📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $纽柯(NUE.US) Nucor Guides Q3 EPS to $5.55–$5.65, Above Q2 Earnings
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗡𝘂𝗰𝗼𝗿 expects Q3 diluted EPS between $𝟱.𝟱𝟱 and $𝟱.𝟲𝟱.➤ Q2 diluted EPS was $𝟱.𝟬𝟰, with adjusted EPS of $𝟰.𝟴𝟰.➤ Year-ago Q3 earnings were $𝟮.𝟲𝟯 per diluted share.➤ 𝗦𝘁𝗲𝗲𝗹 𝗺𝗶𝗹𝗹𝘀 earnings expected to rise on higher selling prices.➤ 𝗦𝘁𝗲𝗲𝗹 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝘀 earnings expected higher on volumes and realized pricing.➤ 𝗥𝗮𝘄 𝗺𝗮𝘁𝗲𝗿𝗶𝗮𝗹𝘀 earnings expected to decline on pricing and shipments.➤ Q2 included $𝟭𝟯𝟬 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 in prior-period procurement cash refunds.➤ Nucor repurchased approximately 𝟮.𝟬𝟯 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 shares during Q3.➤ Average Q3 repurchase price was $𝟮𝟰𝟳.𝟬𝟰 per share.➤ Year-to-date shareholder returns total approximately $𝟭.𝟯𝟲 𝗯𝗶𝗹𝗹𝗶𝗼𝗻.👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ Guidance indicates a 𝘀𝗲𝗾𝘂𝗲𝗻𝘁𝗶𝗮𝗹 𝗲𝗮𝗿𝗻𝗶𝗻𝗴𝘀 𝗶𝗺𝗽𝗿𝗼𝘃𝗲𝗺𝗲𝗻𝘁 despite losing Q2 one-time benefits.➤ Higher steel pricing and volumes signal stronger operating conditions in core segments.➤ Continued 𝘀𝗵𝗮𝗿𝗲 𝗿𝗲𝗽𝘂𝗿𝗰𝗵𝗮𝘀 demonstrate substantial ongoing capital returns to shareholders.➤ Raw-material weakness remains a potential offset to improving steel profitability.
Hardik Shah
H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: OpenAI 发布 GPT-6 Astra Law，一款面向法律工作的全新 AI 模型与平台
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗢𝗽𝗲𝗻𝗔𝗜 推出 𝗔𝘀𝘁𝗿𝗮 𝗳𝗼𝗿 𝗟𝗮𝘄，专为律师事务所和法律科技公司打造。➤ 该平台将 𝗚𝗣𝗧-𝟲 𝗔𝘀𝘁𝗿𝗮 与专业的法律工具和上下文相结合。➤ 法律搜索覆盖超过 𝟮𝟯𝟬 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 个 URL，并每日更新来源。➤ Astra for Law 的正确率达到 𝟱𝟰.𝟬%，而使用网络搜索的 Astra 仅为 𝟯𝟴.𝟳%。➤ 这代表在 OpenAI 的法律研究评估中实现了 𝟰𝟬% 的相对提升。➤ 在最高推理力度下，案例法搜索找到的参考案例增加了 𝟮𝟰%。➤ Astra 从正确意见中检索到了多至 𝟱𝟰% 的相关段落。➤ 𝗛𝗮𝗿𝘃𝗲𝘆 和 𝗟𝗲𝗴𝗼𝗿𝗮 将能够基于 Astra for Law 进行开发。➤ OpenAI 推出了 𝟮𝟲 款由合作伙伴开发的插件，将 ChatGPT 与专业法律工具连接起来。➤ 精选的律所将通过 𝗧𝗿𝘂𝘀𝘁𝗲𝗱 𝗔𝗰𝗰𝗲𝘀𝘀 𝗣𝗿𝗼𝗴𝗿𝗮𝗺 获得早期访问权限。👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ 推动前沿 AI 更深入地融入高价值的 𝗽𝗿𝗼𝗳𝗲𝘀𝘀𝗶𝗼𝗻𝗮𝗹 𝗹𝗲𝗴𝗮𝗹 工作流。➤ 专业的法律搜索可能提升权威检索能力和研究基础。➤ 插件允许律所将 AI 与现有的 𝗹𝗲𝗴𝗮𝗹 𝗱𝗮𝘁𝗮 和工作流结合。➤ 隐私和治理控制措施解决了机密、高风险客户工作的需求。➤ 大型律所的采用可能会加速 𝗹𝗲𝗴𝗮𝗹-𝗔𝗜 市场的竞争。👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁𝘀:𝗝𝗼𝗵𝗻 𝗦𝗮𝘃𝘃𝗮, Sullivan & Cromwell 合伙人：“我们很感激能预览 Astra for Law 的早期版本，这些版本是围绕法律用例构建的。在诉讼和交易事务方面，模型展示了令人印象深刻的研究深度和对权威性的敏感度。即使在早期阶段，它们也感觉是朝着专注于法律的 AI 迈出的重要一步，这种 AI 扎根于既全面又及时的研究。”𝗠𝗶𝗰𝗵𝗮𝗲𝗹 𝗥𝘂𝗯𝗶𝗻, Latham AI 战略委员会主席：“随着 AI 能力的增强，将其部署在需要严格治理、监督和问责的环境中的能力也随之提升。此次合作建立在 Latham 在 AI 开发、治理和基础设施方面的更广泛投资之上，这些投资支撑了我们负责任 AI 的企业级战略。”𝗗𝗮𝘃𝗲 𝗣𝗲𝗶𝗻𝘀𝗶𝗽𝗽, Cooley 全球资本市场组联合主席兼合伙人：“我们与 OpenAI 的合作让我们重新思考这项工作的完成方式——让律师和管理团队更快地度过密集的准备阶段，进入需要判断力、市场经验和战略思维的问题。”𝗝𝗼𝗲𝗹 𝗛𝗿𝗼𝗻, Thomson Reuters 首席技术官：“随着 AI 变得更加开放和互操作，其价值不仅仅在于连接性。法律专业人士需要的不仅仅是信息访问。他们需要可信的情报、相关的企业和案件上下文、专门构建的法律能力，以及处理高风险工作所需的治理。这就是 Thomson Reuters 通过 HighQ 和 CoCounsel Legal 所提供的。CoCounsel 仍然是受信任的专业 AI 系统，旨在帮助完成这项工作。我们与 OpenAI 的合作有助于将这些能力提供到客户选择的环境中，同时保持他们依赖的准确性、保密性和问责制。”𝗪𝗮𝗰𝗵𝘁𝗲𝗹𝗹, 𝗟𝗶𝗽𝘁𝗼𝗻, 𝗥𝗼𝘀𝗲𝗻 & 𝗞𝗮𝘁𝘇：“我们非常感谢 OpenAI 的前端工程团队在与 AI 在法律工作中发挥作用的讨论中所展现的技术深度和倾听意愿，并对 AI 塑造法律实践下一个时代的可能性感到兴奋。”H
📢 𝐉𝐔𝐒𝐓 𝐈𝐍: OpenAI 聘请 SpaceX 的 McCarthy 担任全球销售副总裁 - Fortune
H
⚡ 更新：美国证券交易委员会（SEC）批准美国代币化股票交易五年豁免期 - $Robinhood(HOOD.US) $Circle(CRCL.US)
H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $Coreweave(CRWV.US) CoreWeave Contracts New AI Compute Capacity at Higher Prices as Commitments Surge
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗖𝗼𝗿𝗲𝗪𝗲𝗮𝘃𝗲 continues contracting new compute capacity at 𝗵𝗶𝗴𝗵𝗲𝗿 𝗽𝗿𝗶𝗰𝗲𝘀.➤ Q3 contracts carry pricing near $𝟰𝟬 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 𝗽𝗲𝗿 𝗺𝗲𝗴𝗮𝘄𝗮𝘁𝘁 annualized.➤ New customer contracts generally span approximately 𝟯–𝟲 𝗺𝗼𝗻𝘁𝗵𝘀.➤ CoreWeave added over $𝟮𝟱 𝗯𝗶𝗹𝗹𝗶𝗼𝗻 in net new customer commitments.➤ Commitments were secured early in the 𝘁𝗵𝗶𝗿𝗱 𝗾𝘂𝗮𝗿𝘁𝗲𝗿 𝗼𝗳 𝟮𝟬𝟮𝟲.➤ New commitments aren't included in June 30 revenue backlog.➤ Total contracted power reached approximately 𝟰.𝟮 𝗴𝗶𝗴𝗮𝘄𝗮𝘁𝘁𝘀 by August 11.➤ Contracted power increased from approximately 𝟯.𝟳 𝗴𝗶𝗴𝗮𝘄𝗮𝘁𝘁𝘀 at June-end.👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ Higher pricing indicates 𝘀𝘁𝗿𝗼𝗻𝗴 𝗱𝗲𝗺𝗮𝗻𝗱 for CoreWeave's AI compute capacity.➤ Over $𝟮𝟱 𝗯𝗶𝗹𝗹𝗶𝗼𝗻 in commitments provides additional future revenue visibility.➤ Rising contracted power signals continued expansion of CoreWeave's 𝗔𝗜 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 footprint.➤ Execution remains exposed to financing, power delivery, supply-chain, and customer risks.
H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $Fluence Energy(FLNC.US) Fluence Energy Cuts FY2026 Revenue and EBITDA Guidance on Houston Manufacturing Delays
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗙𝗹𝘂𝗲𝗻𝗰𝗲 𝗘𝗻𝗲𝗿𝗴𝘆 cuts fiscal 𝟮𝟬𝟮𝟲 financial guidance.➤ FY2026 revenue forecast falls to approximately $𝟮.𝟰 𝗯𝗶𝗹𝗹𝗶𝗼𝗻.➤ Previous revenue guidance midpoint was approximately $𝟯.𝟬 𝗯𝗶𝗹𝗹𝗶𝗼𝗻.➤ Adjusted EBITDA outlook drops to approximately $𝟮𝟬𝟬 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 loss.➤ Previous adjusted EBITDA forecast called for approximately $𝟭𝟬 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 loss.➤ Houston contract manufacturing delays are the primary guidance-cut driver.➤ Corrective actions have already increased 𝗱𝗮𝗶𝗹𝘆 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 levels.➤ Fluence is restructuring its 𝗼𝗽𝗲𝗿𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗼𝗿𝗴𝗮𝗻𝗶𝘇𝗮𝘁𝗶𝗼𝗻.➤ Company targets 𝗻𝗲𝘂𝘁𝗿𝗮𝗹-𝘁𝗼-𝗽𝗼𝘀𝗶𝘁𝗶𝘃𝗲 operating cash flow in FY2027.➤ Fluence says domestic and international 𝗱𝗲𝗺𝗮𝗻𝗱 remains strong.👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ Revenue guidance was reduced by roughly $𝟲𝟬𝟬 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 versus the prior midpoint.➤ EBITDA deterioration signals substantial near-term pressure from manufacturing disruptions.➤ Houston execution is critical for converting backlog into 𝗿𝗲𝘃𝗲𝗻𝘂𝗲 and cash.➤ FY2027 cash-flow target highlights management's focus on avoiding 𝗲𝘅𝘁𝗲𝗿𝗻𝗮𝗹 𝗰𝗮𝗽𝗶𝘁𝗮𝗹.👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁𝘀:𝗝𝘂𝗹𝗶𝗮𝗻 𝗡𝗲𝗯𝗿𝗲𝗱𝗮, President and Chief Executive Officer:"We continue to experience delays in the ramp up of our contract manufacturing facility in Houston which is the primary reason we are now lowering our fiscal year 2026 financial guidance."𝗔𝗵𝗺𝗲𝗱 𝗣𝗮𝘀𝗵𝗮, Chief Financial Officer:The provided source summarizes Pasha's comments but does not provide a direct quotation, so his statement cannot be reproduced verbatim.H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $Bitdeer Tech(BTDR.US) Bitdeer Announces August 2026 Production and Operations Update
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗕𝗶𝘁𝗱𝗲𝗲𝗿's 9.5 MW A102 facility reaches 𝟭𝟬𝟬% contracted GPU capacity.➤ A102 commitments represent more than $𝟴𝟬𝟬 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 expected revenue.➤ AI Cloud pipeline reaches 𝟮𝟬𝟲.𝟱 𝗠𝗪 across multiple facilities.➤ AI Cloud ARR increased to approximately $𝟴𝟲 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 in August.➤ A202 agreement adds 𝟲𝟱.𝟭 𝗠𝗪 under a 𝟭𝟬-𝘆𝗲𝗮𝗿 contract.➤ Bhutan AI project targets 𝟯𝟬 𝗠𝗪 initially, expandable to 𝟱𝟬𝟬 𝗠𝗪.➤ Tydal's 𝟭𝟲-𝘆𝗲𝗮𝗿 lease represents approximately $𝟰.𝟳 𝗯𝗶𝗹𝗹𝗶𝗼𝗻 contract value.➤ Tydal contract value could reach approximately $𝟴.𝟬 𝗯𝗶𝗹𝗹𝗶𝗼𝗻 with extension.➤ Bitcoin production surged 𝟮𝟰𝟵% year-over-year to 𝟭,𝟯𝟭𝟬 𝗕𝗧𝗖.➤ Self-mining hash rate reached approximately 𝟳𝟵.𝟵 𝗘𝗛/𝘀 in August.👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ Long-term AI contracts diversify 𝗕𝗶𝘁𝗱𝗲𝗲𝗿 beyond Bitcoin-mining revenue.➤ Fully contracted A102 provides revenue visibility before planned 𝗡𝗼𝘃𝗲𝗺𝗯𝗲𝗿 𝟮𝟬𝟮𝟲 energization.➤ Malaysia expansion materially increases Bitdeer's exposure to high-density 𝗔𝗜 𝗰𝗼𝗺𝗽𝘂𝘁𝗲 demand.➤ Rising Bitcoin production strengthens Bitdeer's core mining operations alongside AI expansion.👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁:𝗠𝗶𝗰𝗵𝗮𝗲𝗹 𝗚. 𝗣𝗼𝘁𝘁𝗲𝗿, Chief Financial Officer of Bitdeer:"Our AI-cloud strategy continues to validate the vertically integrated model we've built at Bitdeer, converting powered land into large-scale, contracted AI/HPC revenue in addition to simply mining Bitcoin,""Tydal and the full contracting of our A102 facility are proof points that customers value the reliability and scale of our platform.""We remain focused on turning that demand into durable, high-margin revenue that complements our mining business, and we're encouraged by the pipeline of opportunities ahead."
H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $NU Holdings(NU.US) Nu Holdings Launches U.S. Banking Operations and Introduces Nu Global
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗡𝘂 𝗛𝗼𝗹𝗱𝗶𝗻𝗴𝘀 begins banking operations in the 𝗨𝗻𝗶𝘁𝗲𝗱 𝗦𝘁𝗮𝘁𝗲𝘀.➤ Launch marks Nu's expansion beyond its core 𝗟𝗮𝘁𝗶𝗻 𝗔𝗺𝗲𝗿𝗶𝗰𝗮𝗻 markets.➤ Nu launches products targeting everyday 𝗨.𝗦. 𝗯𝗮𝗻𝗸𝗶𝗻𝗴 needs.➤ Company reports more than 𝗟𝟰𝟬 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 customers across Latin America.➤ Nu also introduces 𝗡𝘂 𝗚𝗹𝗼𝗯𝗮𝗹, a multi-currency digital account.➤ Nu Global targets customers with 𝗶𝗻𝘁𝗲𝗿𝗻𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗯𝗮𝗻𝗸𝗶𝗻𝗴 needs.➤ U.S. entry extends Nu's 𝗱𝗶𝗴𝗶𝘁𝗮𝗹-𝗳𝗶𝗿𝘀𝘁 retail banking strategy globally.➤ Expansion represents a significant new-market milestone for 𝗡𝘂 𝗛𝗼𝗹𝗱𝗶𝗻𝗴𝘀.👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ U.S. launch opens a major new addressable market for 𝗡𝘂.➤ Expansion could diversify growth beyond Nu's established 𝗟𝗮𝘁𝗶𝗻 𝗔𝗺𝗲𝗿𝗶𝗰𝗮 footprint.➤ 𝗡𝘂 𝗚𝗹𝗼𝗯𝗮𝗹 broadens the company's offering into cross-border banking.➤ U.S. execution will test whether Nu's digital model scales internationally.👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁:𝗗𝗮𝘃𝗶𝗱 𝗩𝗲́𝗹𝗲𝘇, Founder and CEO of Nu Holdings:"This is the next chapter of something we started thirteen years ago in a small house in São Paulo. Since our founding, we have believed that consumer-obsessed digital banking is the future of retail banking globally, and after proving this hypothesis with over 140 million fanatical customers in Latin America, we are excited to start executing our thesis beyond our core region."
H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: CoreWeave 将多机架 NVIDIA Vera Rubin NVL72 集群引入其 AI 云 - $Coreweave(CRWV.US) $英伟达(NVDA.US)
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗖𝗼𝗿𝗲𝗪𝗲𝗮𝘃𝗲将其𝗡𝗩𝗜𝗗𝗜𝗔 𝗩𝗲𝗿𝗮 𝗥𝘂𝗯𝗶𝗻 𝗡𝗩𝟳𝟮部署至云端。➤ 集群连接了数百个𝗥𝘂𝗯𝗶𝗻 𝗚𝗣𝗨，以支持𝗮𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜工作负载。➤ 每个 NVL72 机架将𝟳𝟮个𝗥𝘂𝗯𝗶𝗻 𝗚𝗣𝗨与𝟯𝟲个𝗩𝗲𝗿𝗮 𝗖𝗣𝗨相结合。➤ 𝗡𝗩𝗜𝗗𝗜𝗔 𝗦𝗽𝗲𝗰𝘁𝗿𝘂𝗺-𝗫以太网将多个机架连接为扩展式集群。➤ 每个 Rubin GPU 可获得高达𝟭.𝟲 𝗧𝗯/𝘀的扩展连接速度。➤ 𝗖𝗼𝗿𝗲𝗪𝗲𝗮𝘃𝗲 𝗠𝗶𝘀𝘀𝗶𝗼𝗻 𝗖𝗼𝗻𝘁𝗿𝗼𝗹自动化了机架设置、验证、供电和冷却流程。➤ LOTA 提供本地 NVMe 级读取速度，单 GPU 可达𝟳 𝗚𝗕/𝘀。➤ 与传统存储集群读取相比，LOTA 将延迟降低了𝟴倍。➤ 新的𝗰𝗿𝗼𝘀𝘀-𝗿𝗲𝗴𝗶𝗼𝗻 𝘄𝗿𝗶𝘁𝗲 𝗮𝗰𝗰𝗲𝗹𝗲𝗿𝗮𝘁𝗶𝗼𝗻最大限度地减少了 AI 检查点延迟。➤ 新的𝗔𝗿𝗰𝗵𝗶𝘃𝗲层级旨在降低长期 AI 数据存储成本。👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ 扩大了 CoreWeave 进行大规模𝗮𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜训练和推理的能力。➤ 更快的存储访问可以减少 GPU 空闲时间并提高利用率。➤ 多机架 Rubin 支持更大规模的模型和𝗿𝗲𝗶𝗻𝗳𝗼𝗿𝗰𝗲𝗺𝗲𝗻𝘁 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴规模化应用。➤ 跨区域加速解决了分布式 AI 基础设施中的一个关键瓶颈。👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁𝘀:𝗖𝗵𝗲𝗻 𝗚𝗼𝗹𝗱𝗯𝗲𝗿𝗴, CoreWeave 产品与工程执行副总裁：“CoreWeave 是首家验证并上线 Vera Rubin NVL72 的 AI 云服务提供商，证明了这种先进的机架级架构可以作为可靠的高性能云服务运行。通过多机架 Vera Rubin，我们将数百个 Rubin GPU 连接为一个单一的扩展集群。对于构建 Agentic AI 的客户而言，这意味着更大的规模、更快的迭代以及更高的生产力，因为模型和智能体在不断学习和改进。”𝗖𝗲́𝗰𝗶𝗹𝗲 𝗥𝗼𝗯𝗲𝗿𝘁-𝗠𝗶𝗰𝗵𝗼𝗻, Cohere 内部基础设施总监：“我们的数据集跨越多个区域，我们无法承受因跨区域检索延迟而打乱训练计划。CoreWeave AI 对象存储为我们提供了跨区域的统一数据集足迹，并在本地缓存读取数据，因此没有任何内容会等待网络传输。这是在围绕数据进行规划与直接进行训练之间的区别。”H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: Blockfusion 与 $Coreweave(CRWV.US) CoreWeave 签署为期 15 年的 AI 数据中心锚定租约
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ Blockfusion 在尼亚加拉瀑布园区与 𝗖𝗼𝗿𝗲𝗪𝗲𝗮𝘃𝗲 签署了具有约束力的锚定租约。➤ 协议将 6 月 30 日的 𝗻𝗼𝗻-𝗯𝗶𝗻𝗱𝗶𝗻𝗴 𝗟𝗢𝗜（意向书）转化为正式合同。➤ 租约初始期限为𝟭𝟱年，并包含两次五年续租权。➤ 该园区正在升级以支持 𝗵𝗶𝗴𝗵-𝗱𝗲𝗻𝘀𝗶𝘁𝘆、液冷 AI 基础设施。➤ 设施主要由 𝗵𝘆𝗱𝗿𝗼𝗲𝗹𝗲𝗰𝘁𝗿𝗶𝗰 𝗽𝗼𝘄𝗲𝗿（水电）供电。➤ 场地具备现有的电网互联和 𝗿𝗲𝗱𝘂𝗻𝗱𝗮𝗻𝘁 𝗳𝗶𝗯𝗲𝗿（冗余光纤）基础设施。➤ 协议推进了 Blockfusion 拟议的与 𝗕𝗹𝘂𝗲 𝗔𝗰𝗾𝘂𝗶𝘀𝗶𝘁𝗶𝗼𝗻 𝗖𝗼𝗿𝗽. 的组合交易。➤ 合并后的公司预计将以 𝗕𝗗𝗜 为代码寻求纳斯达克上市。👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ CoreWeave 的长期承诺为该 AI 园区提供了 𝗮𝗻𝗰𝗵𝗼𝗿 𝗰𝘂𝘀𝘁𝗼𝗺𝗲𝗿（锚定客户）。➤ 正式协议降低了此前宣布的 𝗖𝗼𝗿𝗲𝗪𝗲𝗮𝘃𝗲 LOI 的不确定性。➤ 水电供应和现有基础设施支持 𝗵𝗶𝗴𝗵-𝗱𝗲𝗻𝘀𝗶𝘁𝘆 AI 部署。➤ 这笔交易是 Blockfusion 拟议的 𝗡𝗮𝘀𝗱𝗮𝗾 上市的重要一步。👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁𝘀:𝗔𝗹𝗲𝘅 𝗠𝗮𝗿𝘁𝗶𝗻𝗶-𝗟𝗼 𝗠𝗮𝗻𝘁𝗼, Blockfusion 首席执行官："这项协议是 Blockfusion 的一个决定性里程碑，也是对我们尼亚加拉瀑布园区的有力验证。我们将一个前工业遗址转变为下一代计算的基础设施，我们相信这可以成为该地区长期的技术和经济锚点。"𝗥𝗼𝗯𝗲𝗿𝘁 𝗥𝗲𝘀𝘁𝗮𝗶𝗻𝗼, 尼亚加拉瀑布城市市长：“关于数据中心的讨论最终关乎确保技术投资转化为当地社区的切实利益。Blockfusion 对尼亚加拉瀑布的承诺为居民、企业和后代创造了机会。我们很高兴看到有助于加强当地经济并为尼亚加拉瀑布的长期增长定位的投资。”H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: Intuitive Machines 完成 SXM-11 卫星向 SiriusXM 的移交 - $Intuitive Machines(LUNR.US) $天狼星XM广播(SIRI.US)
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗜𝗻𝘁𝘂𝗶𝘁𝗶𝘃𝗲 𝗠𝗮𝗰𝗵𝗶𝗻𝗲𝘀 完成 𝗦𝗫𝗠-𝟭𝟭 卫星向 SiriusXM 的移交。➤ 𝗦𝗫𝗠-𝟭𝟭 在成功进行在轨测试后，现已全面投入运营。➤ 该卫星于 2026 年 6 月 28 日搭乘 𝗦𝗽𝗮𝗰𝗲𝗫 𝗙𝗮𝗹𝗰𝗼𝗻 𝟵 发射升空。➤ SXM-11 采用 Intuitive Machines 的 𝗜𝗠 𝟭𝟯𝟬𝟬™ 系列平台。➤ 双方合作在二十多年间已制造了 𝟭𝟯 艘航天器。➤ SiriusXM 的下一颗卫星的研发工作已经启动。➤ SXM-11 增强了 SiriusXM 网络的 𝗿𝗲𝗮𝗰𝗵、可靠性和韧性。➤ 最近的收购扩大了 Intuitive Machines 的 𝗕𝘂𝗶𝗹𝗱-𝗖𝗼𝗻𝗻𝗲𝗰𝘁-𝗢𝗽𝗲𝗿𝗮𝘁𝗲 框架。➤ 公司正扩展为集成的 𝘀𝗽𝗮𝗰𝗲 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 主承包商。👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ 成功的移交证明了 Intuitive Machines 的 𝗰𝗼𝗺𝗺𝗲𝗿𝗰𝗶𝗮𝗹 𝘀𝗮𝘁𝗲𝗹𝗹𝗶𝘁𝗲 执行能力。➤ 持续的 SiriusXM 订单提供了 𝗿𝗲𝗽𝗲𝗮𝘁 𝗰𝘂𝘀𝘁𝗼𝗺𝗲𝗿 需求的证据。➤ 运营的 SXM-11 增强了 SiriusXM 的 𝘀𝗮𝘁𝗲𝗹𝗹𝗶𝘁𝗲 𝗻𝗲𝘁𝘄𝗼𝗿 韧性和覆盖范围。➤ 收购拓宽了 LUNR 的业务范围，从月球任务扩展到 𝘀𝗽𝗮𝗰𝗲 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲。👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁𝘀:𝗦𝗲𝗮𝗻 𝗚𝗶𝗯𝗯𝗼𝗻𝘀, SiriusXM 首席产品与技术官：“二十多年来，SiriusXM 和 Intuitive Machines 团队一直合作设计和交付多代大功率卫星，” SiriusXM 首席产品与技术官 Sean Gibbons 表示。“这颗最新卫星建立在 SXM-9 和 SXM-10 的成功基础之上，下一代卫星的研发工作已经在进行中。这种持续的合作正在加强我们下一代卫星网络的覆盖范围、可靠性和韧性，帮助我们为订阅者提供他们信赖的高端车载体验。”𝗖𝗵𝗿𝗶𝘀 𝗝𝗼𝗵𝗻𝘀𝗼𝗻, Intuitive Machines Space Systems 总裁：“SXM-11 标志着我们与 SiriusXM 长期合作的另一个里程碑。在我们的合作关系中，我们共同建造了 13 艘航天器，” Intuitive Machines Space Systems 总裁 Chris Johnson 表示。“这次移交反映了我们对任务成功的共同承诺，以及 SiriusXM 对我们团队提供高性能、可靠太空系统的信心。我们继续证明了我们设计、制造和交付满足当今商业太空市场严苛要求的卫星系统的能力，拥有强大、可扩展的大功率平台。”
H
📊 𝐈𝐍𝐕𝐄𝐒𝐓𝐎𝐑 𝐍𝐎𝐓𝐄: AI 基础设施：在 AI 革命背后增长的万亿美元市场 - $Applied Digital(APLD.US) $Hut 8 Mining(HUT.US) $IREN(IREN.US) $Nebius(NBIS.US)
H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $Credo Tech(CRDO.US) Credo Launches 1.6T ZeroFlap Optical Transceivers for AI Networks
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗖𝗿𝗲𝗱𝗼 launches 𝟮𝟮𝟰𝗚-based 𝟭.𝟲𝗧 ZeroFlap optical transceivers for AI networks.➤ Platform combines 𝗭𝗲𝗿𝗼𝗙𝗹𝗮𝗽 𝗗𝗦𝗣, 𝗞𝗳𝗶𝗿𝟮𝟬𝟬 silicon photonics, and PILOT diagnostics.➤ 𝗣𝗜𝗟𝗢𝗧 continuously monitors optical-link health and streams real-time telemetry.➤ Technology detects fiber issues and 𝗹𝗶𝗻𝗸 𝗳𝗹𝗮𝗽𝘀 before AI workloads suffer.➤ ZeroFlap supports 𝟮𝘅𝗗𝗥𝟰, 𝟮𝘅𝗙𝗥𝟰, and 𝗗𝗥𝟴 AI interconnect architectures.➤ Platform targets faster 𝘁𝗶𝗺𝗲 𝘁𝗼 𝗳𝗶𝗿𝘀𝘁 𝘁𝗼𝗸𝗲𝗻 and improved cluster uptime.➤ Telemetry supports 𝗦𝗢𝗡𝗶𝗖 and other leading switch operating systems.➤ Remote telemetry and firmware updates simplify large-scale AI network maintenance.➤ Cignal AI forecasts over 𝟭𝟭 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 1.6T transceivers shipping in 2026.➤ Cignal AI expects 2027 shipments to 𝗺𝗼𝗿𝗲 𝘁𝗵𝗮𝗻 𝗱𝗼𝘂𝗯𝗹𝗲 2026 levels.👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ 𝗖𝗿𝗲𝗱𝗼 expands its opportunity in rapidly scaling AI data-center connectivity.➤ 𝟭.𝟲𝗧 networking is emerging as critical infrastructure for larger AI clusters.➤ Proactive diagnostics could reduce 𝗱𝗼𝘄𝗻𝘁𝗶𝗺𝗲 and accelerate AI cluster deployment.➤ Forecast shipment growth signals potentially significant demand for 𝟭.𝟲𝗧 optical connectivity.👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁𝘀:𝗔𝗺𝗲𝗲𝘁 𝗦𝘂𝗿𝗶, VP of Product at Credo:“Getting an AI cluster online—and keeping it running—starts with a stable network,” said Ameet Suri, VP of Product at Credo. “ZeroFlap Optics constantly monitor every link with built-in telemetry that detects warning signs to notify operators and avoid costly link flaps. That means customers get their clusters running faster and keep them running longer.”𝗦𝗰𝗼𝘁𝘁 𝗪𝗶𝗹𝗸𝗶𝗻𝘀𝗼𝗻, Lead Analyst, Optical Components at Cignal AI:“The 1.6T optical transceiver ramp has begun, with over 11 million forecast to ship in 2026 and more than double that number shipping in 2027," said Scott Wilkinson, Lead Analyst, Optical Components at Cignal AI. “Credo's ZeroFlap transceivers address two critical priorities in scaling 1.6T AI infrastructure deployments reliably: optical link resilience and operational visibility. By harnessing real-time telemetry in the transceiver, Credo is enabling networks to predict and prevent failures before they impact services.”H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $Astera Labs(ALAB.US) Astera Labs Launches New Leo Memory Solutions for Agentic AI Infrastructure
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗔𝘀𝘁𝗲𝗿𝗮 𝗟𝗮𝗯𝘀 launches new 𝗟𝗲𝗼 𝗫-𝗦𝗲𝗿𝗶𝗲𝘀 and 𝗟𝗲𝗼 𝟮 memory controllers.➤ Leo X-Series targets 𝗞𝗩-𝗰𝗮𝗰𝗵𝗲-intensive, long-context agentic AI inference workloads.➤ Leo X-Series delivers up to 𝟲𝟮% faster time-to-first-token.➤ Platform delivers up to 𝟮𝟮% more tokens per second.➤ Leo 2 doubles 𝗺𝗲𝗺𝗼𝗿𝘆 𝗯𝗮𝗻𝗱𝘄𝗶𝗱𝘁𝗵 and capacity versus previous generation.➤ New controllers support 𝗗𝗗𝗥𝟰, 𝗗𝗗𝗥𝟱, 𝗣𝗖𝗜𝗲 𝟲 and 𝗖𝗫𝗟 𝟯.𝟮.➤ Leo 2 enables memory expansion, pooling, sharing, and 𝗗𝗗𝗥𝟰 reuse.➤ Astera reports growing design wins across 𝗔𝗜 𝗹𝗮𝗯𝘀, hyperscalers, and neoclouds.➤ New Leo products are 𝘀𝗮𝗺𝗽𝗹𝗶𝗻𝗴 𝘁𝗼𝗱𝗮𝘆 with hyperscaler customers.➤ Development involved 𝗔𝗠𝗗, 𝗔𝗿𝗺, 𝗜𝗻𝘁𝗲𝗹 and major memory suppliers.👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ Agentic AI is rapidly increasing demand for 𝗺𝗲𝗺𝗼𝗿𝘆 𝗰𝗮𝗽𝗮𝗰𝗶𝘁𝘆 and bandwidth.➤ Faster KV-cache access can improve 𝗔𝗜 𝗶𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲 economics and responsiveness.➤ Memory pooling and reuse could reduce expensive 𝗗𝗥𝗔𝗠 over-provisioning.➤ Growing design wins could expand Astera's exposure to 𝗿𝗮𝗰𝗸-𝘀𝗰𝗮𝗹𝗲 𝗔𝗜 infrastructure.👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁𝘀:𝗠𝗮𝘁𝘁 𝗞𝗶𝗺𝗯𝗮𝗹𝗹, Vice President and Principal Analyst at Moor Insights & Strategy:“Agentic AI adoption is outpacing the industry's ability to provision memory for it, and that gap is widening every quarter," said Matt Kimball, vice president and principal analyst at Moor Insights & Strategy. "That growth is landing at the worst possible time for memory supply, with DDR5 tight and pricing climbing, so every gigabyte already deployed has to work harder. The new Leo X-Series and Leo 2 E and P-Series Smart Memory Controllers immediately address KV-cache-intensive AI workloads, improve memory utilization, and reliably reuse previously deployed memory across both AI and general-purpose cloud infrastructure, instead of simply buying more of it."𝗧𝗵𝗮𝗱 𝗢𝗺𝘂𝗿𝗮, Senior Vice President, Compute Connectivity Group at Astera Labs:“Agentic AI is where the economics of AI infrastructure are being decided, and those economics depend on putting every usable gigabyte of memory to work,” said Thad Omura, senior vice president, Compute Connectivity Group at Astera Labs. “The enhanced Leo family gives infrastructure providers purpose-built ways to connect memory to accelerators, CPUs, and hosts across the rack—turning previously deployed and stranded capacity into a resource that new AI and cloud workloads can use. We're seeing that translate into broader customer engagement across AI labs, hyperscalers, and neoclouds.”𝗥𝗼𝗯𝗲𝗿𝘁 𝗛𝗼𝗿𝗺𝘂𝘁𝗵, Corporate Vice President, Architecture and Strategy at AMD:“Agentic AI is increasing demand for balanced systems that can move and process more data without sacrificing efficiency. AMD EPYC processors deliver leadership performance, exceptional memory bandwidth and capacity, and high-speed I/O across the broadly deployed x86 platform. Our collaboration with Astera Labs around CXL extends that foundation, giving customers greater flexibility to scale memory-intensive AI and cloud workloads and improve infrastructure utilization.”𝗘𝗱𝗱𝗶𝗲 𝗥𝗮𝗺𝗶𝗿𝗲𝘇, Vice President of Go-To-Market, Cloud AI, Arm:“As AI infrastructure evolves, agentic AI, reinforcement learning and database workloads are driving demand for higher-capacity memory. Arm and Astera Labs are working together to pair Arm AGI CPU with Leo CXL Smart Memory Controllers, helping customers meet these growing requirements while maintaining scalable performance.”𝗦𝗿𝗶𝗻𝗶 𝗞𝗿𝗶𝘀𝗵𝗻𝗮, Fellow at Intel Data Center Group:"CXL was created to bring greater flexibility and efficiency to memory across the data center. Technologies like Astera Labs' Leo Smart Memory Controllers help turn that vision into reality, demonstrating how ecosystem innovation can help customers get more out of their AI infrastructure and support the next generation of data-intensive workloads."𝗝𝗮𝗻𝗴𝘀𝗲𝗼𝗸 𝗖𝗵𝗼𝗶, Vice President of Product Planning Team at Samsung Electronics:"We are contributing to the evolution of AI infrastructure through the development of CXL memory solutions designed to help address the growing memory demands of modern data centers. We are excited to collaborate with Astera Labs to help ensure our DRAM solution offer strong interoperability within the evolving CXL ecosystem, and remain committed to supporting a scalable, high-performance memory architecture for the AI era."
H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $Grab(GRAB.US) GRAB 将以 14.9 亿美元收购 Atome Financial 60% 的股权
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗚𝗿𝗮𝗯 将收购 Atome Financial 的 𝗺𝗮𝗷𝗼𝗿𝗶𝘁𝘆 60% 股权。➤ 该交易使所收购股权的估值约为 $𝟭.𝟰𝟵 𝗯𝗶𝗹𝗹𝗶𝗼𝗻（14.9 亿）。➤ 股权将从 𝗔𝗱𝘃𝗮𝗻𝗰𝗲 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗚𝗿𝗼𝘂𝗽 及其他方手中收购。➤ 𝗔𝘁𝗼𝗺𝗲 在多个消费类别中提供先买后付 (BNPL) 服务。➤ Grab 目前目标是在 2028 年前将贷款规模提升至超过 $𝟲 𝗯𝗶𝗹𝗹𝗶𝗼𝗻（60 亿）。➤ 此前的目标是年底前超过 $𝟯 𝗯𝗶𝗹𝗹𝗶𝗼𝗻（30 亿）。➤ Atome 在 2025 年产生 $𝟰𝟳𝟬 𝗺𝗶𝗹𝗹𝗶𝗼𝗻（4.7 亿）收入，增长 80%。➤ Atome 实现了其 𝘀𝗲𝗰𝗼𝗻𝗱 𝗰𝗼𝗻𝘀𝗲𝗰𝘂𝘁𝗶𝘃𝗲（连续第二年）税前利润。➤ 此交易紧随 Grab 近期收购 𝗦𝘁𝗮𝘀𝗵 和 𝗙𝗼𝗼𝗱𝗽𝗮𝗻𝗱𝗮 𝗧𝗮𝗶𝘄𝗮𝗻 之后。👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ 此次收购显著扩大了 Grab 在 𝗰𝗼𝗻𝘀𝘂𝗺𝗲𝗿 𝗹𝗲𝗻𝗱𝗶𝗻𝗴（消费借贷）领域的敞口。➤ 新的 $𝟲 𝗯𝗶𝗹𝗹𝗶𝗼𝗻（60 亿）目标信号表明其金融服务野心大幅增强。➤ Atome 增加了一个快速增长且已 𝗽𝗿𝗼𝗳𝗶𝘁𝗮𝗯𝗹𝗲（盈利）的 BNPL 平台。➤ 该交易进一步使 Grab 的业务多元化，超越 𝗿𝗶𝗱𝗲-𝗵𝗮𝗶𝗹𝗶𝗻𝗴（网约车）和外卖配送。
H
吉姆·克莱默谈$Gemini Space Station(GEMI.US) Gemini Space Station："这是一个加密货币投机。它真的跌得太低了。我会把它作为你的投机标的。"
H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $Cipher Digital(CIFR.US) Cipher Digital, Fluidstack and Anthropic Commit $𝟭𝟬 𝗠𝗶𝗹𝗹𝗶𝗼𝗻 to Texas Water Infrastructure
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗖𝗶𝗽𝗵𝗲𝗿, 𝗙𝗹𝘂𝗶𝗱𝘀𝘁𝗮𝗰𝗸 and 𝗔𝗻𝘁𝗵𝗿𝗼𝗽𝗶𝗰 jointly commit $𝟭𝟬 𝗺𝗶𝗹𝗹𝗶𝗼𝗻.➤ Funding targets 𝗖𝗼𝗹𝗼𝗿𝗮𝗱𝗼 𝗖𝗶𝘁𝘆's aging water infrastructure in Texas.➤ Projects include additional 𝘄𝗮𝘁𝗲𝗿 𝘀𝘁𝗼𝗿𝗮𝗴𝗲 and replacement of aging valves.➤ Companies established the nonprofit 𝗖𝗼𝗹𝗼𝗿𝗮𝗱𝗼 𝗖𝗶𝘁𝘆 𝗪𝗮𝘁𝗲𝗿 𝗜𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 𝗙𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻.➤ Foundation will coordinate projects with 𝗖𝗼𝗹𝗼𝗿𝗮𝗱𝗼 𝗖𝗶𝘁𝘆 officials and engineers.➤ Investment follows chronic outages and July storm damage to water infrastructure.➤ Cipher's 𝗕𝗮𝗿𝗯𝗲𝗿 𝗟𝗮𝗸𝗲 data center uses its own water supply.➤ Facility draws 𝗻𝗼 𝘄𝗮𝘁𝗲𝗿 from Colorado City's municipal system.➤ Cipher says operations won't impact local drinking water or residential utility bills.👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ The $𝟭𝟬 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 commitment could improve local water reliability and resilience.➤ Investment addresses community infrastructure concerns surrounding expanding 𝗔𝗜 𝗱𝗮𝘁𝗮 𝗰𝗲𝗻𝘁𝗲𝗿 development.➤ Independent on-site water reduces potential competition with 𝗿𝗲𝘀𝗶𝗱𝗲𝗻𝘁𝗶𝗮𝗹 𝘄𝗮𝘁𝗲𝗿 demand.➤ Partnership highlights community investment alongside growing 𝗔𝗜 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 deployment.👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁𝘀:𝗗𝘂𝘀𝘁𝗶𝗻 𝗕𝘂𝗿𝗿𝗼𝘄𝘀, Texas House Speaker:"As Texas continues to grow, it is important that the companies helping drive that growth are also good neighbors and partners in our communities. This donation for Colorado City's water infrastructure will strengthen the region's water future while demonstrating a commitment to the Texans who call this community home. I want to thank the companies for their generosity and for investing in the long-term future of Colorado City and West Texas."𝗥𝘂𝗯𝗲𝗻 𝗛𝘂𝗿𝘁, Colorado City Mayor:"When we welcomed these companies to our community, they told us they would be good neighbors. This $10 million contribution for critical repairs to our local water system demonstrates they meant every word."𝗧𝘆𝗹𝗲𝗿 𝗣𝗮𝗴𝗲, Cipher's Chief Executive Officer:"Being a good neighbor means showing up for the issues that matter most to the communities where we operate, even when those issues predate us. In Colorado City, we wanted to invest directly in solutions that will benefit the community for years to come. We're grateful to Speaker Burrows and local leaders for their partnership, and proud that Colorado City is home to one of our data centers."𝗞𝗮𝘁𝗲 𝗙𝗿𝗮𝗻𝗸𝗼, Public Affairs Lead for Fluidstack:"Everywhere Fluidstack operates, we try to listen first and understand what the community actually needs. This summer, our neighbors here had a problem. We asked if we could help, and we're glad the answer was yes."𝗥𝘆𝗮𝗻 𝗪𝗶𝗹𝘀𝗼𝗻, Head of Economic Development for Anthropic:"Communities that host AI infrastructure should be better off for it. Colorado City's leaders identified what their water system needs, and we're glad to help fund it.”
H
瑞穗证券关于半导体板块回调的观点：“我们认为这是一个买入机会，$美光科技(MU.US)、$Lumentum控股(LITE.US)、$Credo Tech(CRDO.US)、$闪迪(SNDK.US)、$博通(AVGO.US)、$戴尔科技-C(DELL.US) 和 $泛林集团(LRCX.US) 依然处于有利地位。”
“在 Anthropic 和 OpenAI 的首席执行官发表文章呼吁在 Dario Amodei 的文章之后放缓 AI 发展，并推动对前沿模型采用的监管护栏后，我们审视了 AI 半导体领域。要点：1) Anthropic 和 OpenAI 以安全为由寻求可能放缓 AI 发展，美国参议员 David Sacks 对此的利他主义理由提出反驳；2) 我们认为，实施 RSI 限制加上缺乏全球 AI 共识或协调，意味着潜在的 AI 节奏控制无效，因为这存在将市场份额让给高效的中国竞争对手的风险（见第 2 页）；3) 关键观察点——AI 实验室和云服务商（CSPs）的 AI 数据中心资本支出/GW 计算支出没有变化，尽管 AI 采用仍处于早期阶段，但对 AI 半导体受益者仍然保持积极态度。随着 SOX 指数今早下跌约 6%，我们认为这是一个买入机会，MU、LITE、CRDO、SNDK、AVGO、DELL 和 LRCX 依然处于有利地位。”H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $Ondas(ONDS.US) Ondas（奥达斯）以高达 3.9 亿美元收购 GATE Technologies
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ Ondas 收购了𝗚𝗔𝗧𝗘 𝗧𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝗶𝗲𝘀和 Bron Technologies。➤ 交易包括 2.05 亿美元的预付款，以及 1.85 亿美元的潜在盈利支付（earn-out）。➤ 对价由 1.05 亿美元现金和 1 亿美元 Ondas 股票组成。➤ GATE 的技术已集成到超过 90 种武器系统中。➤ GATE 专注于𝗲𝗹𝗲𝗰𝘁𝗿𝗼𝗻𝗶𝗰 𝘀𝗮𝗳𝗲-𝗮𝗻𝗱-𝗮𝗿𝗺（电子保险与解保）和先进引信技术。➤ Ondas 预计 GATE 的 2026 年营收将达到 6500 万美元。➤ 预计到 2028 年营收将达到 1.8 亿美元。➤ 被收购业务预计在 2028 年前产生超过 1.3 亿美元的调整后 EBITDA。➤ 美国生产的成品预计将在 2027 年上半年到位。➤ 巡飞弹市场到 2030 年可能达到 133 亿美元。👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ 为 Ondas 的精确打击技术组合增添了𝗺𝗶𝘀𝘀𝗶𝗼𝗻-𝗰𝗿𝗶𝘁𝗶𝗰𝗮𝗹（任务关键型）组件。➤ 扩大了对𝗨.𝗦.和欧洲国防项目及客户的准入。➤ 波兰的生产提供了基于𝗡𝗔𝗧𝗢-𝗯𝗮𝘀𝗲𝗱、非 ITAR 的制造足迹。➤ 此次收购增加了 Ondas 对 rapidly expanding 𝗽𝗿𝗲𝗰𝗶𝘀𝗶𝗼𝗻-𝘀𝘁𝗿𝗶𝗸𝗲需求的敞口。➤ GATE 的合格设计在武器生产规模扩大时创造了潜在的经常性需求。👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁𝘀:𝗘𝗿𝗶𝗰 𝗕𝗿𝗼𝗰𝗸, Ondas 董事长兼 CEO："GATE 是 Ondas 极具战略意义的补充。全球正在重建弹药库存，新一代低成本自主弹药正成为战场日益重要的一部分。同时，市场正从传统的机械引信转向电子保险与解保装置，这对于更小、更智能且软件定义的武器至关重要。GATE 的 RUBI 产品线及其工程团队通过数十年的资格认证和集成到 90 多种武器系统中，在这个市场中确立了强劲的地位。凭借 Ondas 的全球营销和运营平台支持，我们相信 GATE 能够捕获有意义的市场份额。"𝗘𝗿𝗶𝗰 𝗕𝗿𝗼𝗰𝗸, Ondas 董事长兼 CEO："目前 ESAD 组件供应紧张，合格产能是整个精确打击供应链的瓶颈。GATE 赢得了全球广泛国防原始设备制造商（OEM）和武器项目的信任，其客户群优先。Ondas 致力于以低成本、先进的引信能力服务整个市场，我们将投资 GATE 的产能、工程和全球覆盖范围，使其能够支持更多客户和项目。结合 GATE 有吸引力的增长和利润率前景，我们预计此次收购将为我们的股东带来高额的投入资本回报。"𝗢𝘀𝗵𝗿𝗶 𝗟𝘂𝗴𝗮𝘀𝘀𝘆, Ondas 自主系统联席 CEO："电子保险与解保及引信系统是现代化精确武器的任务关键型组件。GATE 通过数十年的资格认证和在广泛系统中的集成，在该市场中建立了战略地位。其技术深入设计于客户平台中，形成了强大的进入壁垒，并使 GATE 能够从全球精确打击系统生产的扩张中受益。"
H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $英伟达 (NVDA.US) NVIDIA 扩展 CUDA-Q，新增用于容错量子计算的逻辑层
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗡𝗩𝗜𝗗𝗜𝗔 推出 𝗖𝗨𝗗𝗔-𝗤 𝗟𝗼𝗴𝗶𝗰𝗮𝗹，用于容错量子应用开发。➤ 该平台协调算法、纠错、架构和量子硬件。➤ 𝗙𝗲𝗿𝗺𝗶𝗹𝗮𝗯 将开发周期从五个月缩短至三周。➤ Fermilab 使用 CUDA-Q Logical 实现了 7 倍加速。➤ Iceberg 仅使用 150,000 个物理量子比特就模拟了 1,000 个逻辑量子比特。➤ 所需的物理量子比特数量比之前的估计降低了约 10 倍。➤ Sandia 的 𝗤𝗨𝗢𝗣𝗦 基准测试现已通过 NVIDIA CUDA-Q 提供。➤ QUOPS 衡量硬件对实际 𝗳𝗮𝘂𝗹𝘁-𝘁𝗼𝗹𝗲𝗿𝗮𝗻𝘁 量子应用的准备情况。➤ CUDA-Q Logical 现已通过 𝗚𝗶𝘁𝗛𝘂𝗯 公开可用。👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ 可能显著缩短 𝗽𝗿𝗮𝗰𝘁𝗶𝗰𝗮𝗹 𝗾𝘂𝗮𝗻𝘁𝘂𝗺 𝗰𝗼𝗺𝗽𝘂𝘁𝗶𝗻𝗴 的开发时间表。➤ 降低物理量子比特需求可能会减少未来量子硬件的成本。➤ 加强了 NVIDIA 在连接 𝗚𝗣𝗨 𝘀𝘂𝗽𝗲𝗿𝗰𝗼𝗺𝗽𝘂𝘁𝗶𝗻𝗴 与量子处理器方面的地位。➤ QUOPS 提供了一个开放基准，用于跟踪 𝗾𝘂𝗮𝗻𝘁𝘂𝗺 𝘂𝘁𝗶𝗹𝗶𝘁𝘆 进展。👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁𝘀:𝗧𝗶𝗺𝗼𝘁𝗵𝘆 𝗖𝗼𝘀𝘁𝗮，NVIDIA 量子计算副总裁兼总经理：“量子计算正逐渐进入逻辑量子比特时代，研究人员需要一个开放、可定制的平台，能够代表容错系统的所有方面。CUDA-Q Logical 的加入提供了强大的功能和灵活性，以探索完全集成、协同优化的系统，无论量子比特类型和架构如何——大幅缩短了通向实用量子-GPU 超级计算机的时间表。”𝗔𝗻𝗻𝗮 𝗚𝗿𝗮𝘀𝘀𝗲𝗹𝗹𝗶𝗻𝗼，Fermilab 首席技术官兼超导量子材料与系统中心主任：“容错量子计算是解锁新科学发现的路径，但实现这一目标需要研究人员共同设计算法、纠错、架构和硬件。使用 CUDA-Q Logical，我们的团队在三周内探索了这些资源的组合，而通常情况下，构建专用基础设施需要大约五个月的时间。”𝗧𝗶𝗺𝗼𝘁𝗵𝘆 𝗣𝗿𝗼𝗰𝘁𝗼𝗿，桑迪亚国家实验室量子性能实验室联合主任：“在桑迪亚，我们迫不及待地希望看到容错量子计算机帮助解决美国能源部和美国的国家重大问题。我们目前的使命是通过加速行业合作伙伴的进展来尽快实现这一目标。为此，我们和其他量子计算利益相关者必须能够跟踪和预测量子计算机能力的发展。我们创建了 QUOPS 来做到这一点，我们很高兴看到它被量子计算供应商和客户所使用。”H
📢 𝗝𝗨𝗦𝗧 𝗜𝗡: $EOS Energy Enterprises(EOSE.US) EOS Energy 获得美国能源部 8700 万美元预付款，以扩大美国电池制造规模
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗘𝗼𝘀 𝗘𝗻𝗲𝗿𝗴𝘆 获得 8700 万美元 DOE 贷款预付款。➤ 该预付款报销 Thorn Hill 制造成本的 80%。➤ 自 2024 年以来，Eos 已提取约 1.78 亿美元。➤ Line 2 于 2026 年 6 月进入商业化生产。➤ Line 2 目标年产能约为 2 GWh。➤ Thorn Hill 最终可能支持每年约 4 GWh 的产能。➤ 计划中的 Line 1 搬迁仍取决于贷款批准。➤ Eos 预计将为 Line 2 的 4 个班次增加招聘。➤ 资金用于支持扩大美国制造的锌基长时储能规模。👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ 8700 万美元的退款将之前投入 Eos 的资本返还给公司。➤ 额外的流动性可支持进一步的制造扩张和业务投资。➤ 4 GWh 的产能将显著扩大 Eos 的国内生产足迹。➤ 整合两条产线可提高生产力并降低制造成本。👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁𝘀:𝗔𝗹𝗲𝘀𝘀𝗮𝗻𝗱𝗿𝗼 𝗟𝗮𝗴𝗶, Eos 首席财务官：“来自能源主导融资办公室的贷款资金对于扩大 Eos 规模至关重要，” Eos 首席财务官 Alessandro Lagi 表示。“这笔预付款报销了我们在 Line 2 上已经做出的大量投资，并将这些资本返还至资产负债表，使我们在保持稳健增长方法的同时，有更多的空间进行业务投资。”𝗝𝗼𝗵𝗻 𝗠𝗮𝗵𝗮𝘇, Eos 首席运营官：“在同一屋檐下运行两条产线能更有效地利用我们的工程、支持资源和劳动力，” Eos 首席运营官 John Mahaz 表示。“这使得运营能够提高生产力并优化制造成本。”
H![图片 1，共 1 张]()
$Bloom Energy(BE.US) | 瑞穗维持对 bloom energy corp. 的跑赢大盘评级，将目标价从 $242 上调至 $351
分析师认为，更强劲的价格和需求，以及更好的交付可见性，将支撑更高的长期平均售价 (ASP) 和 EBITDA 利润率。
H
⚡ 𝐔𝐏𝐃𝐀𝐓𝐄: $Bitdeer Tech(BTDR.US) Bitdeer AI Secures 65.1MW Malaysia AI Data Center Under 10-Year Agreement
👉 𝗞𝗲𝘆 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀:➤ 𝗕𝗶𝘁𝗱𝗲𝗲𝗿 𝗔𝗜 secures 𝗔𝟮𝟬𝟮, a 𝟲𝟱.𝟭𝗠𝗪 Malaysia AI data center.➤ A202 is secured under a 𝟭𝟬-𝘆𝗲𝗮𝗿 data center services agreement.➤ Energization is expected in 𝗤𝟯 𝟮𝟬𝟮𝟳.➤ A202 supports 𝗡𝗩𝗜𝗗𝗜𝗔 𝗚𝗕𝟯𝟬𝟬 𝗡𝗩𝗟𝟳𝟮 and 𝗩𝗲𝗿𝗮 𝗥𝘂𝗯𝗶𝗻 platforms.➤ Johor campus capacity will reach 𝟴𝟲.𝟴𝗠𝗪 including A201.➤ Secured AI Cloud capacity now totals approximately 𝟮𝟬𝟲.𝟱𝗠𝗪 globally.➤ That represents approximately 𝟱𝟵% of Bitdeer's 𝟯𝟱𝟬𝗠𝗪 Q1 2028 target.➤ Bitdeer AI's active AI cloud pipeline now exceeds $𝟳 𝗯𝗶𝗹𝗹𝗶𝗼𝗻.➤ GPU deployment financing is expected to include 𝟱𝟬%+ customer prepayments.➤ A102's 9.5MW commitments represent over $𝟴𝟬𝟬 𝗺𝗶𝗹𝗹𝗶𝗼𝗻 expected revenue.👉 𝗪𝗵𝘆 𝗧𝗵𝗶𝘀 𝗠𝗮𝘁𝘁𝗲𝗿𝘀:➤ A202 marks Bitdeer AI's 𝗹𝗮𝗿𝗴𝗲𝘀𝘁 Southeast Asian capacity addition yet.➤ Expansion significantly advances Bitdeer's 𝟯𝟱𝟬𝗠𝗪 AI infrastructure target.➤ Existing Johor infrastructure could accelerate deployment while limiting new-site complexity.➤ $𝟳+ 𝗯𝗶𝗹𝗹𝗶𝗼𝗻 pipeline signals substantial demand for AI cloud capacity.➤ NVIDIA-ready infrastructure positions Bitdeer for next-generation 𝗔𝗜 𝗰𝗼𝗺𝗽𝘂𝘁𝗲 demand.👉 𝗘𝘅𝗽𝗲𝗿𝘁 𝗦𝘁𝗮𝘁𝗲𝗺𝗲𝗻𝘁𝘀:𝗠𝗶𝗰𝗵𝗮𝗲𝗹 𝗚. 𝗣𝗼𝘁𝘁𝗲𝗿, Chief Financial Officer:“The signal from customers is unambiguous: demand for liquid-cooled, rack-scale AI Cloud capacity in 2027 is running well ahead of what the market can supply,” said Michael G. Potter, Chief Financial Officer. “A102 sold out ahead of energization, A201 is in advanced negotiations, and our secured AI Cloud data center capacity now stands at approximately 206.5MW. Securing A202 is a direct response to that demand — an incremental 65.1MW on a campus we already have contracted for.”𝗥𝗲𝘁𝗮𝗶𝗻𝗻𝗮 𝗟𝗶𝗻, VP of AI Cloud:“Johor is where we can move fastest,” said Retainna Lin, VP of AI Cloud. “A201 is already secured at the data center level and is scheduled to be ready for service in January 2027, and A202 sits on the same campus — the same power, the same cooling, the same network. That is how we added 65.1MW without originating a new site. For our customers, it means liquid-cooled, rack-scale capacity available in Southeast Asia on a timeline that few providers in this region can offer.”