stock prices. According to market analysts, the core trigger for this round of sell-off comes from the rapid rise in global long-term bond yields. The 30-year US Treasury yield rose to 5.33% intraday, a record high since 2007. Meanwhile, French, German, British, and Japanese long-term treasury yields have all risen to multi-year highs, and global long-term capital costs are being repriced. According to data from the US Treasury Department, the amount of US debt held by overseas investors fell to 9.299 trillion US dollars in June, a decrease of about 72 billion US dollars compared to May. Among them, Japan's holdings fell to 1.116 trillion US dollars, reducing their holdings by 26.4 billion US dollars in a single month; UK holdings fell to 939.9 billion US dollars. As risk-free interest rates rise, the market is beginning to re-evaluate the AI industry's high capital investment model. Investors are concerned that continued expansion of data centers, GPU procurement, and infrastructure construction will require significant financing, and that higher capital costs may reduce future cash flow estimates for technology companies. Currently, the market is concerned about three major variables: 30-year US Treasury yield trends, the Federal Reserve's judgment on long-term interest rate paths, and whether subsequent earnings reports from tech giants such as Nvidia and Broadcom can verify AI return expectations. The US Treasury Department will issue 16 billion US dollars of 20-year US bonds at 1 a.m. Beijing time on Thursday. The market will pay close attention to the results of this issuance.

3d ago

Nomura covered YuShu Technology for the first time to give a purchase rating. There are advantages in cost, market share, and iteration speed

According to comparative news, Nomura Securities included Yuju Technology for the first time in its report today and gave a purchase rating. Nomura's core judgment is that Yuju Technology has moved from a technology demonstration company to a large-scale delivery company. According to the report, the company shipped more than 5,500 humanoid robots in 2025, ranking first in the world; the compound revenue growth rate from 2026 to 2028 is expected to reach 122%. In the context of AI applications extending from software models to the physical world, robot companies that can continuously ship, reduce costs, and iterate will be more likely to obtain valuation premiums. The cost advantage is the first main line in Nomura's view of Daewoo. According to the report, Yushu is highly self-developed in key hardware such as motors, speed reducers, drivers, encoders, lidars, and power management, and outsourced components account for only about 10% to 20% of the total cost. This allows the company to launch products at lower prices while maintaining a high gross margin. The gross margin of the company's main business rose to about 60% in 2025, and the gross margin of humanoid robots reached 63.2%. Rapid iteration is the second main line. According to Nomura, Uki launched 4th generation humanoid robot products within 26 months, from H1 and G1 to R1 and H2. The product matrix covers consumer, scientific research, education, and industrial applications. Low-cost products will bring more shipments, and more shipments will also form real physical interaction data to further train models and algorithms. This cycle will be the key for the company to expand its leading edge. At the same time, the report warned that the risks faced by Yu Shu are not low. US regulatory restrictions will affect the entry of new models into the US market, which contributed approximately 13.3% of its revenue in 2025. More importantly, the current demand for humanoid robots is still dominated by scientific research, education, entertainment, and government procurement, and industrial and commercial applications still account for a low proportion. Nomura believes that what we really need to observe in the future is whether industrial customers can form repeated orders, which will determine whether demand for humanoid robots can move from themed investment to actual volume. In terms of financial forecasts, Nomura expects Uki's revenue from 2026 to 2028 to reach 2,687 billion yuan, 5.396 billion yuan, and 13.184 billion yuan, respectively, with year-on-year growth rates of 58%, 101%, and 144%. In terms of valuation, Nomura used the 2027 forecast market sales ratio of 25 times, which is significantly higher than the average of the Chinese robotics industry. The reason is that Uki has higher growth, more pure business attributes, and the scarcity of already achieved profits.

3d ago

$21 billion AI chip upstart Etched questioned: performance has not been verified by a third party so far

Comparing the news, AI News, Etched has caught the attention of the chip community after having just completed $700 million at a valuation of 21 billion US dollars. The Tiny Corp, an AI computing team founded by famous hacker George Hotz, the team behind tinygrad, an open source deep learning framework, publicly questioned Etched's technical propaganda: there are many photos of investors, orders, and hardware, but too little data to actually verify performance. One of Etched's core selling points is LVI, which allows the chip to run AI inference at lower voltages. Etched claims that this allows the trillion-parameter sparse MoE to reach over 80% of its theoretical peak computing power. Chip design practitioner Wesley Yue questioned that a high ratio does not mean absolute performance is strong. MFU (model computing power utilization) measures the ratio of actual computational power to the theoretical peak. If the chip itself has lower peak computing power, even if the utilization rate reaches 80%, it may not be able to outperform its rivals. Etched has yet to disclose full FLOPs, power consumption, and third-party benchmarks. The official website still only writes that early customer tests have reached the leading level, and that detailed performance data will be published later. However, there is currently no evidence that Etched was a fraud. The Wall Street Journal and Reuters have both confirmed that their chips have been shipped. Jane Street got its first complete rack last month, and deployment has already begun. The biggest question now is not whether there is a chip or not, but whether this chip has been advertised that well.

3d ago

Nomura covered YuShu Technology for the first time and gave a 25 times market sales rate estimate for 2027

According to comparative news, Nomura Securities covered YuShu Technology for the first time and gave an estimate of 25 times the 2027 market sales rate. The revenue growth rate for the 2027/28 fiscal year is expected to reach 101% and 144%, respectively, and the compound annual growth rate of revenue from 2026 to 2028 is about 122%. The research report points out two major investment logics: full-stack mechanical design and self-developed hardware gave Yushu a structural cost advantage; outsourced components accounted for only 10% to 20% of the total cost, and gross margin increased from 44% in FY2022 to 60% in FY2025; rapid product iteration enabled Yushu to repeatedly lead the industry into new application scenarios and launch four humanoid product lines within 26 months. The research report also suggests that on July 28, 2026, the US FCC added it to the Covered List, and the models that have been authorized for sale can still be sold, but the newly developed models will be restricted, which will constitute structural entry restrictions. Currently, 73.6% of humanoid robot revenue still comes from research institutions, and catalysts for growth in the new terminal market are still scarce. Nomura suggests focusing on “brain” advances—the commercialization of WVLA2.0 and the industrial deployment of UniForm-X1-0, which may be the key to unlocking new application scenarios. The opinion in this research report is Nomura Securities research and analysis, and does not constitute investment advice.

3d ago