第四款是Intel 酷睿 i5-12500H，基于12代混合架构设 计，配备12核16线程（4个性能核+8个能效核），P核最 高睿频4.5GHz，E核可达3.3GHz，三级缓存达18MB，确 保数据吞吐畅通无阻。TDP为45W，最大睿频功耗 95W，适合高性能轻薄本或小型游戏本使用。集成Iris Xe核显，拥有80个执行单元，动态频率1.3GHz， 支持四屏输出与DirectX 12.1，无论是多屏办公还是影音娱乐都能轻松应对。支持DDR5、LPDDR5等多 种内存规格，最大容量64GB，未来升级路径清晰，是一款兼具性能与前瞻性的全能型处理器。 夜深人静，键盘轻响，散热风扇轰鸣如战车启动，超频玩家的世界从不追求平凡。他们要的是极致性 能、精准调度与无限潜力，哪怕只是0.1GHz的频率提升，也足以点燃整晚的激情。在笔记本有限的空 间里，CPU不仅是大脑，更是灵魂——决定着帧率、响应速度、多任务流畅度乃至AI运算能力。当一款 处理器既能压得住日常负载，又能在极限时刻爆发出惊人能量，这才是真正属于发烧级玩家的选择。今 天推荐的这几款笔记本CPU，不仅定位高端、性能强劲，更关键的是——它们全部"到手价0.0元"，意 ...

Core Insights - AMD has unveiled its latest generation of mobile and desktop processors, the Ryzen AI 400 Series and Ryzen AI PRO 400 Series, which enhance AI capabilities, gaming performance, and commercial features across a wider range of systems [2][3][4] Product Launch - The new AMD Ryzen AI 400 Series includes processors designed for Copilot+ PCs, while the Ryzen AI Max+ processors target premium ultra-thin notebooks and small form-factor desktops [3][4] - The Ryzen AI PRO 400 Series focuses on AI acceleration, modern security, and enterprise-class manageability for business laptops [3][7] AI Integration - AMD introduced the Ryzen AI Halo, its first AI developer platform, along with ROCm 7.2 software support for the Ryzen AI 400 Series, facilitating seamless AI adoption and development [4][11] - The Ryzen AI 400 Series processors deliver up to 60 TOPS of NPU AI compute, with features designed to meet the requirements for seamless AI experiences [7][11] Gaming Performance - The Ryzen 7 9850X3D processor is positioned as the fastest gaming CPU, offering a 400MHz higher boost clock than its predecessor, the Ryzen 7 9800X3D, and achieving up to 27% better gaming performance compared to Intel's Core Ultra 9 285K [5][19][20] - The new FSR "Redstone" technology enhances gaming experiences through machine learning frame generation and upscaling [5][30] Market Availability - Systems powered by the Ryzen AI 400 Series and Ryzen AI PRO 400 Series will be available starting Q1 2026 from major OEMs, with desktop versions expected in Q2 2026 [9][15][22] Developer Support - The AMD Ryzen AI Halo developer platform is designed to support AI development with capabilities to run large models locally, featuring up to 128GB of unified memory and optimized for AMD ROCm software [17][18] - The ROCm software platform has seen significant improvements, including a fivefold increase in AI performance and expanded compatibility across Windows and Linux [28][27] Ecosystem Growth - AMD is expanding its OEM partnerships, with a growing number of systems powered by AMD processors across consumer, commercial, and gaming segments [24][25] - Collaborations with software developers aim to integrate new AI features into everyday applications, enhancing user experiences [25][26]

Whenever Cathie Wood trims her long-time favorite stock in Tesla (TSLA), it usually grabs headlines. However, what matters more is where that capital goes next. In ARK’s Dec. 22 trades, Wood scaled back exposure to some of her biggest winners, while leaning harder into smaller, more speculative bets such as Oklo (OKLO). She scooped up 107,321 shares, a purchase worth nearly $8.9 million, indicating a fresh conviction following the stock’s recent pullback rather than a short-term trade. Nuclear energy ...

格隆汇12月23日｜据快科技，中科曙光推出开放兼容的scaleX万卡超级集群，国产万卡AI集群首次真机 亮相。中科曙光scaleX万卡超集群由多个scaleX640超节点(单机柜640卡)、scaleFabirc高速网络互连而 成，总计拥有10240块加速卡，总算力超过5EFlops(500亿亿次每秒)。 自研scaleFabric网络芯片提供 400Gb/s带宽、低于1微秒延迟，支持扩展至10万卡，网络成本降30%。采用全浸没式相变液冷，单机柜 算力密度提升20倍，PUE 1.04；三级协同优化使大模型训推效率提升30-40%，GPU利用率最高增55%， 集群可用性达99.99%，30天不可用时间小于4分钟。 ...

TeraWulf (NASDAQ: WULF) and Fluidstack plan to offer $1.275 billion in senior secured notes to fund the development of high-performance computing infrastructure. The duo are making the offering under Flash Compute, a subsidiary of a joint venture that TeraWulf (NASDAQ: WULF) and Fluidstack created earlier this year. TeraWulf holds a 50.1% interest through Big Country Wulf, and Fluidstack owns the remaining 49.9% through its subsidiary Fluidstack CS I. The notes are due in 2030 and include a 50% excess c ...

Core Points - CoreWeave's stock has dropped 4.11% to $84.67 from $88.30 due to pre-market trading pressure [1][3] - The decline is attributed to CoreWeave's announcement of a $2 billion convertible senior notes offering, with a potential expansion of $300 million [3][4] - The notes are expected to have a conversion premium of 25% to 30% and will price after Monday's market close [3][4] Company Financials - The financing aims to provide substantial capital for expanding data-center infrastructure but raises concerns about future equity dilution [4] - Increased leverage and financial risk are indicated by this move, prompting a swift reaction from investors [4] Industry Context - CoreWeave's reliance on Nvidia GPUs positions it as a key player in the AI-computing ecosystem, making its financial health a proxy for Nvidia-linked infrastructure demand [5] - Any signs of strain at CoreWeave could raise questions about sustainability, capital needs, and long-term profitability across the AI-compute supply chain [5][6] - The market's perception of this financing as a necessary step for long-term expansion or a warning about growth pace remains uncertain [6]

Summary of Conference Call Notes on China Data Centres Equities Industry Overview - The focus is on the China data centre industry, specifically companies GDS Holdings (GDS) and VNET Group (VNET) - The industry is expected to experience a slowdown in 2026, with potential recovery in 2027 driven by AI demand and clarity on chip supply issues [2][10] Key Points and Arguments 1. **2026 Slowdown and 2027 Recovery**: - A slowdown in new wholesale orders was noted in 3Q25, attributed to large clients delaying capital expenditures due to uncertainties regarding chip resolutions [2] - GDS's adjusted EBITDA growth is forecasted to slow from 10% in 2025 to 6% in 2026, while VNET's growth is expected to decelerate from 21% to 19% in the same period [2] 2. **Market Resilience**: - Despite disappointing order numbers in 3Q, GDS and VNET's share prices remained resilient, with GDS up 15% and VNET up 8% post-results announcement, indicating that the market has already priced in the anticipated slowdown [3] 3. **REITs as Valuation Benchmarks**: - Both GDS and VNET completed their C-REIT and Private REIT issuances, which are seen as providing valuation benchmarks and future financing channels [4] 4. **Preference for VNET**: - VNET is expected to outperform GDS in growth due to better wholesale capacity utilization and lower electricity costs in Inner Mongolia, which is advantageous for securing large AI orders [5] - VNET is trading at a lower valuation of 10x 2026e EV/adj. EBITDA compared to GDS at 13x [5] 5. **Target Price Adjustments**: - Target prices for both companies have been raised, with GDS's target price increasing from USD 44.10 to USD 46.90 and VNET's from USD 11.40 to USD 14.40 [6] Financial Highlights - **GDS Financials**: - Revenue projections for GDS are CNY 10,322 million for 2024, increasing to CNY 14,053 million by 2027 [11] - Adjusted EBITDA is expected to decline from CNY 6,889 million in 2025 to CNY 5,275 million in 2026 [11] - **VNET Financials**: - VNET's revenue is projected to grow from CNY 8,259 million in 2024 to CNY 14,424 million by 2027 [19] - EBITDA is expected to increase from CNY 2,268 million in 2024 to CNY 4,913 million in 2027 [19] Additional Important Insights - **Risks**: - Potential risks include failure to secure new large orders, chip shortages affecting data centre utilization, and a slowdown in AI data centre investments [34] - **Valuation Methodology**: - GDS is valued using a sum of the parts (SOTP) approach, with a target EV/EBITDA multiple of 13x for its mainland China business and 21x for its international business, DayOne [27][31] - **Market Context**: - The report highlights the competitive landscape, comparing GDS and VNET with peers like Equinix and Digital Realty, indicating a need for strategic positioning in the evolving market [35] This summary encapsulates the critical insights from the conference call regarding the China data centre industry, focusing on GDS and VNET, their financial outlook, market dynamics, and potential risks.

Core Viewpoint - BofA Securities raised its price target on Broadcom to $460 from $400 while maintaining a Buy rating, indicating strong confidence in the company's growth prospects driven by rising TPU adoption [1]. Group 1: Price Target and Rating - BofA Securities increased Broadcom's price target to $460 from $400 and maintained a Buy rating [1]. - The valuation basis was lifted to 33x CY27E, remaining within the historical range of 11x–40x [3]. Group 2: TPU Market Analysis - Rising TPU adoption is seen as a significant positive for Broadcom, which is a core design partner [2]. - BofA estimated that TPU average selling prices (ASPs) could increase from $5,000–$6,000 to $12,000–$15,000 by calendar 2026 [2]. - Unit sales of TPUs are projected to grow from approximately 2 million in CY2025 to over 3 million in CY2026, with potential upside reaching 3.6–3.8 million units depending on demand [2]. Group 3: Financial Estimates - Slight reductions were noted in Broadcom's margin assumptions due to an increased compute/ASIC mix, but earnings per share (EPS) estimates were mostly unchanged [3].

Core Insights - Insta360 has launched the Antigravity A1 panoramic drone in collaboration with a third-party team, featuring the new generation Sunflower 5 intelligent computing chip from Diguo Robotics [1] - The Sunflower 5 chip offers 10 TOPS AI computing power, optimizing algorithms such as BEV and 3D Occupancy, and is compatible with Transformer models, enhancing the drone's obstacle avoidance accuracy and flight stability [1] - In early 2024, Horizon announced the spin-off of its AIoT and robotics division to establish a wholly-owned subsidiary, Diguo Robotics, which will focus on non-automotive business areas including chip sales, product technical services, and cloud services [1]

Summary of AI Computing Power Conference Call Industry Overview - The conference call focuses on the AI computing power industry, highlighting its growth potential compared to traditional telecommunications sectors like 4G and 5G [1][2][3]. Key Points and Arguments 1. Exponential Growth and Scalability - AI computing power is driven by a data flywheel effect, with token usage increasing exponentially. For instance, the Open Router platform saw a 28-fold increase in token calls within a year, contrasting with a mere 60% growth in mobile internet traffic over a decade [1][3]. 2. Shorter Investment Return Period - AI computing power offers a shorter investment return period compared to 4G/5G, which typically requires 8-10 years to recoup costs due to upfront capital investments. In contrast, AI operates on a usage-based billing model, allowing for quicker cash recovery [1][3][9]. 3. Faster Hardware Iteration - The iteration cycle for AI hardware and software is 12-18 months, faster than the 18-24 months for traditional telecom equipment. This rapid iteration reduces unit computing costs and fosters new demand, leading to higher generational value re-pricing [1][5][11]. 4. Market Concentration and Profitability - The AI hardware industry is characterized by a concentrated supply chain, with a few upstream companies holding significant market power and profit margins. Leading firms leverage economies of scale and high-end products to enhance profitability, unlike telecom equipment, which faces buyer power and regulatory pressures [1][5][13]. 5. Incremental Value Creation - AI computing power creates new incremental value through innovative technologies and applications. For example, OpenAI's new POS feature shifts AI from passive applications to actively empowering users, a capability not achievable with traditional technologies [1][6]. 6. Untapped Application Potential - Many potential applications in AI remain underdeveloped, such as various intelligent services and automated processes. As technology advances and applications become more widespread, new scenarios will emerge, further driving market demand [1][6]. 7. Flywheel Effect - The interconnection between models, data, and applications creates a self-reinforcing flywheel mechanism. Continuous upgrades, such as Google's Gemini 2.5 and GPT iterations, enhance user engagement and open new scenarios, accelerating ecosystem development [1][7]. 8. Comparison with 4G/5G Investment Recovery - The lengthy investment recovery period for 4G/5G is attributed to substantial initial capital requirements for infrastructure, such as base station construction and spectrum auctions. For example, Germany's 2019 5G spectrum auction totaled $6.55 billion [8]. 9. AI Technology's Quick Return on Investment - AI technology's return on investment is quicker due to lower initial costs and the ability to monetize through cloud services. For instance, NVIDIA's H100 GPU costs around $30,000, with a payback period of about 400 days [9][10]. 10. Market Performance and Demand Growth - The rapid iteration of AI technology does not diminish demand; rather, it fuels it. For example, Google's Genie 3 model requires 5.2 million tokens for generating a one-minute 360-degree video, indicating a sustained need for high bandwidth and computing power [12]. 11. Stability of AI Hardware Supply Chain - The AI hardware supply chain is more stable and favorable compared to traditional telecom chains. The GPU market is dominated by NVIDIA, while other solutions like ASICs are emerging, contributing to a more stable pricing and competitive environment [13]. 12. Positive Trends in AI Computing Demand - In the first half of 2025, overseas demand for AI computing power is expected to rise, with leading companies in optical modules and PCBs showing increasing profit margins despite normal price declines [14]. 13. Future Development Potential - The AI computing market's growth potential is significantly higher than other tech sectors. Its ability to create societal value suggests that the ceiling for growth is not yet visible, making it one of the most promising areas for investment despite current high valuations [15].