2026 线上黑客松：AMD E2E Model Speedrun 正式吹响集结号！ 你，敢来迎战吗？ 面向所有热爱硬核技术的你：这一次，用代码说话，用性能封神。 责编 | 梦依丹 出品丨AI 科技大本营（ID：rgznai100） 在 DeepSeek-R1 和 Kimi K2.5 等顶级开源模型确立了万亿参数的工业基准后，真正的极限性能压榨才刚刚开始。谁能在极高并发下打破内存墙？谁 能用最优雅的代码实现最高效的算子重构？在当下的 AI 战场，速度即是正义，吞吐量决定生死。 由 AMD 与 GPU MODE 联合发起的极客巅峰对决，正式向全球发榜： 在这里，你将直接操控 AMD 专为大模型打造的顶配云端 GPU 阵列 在这里，没有纸上谈兵的 PPT，只有硬核的绝对速度与吞吐量。 在这里，你的极限优化不仅能赢取巨额美金，更将直接合入主流开源框架，定义下一代 AI 推理的工业标准！ 百万奖金, 名利双收 进入决赛的 Top 10 队伍,保底获得 1 万美元奖金! 【赛道一】 DeepSeek-R1-0528 (FP4 + MTP) | 桂冠: 35 万美元奖金 【赛道二】 Kimi K2.5 1T (FP4) ...

Core Insights - Crane Harbor Acquisition Corp (CHAC) stock is currently trading at depressed levels, raising questions about its future performance [1] Group 1: Xanadu's Technological Advancements - Xanadu utilized AMD's high-performance computing and AI infrastructure on the AMD DevCloud alongside its PennyLane quantum software to conduct computational fluid dynamics simulations in a hybrid quantum-classical environment [2] - The company achieved a significant performance improvement, reducing simulation time by 25 times when transitioning a core quantum algorithm from a traditional CPU to a single AMD GPU [3] - Xanadu's Catalyst compiler successfully translated a 68-qubit quantum circuit into over 15 million hardware-optimized gates, which is a crucial step towards developing future fault-tolerant quantum systems for industrial applications [3] Group 2: Merger and Financial Projections - Xanadu is anticipated to go public following a shareholder vote on March 19 regarding its merger with Crane Harbor, with plans to close the merger in the first quarter of 2026 after receiving SEC clearance and pending final approvals [4] - Crane Harbor Acquisition shares were trading at $10.29 at the time of publication, indicating a flat trading day for the stock [5]

Report's Investment Rating - No information provided on the industry investment rating Core Viewpoints - The global economy has passed its peak in late 2025 and started to decline due to the continuous wrong policies of the United States [3] - The return of the US to the Monroe Doctrine will have a profound and subversive impact on major asset classes such as the global economy, US bonds, US stocks, the US dollar, precious metals, and industrial metals [3] - Wash's combination of interest - rate cuts and balance - sheet reduction indicates a major shift in the Fed's monetary policy, creating a strong expectation of liquidity contraction for equity assets [3] Summary by Related Information Categories Software Industry - JPMorgan Chase CEO Dimon warns that the current high asset prices and blind profit - seeking are reminiscent of the 2008 financial crisis, and the credit cycle reversal may trigger an unexpected wave of defaults, with the AI - disrupted software industry as a potential hard - hit area [1] - Arini's founder warns that even without a substantial impact from AI, market panic can raise financing costs and trigger large - scale defaults in the software industry, as software companies generally operate with high leverage and rely on private credit for financing [1] - Traders are short - selling ETFs holding a large number of software company loans through record - high put options, with a focus on Invesco's BKLN [1] - OpenAI aims to replace giants like Salesforce and Workday, while Anthropic emphasizes collaboration with existing software, leading to a general rebound in enterprise software stocks [1] Global Economy and Finance - Hedge funds have been net - selling US stocks at the fastest pace since March last year [2] - Bridgewater's Dalio warns that the world is on the verge of a "capital war" due to geopolitical tensions and capital market volatility [2] - The expected balance - sheet reduction policy of the Fed's nominee Wash has a strong negative impact on global equity and commodity assets [2] - Nomura predicts that the Fed's uncertainties will peak from July to November 2026, and there may be a trend of "fleeing US assets" [2] - High - end consumer spending remains resilient, but middle - and low - income families are cutting back, and funds are flowing from tech stocks to defensive sectors [2] Technology and Business Cooperation - AMD and Meta have a five - year strategic partnership, with Meta getting 6GW of AMD GPU and AMD offering up to 1.6 billion performance - based warrants (about 10% equity) in exchange for a $100 billion order [1] - Google will build a new data center in Minnesota and cooperate with Xcel Energy to implement innovative energy solutions [1] AI - related Technology Trends - OpenClaw's Steinberger submits 90,000 lines of code in a year with AI, stating that AI can solve problems autonomously, and he predicts a full - scale technological explosion in 2026 [1] US Policy - making and International Relations - The Trump administration plans to use the Pentagon's AI project to set reference prices for key minerals and build a global metal trading group, initially focusing on four metals: germanium, gallium, antimony, and tungsten [1] - The US has released a new National Security Strategy, giving up global hegemony and adjusting economic relations with China to revive its economic autonomy [2] - The US has taken actions such as attempting to arrest the Venezuelan president, control Venezuelan oil, and buy Greenland, bringing great uncertainty to the global economy [2]

当地时间2月24日美股开盘前，Meta和AMD双双在官网宣布了一项为期五年的战略合作伙伴关系，Meta将部署6GW（每吉瓦容量交易金额约达数百亿美 元）的AMD GPU，作为交换，AMD将提供最多1.6亿股（约10%股权）的基于业绩的认股权证。分析人士认为，这份战略合作背后，既表明Meta在AI领域 持续扩张的强烈诉求，也成为AMD与行业龙头英伟达展开正面抗衡的重要一步。 10%股权换1000亿订单 根据协议，Meta将从2026年下半年起，在五年内采购AMD专为AI模型设计的芯片及服务器设备。AMD首席执行官苏姿丰透露，每吉瓦容量的交易金额将 达"数百亿美元"级别。这意味着，协议的总金额将超过600亿美元，有可能达到上千亿美元。 作为协议的一部分，AMD正在设计一款专门为Meta工作负载量身定制的半定制MI450 GPU。AMD预计初始出货将于2026年下半年开始，产品将采用基于 MI450架构的定制Instinct GPU以及代号为"Venice"的第六代EPYC CPU，运行ROCm软件，并基于Helios机架级平台构建。 苏姿丰表示，Meta早已深度参与AMD芯片设计，未来还将获得定制版的MI450加 ...

针对AMD与Meta"10%股权换1000亿订单"，高盛表示，在看到Meta和OpenAI的实际部署进展落地之 前，尚不愿给出更积极的评级。 2月24日，华尔街见闻文章指出，AMD与Meta宣布了一项为期五年的战略合作伙伴关系，Meta将部署 6GW的AMD GPU。作为交换，AMD将提供最多1.6亿股（约10%股权）的基于业绩的认股权证。 2月25日，据追风交易台消息，高盛在最新研报中指出，这一合作显著提升了AMD在加速器市场的地 位，也意味着AMD在与英伟达等巨头的竞争中取得了重要进展，但该行强调，只有在看到Meta和 OpenAI的实际部署时间表后，才会对AMD采取更积极的立场，维持对AMD的"中性"评级，目标价上 调至240美元。 战略合作细节：股权换订单 高盛称，这笔交易的核心逻辑是：Meta用潜在股权换取了AMD的深度定制服务与长期供货承诺，AMD 则用未来股权稀释换取了顶级超大规模云厂商的确定性订单和市场份额背书。 高盛将AMD的12个月目标价从210美元上调至240美元，这是基于28倍的市盈率（低于之前的30倍，原 因是同行市盈率下降）应用于8.50美元的标准化EPS预测（高于之前的7.00美 ...

热点聚焦 按市值排名，明日即将发布财报的英伟达涨0.68%、苹果涨2.24%、谷歌-A跌0.19%、微软涨1.18%、亚马逊涨1.6%、台积电涨4.25%、Meta涨0.32%、博通跌 1.47%、特斯拉涨2.39%、伯克希尔哈撒韦-A涨0.14%、沃尔玛涨0.75%。 中概股多数走强，纳斯达克中国金龙指数收涨1.37%。 截至收盘，阿里巴巴涨0.22%、拼多多涨1.06%、网易涨0.73%、京东涨1.07%、百度跌0.23%、携程涨1.59%、理想汽车涨1.3%、富途控股跌1.89%、哔哩哔 哩跌0.5%、蔚来涨0.19%、世纪互联涨6.68%、万国数据涨6.72%。 1.小米集团董事长兼CEO雷军表示，小米计划未来五年重点攻坚芯片、AI、操作系统等底层核心技术，向着成为全球硬核科技公司的目标不断努力。 2.据报道，Meta与AMD达成AI设备采购协议，将部署6吉瓦AMD GPU。AMD同意向Meta公司出售价值高达600亿美元的AI芯片。此外，AMD将向Meta Platforms发行基于业绩的认股权证，可认购最多1.6亿股AMD普通股。 3.做空机构香橼资本周二发帖称，已经对存储牛股闪迪建立空头头寸。受 ...

据市场消息，Meta与AMD达成AI设备采购协议，将部署6吉瓦AMD GPU。AMD同意向Meta公司出售价 值高达600亿美元的 AI芯片。（格隆汇） ...

Group 1 - The core message of the article highlights that Dr. Lisa Su will be a keynote speaker at CES, discussing AMD's innovations in chip stacking technology and its impact on AI advancements globally [2][3] - AMD emphasizes that Dr. Su will explore various solutions related to AMD's CPUs, GPUs, adaptive computing technologies, and AI software to address critical global AI challenges for customers and partners [2][3]

Report Industry Investment Rating No relevant content provided. Core Viewpoints The report delves into the in - depth analysis of Google clusters, including their Scale - up (3D structure and optical interconnection) and Scale - out aspects, and also compares the architectures of different GPUs such as NVIDIA and AMD [1][2]. Summary by Directory 1. Google Cluster's Scale up: 3D Structure - **TPU Architecture**: The Ironwood architecture of TPU has high - performance computing components like TensorCore, XLU, VPU, etc., and is connected by high - speed ICI. It uses HBM3 and HBM3E memory to achieve scale - up of 9216 chips [11][12]. - **From TPU to TPU Rack**: A TPU Tray contains 4 Ironwood TPUs, and a TPU Rack consists of 16 TPU Trays and 64 TPU chips. The rack has a specific physical structure and cooling system [28][29]. - **Comparison with Other GPUs**: Compares the architectures of NVIDIA (from Hopper to Blackwell) and AMD (from MI350 to MI400) GPUs, highlighting their different interconnect technologies and performance parameters [20][25]. 2. Google Cluster's Scale up Optical Interconnection: Optical Path Switch - **Optical Switch Components**: The optical path switch uses components such as 850nm camera modules, dichroic beam splitters, fiber collimators, and 2D MEMS micromirrors to separate or combine calibration light and signal light [46]. - **TPU SuperPod Structure**: A TPU SuperPod consists of 64 Google racks, divided into 8 groups of 8 racks. It integrates 4096 chips, sharing 256TiB of HBM memory, with a total computing performance of over 1 ExaFLOP. Each group of 8 racks has a CDU for liquid - cooling [60]. 3. TPU Cluster, Proportion of Optical Path Switches and Optical Modules - **TPU V4**: The proportion of optical path switches is 1.1% with 4096 TPUs, and the proportion of optical modules is 1.5 [70][84]. - **TPU V7**: The proportion of optical path switches is 0.52% with 9216 TPUs, and the proportion of optical modules is also 1.5 [75][89]. - **Rack - level Data**: For a single rack, there are 6 * 16 external optical modules, 4 * 16 PCB traces, and 80 copper cables [94]. 4. Google Cluster's Scale out - **Switch Parameters**: The Tomahawk 5 switch has 128 400G ports [103]. - **Communication Outside TPU SuperPod**: Communication outside the TPU SuperPod is carried out through the Data - center Network (DCN), which includes optical circuit switches and physical fibers [106][108]. - **NV Scale - out OCS**: In the NV scale - out, OCS is used in a redundant spine - leaf network structure, which can enhance the resilience of the network [113][114]. - **Comparison of Interconnection Schemes in a 100,000 - card Cluster**: Compares the InfiniBand, NVIDIA Spectrum - X, and Broadcom Tomahawk5 interconnection schemes in terms of switch quantity, optical module quantity, cost, etc. [125].

Core Viewpoint - Amazon has announced a long-term strategic partnership with OpenAI, involving a financial commitment of $38 billion, which is expected to enhance AI processing capabilities through AWS infrastructure [3][4]. Group 1: Partnership Details - OpenAI will utilize Amazon EC2 UltraServers, accessing hundreds of thousands of NVIDIA GPUs, with the potential to scale to tens of millions of CPUs [4]. - The partnership's value of $38 billion is projected to grow over the next seven years [4][5]. - OpenAI is expected to start using AWS computing services immediately, with full deployment of computing capabilities anticipated by the end of 2026 [5]. Group 2: Competitive Landscape - OpenAI is focusing on GPU usage for its computational needs, contrasting with Anthropic, which has opted for Amazon's proprietary AI chips [5]. - Recent collaborations between OpenAI and major GPU manufacturers, including NVIDIA and AMD, indicate a trend of significant investments in AI infrastructure [6]. Group 3: Financial Performance - Amazon reported a 12% increase in net sales to $180.2 billion for Q3 2025, with a net profit of $21.2 billion, reflecting a 38.6% year-over-year growth [7]. - AWS has experienced its highest growth rate since 2022, driven by strong demand for AI and core infrastructure [7]. Group 4: Market Sentiment - There is ongoing debate in the market regarding the potential for an AI bubble, with experts suggesting that the return on investment from massive AI expenditures may not be clear for at least a year [7].