公众号记得加星标⭐️，第一时间看推送不会错过。 来源 ： 内容 编译自 timesofindia 。 埃隆·马斯克透露了特斯拉即将推出的 AI5 和 AI6 芯片的新细节，声称它们可以为汽车乃至更广泛的 AI 计算性能树立新的标杆。 " 今 天 刚 刚 和 特 斯 拉 AI5 芯 片 设 计 团 队 进 行 了 一 次 很 棒 的 设 计 评 审 ！ 这 将 会 是 一 款 史 诗 级 的 芯 片，"马斯克在 X 上说道。"而接下来的 AI6 也有望成为迄今为止最好的 AI 芯片。" 马斯克解释说，特斯拉已经整合了其硅片路线图，从两种芯片架构减少到一种。"从两种芯片架构转 变为一种，意味着我们所有的硅片人才都将专注于打造一款令人难以置信的芯片。现在回想起来，这 简直是理所当然的。"他补充道。 今年夏初，特斯拉正式放弃其 Dojo 超级计算机平台，转而将所有芯片资源集中用于 AI5 和 AI6。马 斯克表示，这些芯片将作为一种"融合架构"，为汽车和数据中心的推理以及模型训练提供支持。这位 特斯拉首席执行官表示，对于参数数量低于约 2500 亿的模型而言，AI5 将成为"同类最佳的推理芯 片"，提供无与伦比的成本 ...

Core Viewpoint - Tesla's recent restructuring of its AI organization reflects a strategic shift towards enhancing the efficiency of its Full Self-Driving (FSD) and Optimus robot technologies, consolidating resources to support a unified chip architecture for training and inference [2][17]. Group 1: Organizational Changes - The restructuring involves integrating the original Dojo team into three main areas: hardware, software, and firmware, aimed at improving collaboration and resource allocation [6][8]. - Aaron Rodgers will oversee the hardware direction, focusing on chip design and semiconductor optimization, while Ashok Elluswamy will lead the software team, emphasizing AI algorithm optimization and software integration [6][8]. - Silvio Brugada will manage firmware development, concentrating on safety protocols for the autonomous driving system [8]. Group 2: Dojo Project Termination - The Dojo project, initiated in 2019 to develop high-performance chips for autonomous driving, has been officially terminated, with Tesla shifting focus to the Cortex supercomputer cluster [9][11]. - The D1 chip, developed under the Dojo project, featured advanced specifications but faced challenges such as high manufacturing costs and complex thermal designs, leading to the decision to pivot towards a unified chip strategy [9][11]. Group 3: New Computing Initiatives - Tesla is investing in the Cortex supercomputer cluster, which will utilize over 100,000 NVIDIA H100 and H200 chips, aimed at training neural networks for FSD and Optimus [2][4]. - The upcoming Cortex 2.0 center, set to begin construction in May 2025, will enhance training efficiency with a specially designed computing architecture [4]. Group 4: Future Developments - Tesla plans to release an updated version of its FSD system by September 2025, featuring significant improvements in AI model scale and video compression technology [13]. - The Robotaxi project is also progressing, with plans to expand testing areas and seek operational permits in California by the end of 2025 [14][16].

Core Insights - The recent dissolution of Tesla's Dojo project has led to significant restructuring within the company's AI divisions, with a focus on reallocating top talent to other critical areas such as autonomous driving and humanoid robotics [2][12][25] Talent Redistribution - The majority of the original Dojo team members are being reassigned primarily to the Robotaxi and humanoid robot sectors, as well as to Tesla's autonomous driving hardware development [3][8] - Software developers from the Dojo team are now reporting to Ashok Elluswamy, who oversees AI research for both Robotaxi and humanoid robots [5] - Engineers specializing in silicon chips and semiconductors have been moved to the autonomous driving hardware division to work on the upcoming AI5 chip, reporting to Aaron Rodgers [8][12] Strategic Shift - The termination of the Dojo project indicates a strategic pivot for Tesla, moving away from a fully self-developed approach to a more focused innovation on core autonomous driving technologies [8][25] - The decision to dissolve Dojo aligns with Musk's vision of vertical integration in AI hardware, marking the beginning of a significant reorganization within Tesla's AI framework [12][25] Historical Context and Future Outlook - The Dojo project, initially launched with ambitious goals, has faced stagnation since the introduction of the D1 chip in 2021, leading to its eventual discontinuation [20][23] - Tesla had invested approximately $500 million in the Dojo project, with Musk acknowledging that maintaining competitiveness in AI would require annual investments of at least several billion dollars [23] - The new AI6 chip is expected to outperform the previous Dojo framework, suggesting that while Dojo may be officially closed, its concepts could be integrated into future developments [25][26]

特斯拉上个月还在倒计时Dojo二代芯片量产，现在却突然宣布解散Dojo团队—— 那个曾被寄予"为特斯拉市值狂增5000亿"厚望的Dojo超级计算机团队，终究没能熬过这个夏天。 消息公布后，特斯拉股价在盘后小幅下跌。 Dojo解散，前负责人也另起炉灶办起了自己的DensityAI，来研发用于驱动汽车和汽车行业人工智能数据中心芯片、硬件和软件。 据知情人士透露，大约有20名前Dojo团队成员已经加入了DensityAI。 甚至有人爆料Dojo的负责人持有专利权，且拒绝特斯拉继续使用Dojo。 超级计算团队解散、负责人出走创办芯片公司、专利也不让用了…… 虽然看上去挺惨，但Dojo的解散也不排除是马斯克在进行资源优化调整，毕竟xAI在亚特兰大的数据中心建得还挺成功的。 Dojo本是特斯拉驶向未来的算力引擎 特斯拉Dojo超级计算项目在自动驾驶发展进程中曾是极具野心的布局。 2019年被马斯克首次提及，他说Dojo是一台用于训练人工智能的超级计算机。 2021年，Dojo在首届AI Day上正式亮相并发布了D1芯片，这颗自研的7mm芯片，成为Dojo的核心动力源。 2022年到2023年，马斯克坦言Dojo是"不太可 ...

闻乐 发自 凹非寺 量子位 | 公众号 QbitAI 特斯拉上个月还在倒计时Dojo二代芯片量产，现在却突然宣布 解散Dojo团队 —— 那个曾被寄予"为特斯拉市值狂增5000亿"厚望的Dojo超级计算机团队，终究没能熬过这个夏天。 消息公布后，特斯拉股价在盘后小幅下跌。 Dojo解散，前负责人也另起炉灶办起了自己的 DensityAI ，来研发用于驱动汽车和汽车行业人工智能数据中心芯片、硬件和软件。 据知情人士透露，大约有20名前Dojo团队成员已经加入了DensityAI。 甚至有人爆料Dojo的负责人持有专利权，且拒绝特斯拉继续使用Dojo。 超级计算团队解散、负责人出走创办芯片公司、专利也不让用了…… 虽然看上去挺惨，但Dojo的解散也不排除是马斯克在进行资源优化调整，毕竟xAI在亚特兰大的数据中心建得还挺成功的。 而且少一个团队还少一些财务上的压力，马斯克可能又在打算盘了。 话虽如此，我们还是重点来了解一下Dojo的情况。 Dojo本是特斯拉驶向未来的算力引擎 特斯拉Dojo超级计算项目在自动驾驶发展进程中曾是极具野心的布局。 2019年被马斯克首次提及，他说Dojo是一台用于训练人工智能的超级计算 ...

知情人士透露，特斯拉将解散其 Dojo 超级计算机团队，其负责人也将离职，这将颠覆该汽车制造商开发无人驾驶技 术内部芯片的努力。 据不愿透露姓名、因讨论内部事务而不愿透露姓名的消息人士透露，Dojo 项目负责人彼得·班农 (Peter Bannon) 即将离职，首席 执行官埃隆·马斯克 (Elon Musk) 已下令停止该项目。消息人士称，该团队最近有约 20 名员工跳槽至新成立的 DensityAI ，剩余 的 Dojo 员工将被调往特斯拉内部的其他数据中心和计算项目。 消息人士称，特斯拉计划增加对外部技术合作伙伴的依赖，包括用于计算的 Nvidia 和 Advanced Micro Devices，以及用于芯片制 造的三星电子。 据彭博社本周报道，即将正式亮相的DensityAI正在研发芯片、硬件和软件，用于驱动人工智能数据中心，这些数据中心可用于 机器人、人工智能代理、汽车应用等多个领域。该公司由Dojo前负责人Ganesh Venkataramanan以及特斯拉前员工 Bill Chang 和 Ben Floering创立。 彭博社报道后，特斯拉股价延续盘后跌势，截至纽约时间下午 5 点 29 分，跌 ...

特斯拉的 Dojo 是目前全球最大的两款处理器之一。这些巨大的晶圆级芯片使用整块 300 毫米晶圆， 这意味着一次性构建更大的计算能力根本不可能。每个 Dojo 晶圆级处理器最多可容纳 8,850 个核 心，但其中一些核心在部署后可能会引发静默数据损坏 (SDC)，从而破坏大规模训练运行的结果。 大型处理器 鉴于 Dojo Training Tile（大型晶圆尺寸芯片）的极端复杂性，即使在制造过程中也不容易检测到有 缺陷的芯片，但是当涉及到静默数据损坏（SDC）时，事情会变得更加复杂。 请记住，所有类型的硬件都不可避免地会出现 SDC，但 Dojo 处理器的电流消耗高达 18,000 安培， 功耗高达 15,000 瓦，这会产生影响。不过，所有核心都应该按预期运行，否则特斯拉的 AI 训练将 变得更加复杂，因为数据损坏导致的一个错误就可能使数周的 AI 训练付诸东流。 公众号记得加星标⭐️，第一时间看推送不会错过。 来源：内容 编译自 tesla 。 在大型处理器上检测故障核心并将其禁用是一项挑战，但特斯拉开发了 Stress 工具，该工具不仅可 以在 Dojo 处理器上检测容易出现静默数据损坏的核心，还可以 ...

Core Viewpoint - Tesla is undergoing a significant transformation from an automotive manufacturer to an artificial intelligence company, facing challenges in both its automotive sales and the development of its AI-related business [8][21][24]. Group 1: Financial Performance - Tesla reported its worst quarterly results, with net profit plummeting by 71% and automotive revenue dropping by 20% year-on-year, leading to a state of operational loss [4][26]. - In 2023, Tesla's total vehicle deliveries saw a rare decline of 1%, with revenue growth stagnating at just 1% year-on-year, highlighting the transitional difficulties from automotive to AI [26][49]. Group 2: Market Position and Strategy - Tesla's market valuation has become increasingly decoupled from its actual business progress, as the allure of its AI roadmap has not been matched by tangible advancements in its operations [21][25]. - The company aims to shift its competitive edge from manufacturing to software capabilities, focusing on autonomous driving and robotaxi services, which are expected to drive future profitability [14][16]. Group 3: Product Development Challenges - The highly anticipated Cybertruck has faced multiple production delays, with a significant drop in sales to just 6,406 units in the first quarter of 2023 due to insufficient demand [36][37]. - The Model 2/Q, a budget vehicle aimed at a price point of $25,000, has seen its development timeline extended indefinitely, with no concrete updates on its release [41][44]. Group 4: Competitive Landscape - Tesla's reliance on the Chinese market is increasing, with China being the only market showing positive growth for Tesla in 2024, although its market share has significantly decreased from 15.8% in Q1 2021 to 5.6% in Q1 2023 [49]. - The competitive landscape in the electric vehicle market is intensifying, with numerous alternatives emerging in the same price range as the anticipated Model 2/Q, complicating Tesla's market position [47].