Alphabet (GOOG 0.34%) (GOOGL 0.35%) is currently out of favor with the market. Investors are concerned about the impact of generative AI technologies on Alphabet's core business, as evidenced by the decline in its stock. However, I think there's an under-the-radar reason that Alphabet's stock could soar in the second half of 2025, and investors would be wise to get in on it now before the stock really takes off. We'll hear more from Alphabet about the Google Search engine's health during its Q2 earnings rep ...

高盛发表研报，首次覆盖混合云服务商Nutanix(NTNX.US)，给予"买入"评级，目标价为95美元。 高盛指出，Nutanix是企业从传统基础设施向现代化转型的核心标的，其超融合基础设施(HCI)平台通过 将计算、存储和网络整合为软件定义堆栈，为本地环境提供类云服务。相较于传统架构，Nutanix解决 方案可降低40%以上的总体拥有成本，实现更快的资源调配和混合部署灵活性。 高盛认为，在博通(AVGO.US)收购威睿导致产品定价、支持服务和软件包变更后，Nutanix有望在未来 四年(2025至2029财年)每年获得3-6亿美元的经常性收入，承接VMware客户迁移需求。 报告还提及Nutanix与戴尔(DELL.US)、思科(CSCO.US)及Pure Storage(PSTG.US)日益扩大的合作伙伴生 态，以及其向混合多云编排平台的演进。该平台正在解决云迁移、基础设施自动化和AI支持等企业IT核 心需求。 近期动态显示，Nutanix2025财年第三季度业绩超预期，每股收益0.42美元(预期0.38美元)，营收6.39亿 美元(预期5.9444亿美元)，同比增长22%。Piper Sandler将其 ...

机器之心报道 编辑：Panda 然而，相比于预训练大模型，目前通常用于后训练大模型的强化学习存在自己独有的困难：不仅需要高频的数据交互和环境反馈，还要求大规模算力集群具备稳 定、弹性的调度能力。而传统的云计算平台，大多仍以静态推理负载为主，难以适配强化学习训练中动态多阶段、资源高耦合的运行特性。 这一背景下，谁能率先打造出真正适配强化学习的大规模智能计算平台，谁就有机会在新一轮的 AI 基础设施竞争中占据高地。2025 年 6 月，九章云极正式发布 业界首个工业级强化学习云平台 AgentiCTRL ，这也是 全球首个支持万卡级异构算力调度的强化学习基础设施平台 。 AgentiCTRL 基于混合专家（MoE）架构，仅需极少代码即可完成 AI 智能体的训推流程，从而显著增强大模型推理能力。相比于传统强化学习方案，AgentiCTRL 可将端到端训练效率提升 500%，综合成本下降 60%。 在这个云服务纷纷向「 AI 云原生 」转型的时代，九章云极率先跑通了强化学习大规模云化的全链路路径，可以说是为更进一步的「 智能体原生云 」树立了行业 新范式。 为什么是九章云极？ 领先源自系统级重构 本质上讲， 强化学习 ...

Core Insights - Vonage has announced a collaboration with Amazon Web Services (AWS) to integrate Amazon Nova Sonic, a speech-to-speech foundation model, with the Vonage Voice API, enhancing customer engagement and operational efficiency [1][3] - The integration will be showcased at the AWS Summit in New York City on July 16, 2025, allowing businesses to deploy real-time, natural-sounding AI voice agents across various channels [1][7] Group 1: Technology Integration - The next-generation AI voice technology simplifies the creation of responsive conversational agents by combining Vonage's infrastructure with Amazon Nova Sonic's capabilities [2][5] - The partnership eliminates challenges associated with building AI-driven conversational platforms, providing an all-in-one solution for businesses [5][6] Group 2: Business Applications - The integrated solution allows businesses to modernize IVR systems and deploy smart voice-enabled AI agents, improving customer experiences with immediate, natural language responses [4][6] - Use cases include customer support automation, proactive outbound engagement, and multilingual communication, enhancing overall customer outreach and satisfaction [6][5] Group 3: Strategic Importance - This collaboration positions businesses to accelerate their AI strategy, enhance customer satisfaction, and maintain a competitive edge in the market [7][3] - Vonage's offerings, including UCaaS and CCaaS, enable companies to transform their communication and operations, further supported by the integration with AWS [9][8]

Key Takeaways IBM benefits from hybrid cloud and AI demand, boosting Software and Consulting segment performance. IBM trades at a lower P/E than AMZN and has outperformed the industry with a 52.6% stock gain in a year. AMZN projects 2025 EPS growth of 12.7%, aided by AWS expansion and rising enterprise AI adoption.International Business Machines Corporation (IBM) and Amazon.com, Inc. (AMZN) are leading players in the global cloud computing industry. IBM offers cloud and data solutions that aid enterprises ...

Most financial services leaders believe GenAI solutions help improve levels of productivity, automation, and efficiency. Real value lies beyond chatbots and content tools GenAI Adoption Is Widespread but Not Without Risk: Nearly all industry respondents report using some form of GenAI today, with current use cases focused on customer support, content generation, and automation. However, data privacy and security stand out as the top concerns, with 97% agreeing their organizations must do more to secure GenA ...

Core Insights - The report by IDC indicates that the Chinese edge cloud market is projected to reach a total scale of 7.39 billion RMB in the second half of 2024, with a year-on-year growth rate of 18.6% [1] - Key drivers for market growth include internet audio and video distribution, real-time interaction, diversified AI training and inference, cloud network services, cloud gaming, and lightweight IT architecture migration for local industry clients [1][2] Market Overview - In the second half of 2024, lightweight model training and related processing in edge environments have become focal points for clients, stimulating rapid growth in the public and dedicated edge cloud service segments, maintaining an annual growth rate close to 20% despite a slowdown in existing demand from the internet sector [3] - Although large-scale data centers have deployed significant heterogeneous computing resources, the actual use of edge heterogeneous resources and products by industry clients remains in the early stages [3] - Major service providers are accelerating the expansion of edge nodes in hotspot areas, offering basic heterogeneous computing leasing, scheduling, and inference API services, while developing two clear growth paths: building large model API aggregation platforms and enhancing edge AI ecosystem collaboration [3][4] Market Outlook - The IDC has updated its long-term growth expectations for the edge cloud market, forecasting a compound annual growth rate of 20.3% from 2024 to 2029 [7] - The focus of clients and service providers has shifted towards edge AI and overseas edge scenarios, which are expected to bring significant incremental growth to the edge cloud market, potentially impacting existing resource structures and service output systems [9] - If services based on multimodal large models and audio-video AI can achieve large-scale promotion in the consumer market, edge cloud will continue to play a unique role in alleviating bandwidth pressure on core backend architecture and reducing long-term operational costs [9]

Core Viewpoint - Microsoft is restructuring its digital native business in the Asia-Pacific region, with Tian Zhuo taking charge, excluding India, indicating a consolidation of operations and a trend of centralizing authority within the company [1][2][3]. Group 1 - Tian Zhuo has been appointed as the head of Microsoft's digital native line in the Asia-Pacific region, overseeing all areas except India [1][2]. - Tian Zhuo has a background in AWS and Google Cloud, where he focused on gaming industry clients before joining Microsoft to lead the digital native division [2]. - The digital native business in Greater China remains under Tian Zhuo's leadership, with no new successor appointed, suggesting a merging of operations with the Asia-Pacific division [2]. Group 2 - Since last year, Microsoft has been gradually transferring parts of the digital native business from various Asia-Pacific countries to the Asia-Pacific region, reflecting a global trend of centralizing power [3].

Investors shouldn't sleep on one of the leading neocloud operators.It's no secret that artificial intelligence (AI) has come into its own over the past few years, and some of the biggest names in technology have benefited greatly from advances in generative AI. Palantir and Nvidia are among the biggest winners, soaring 2,180% and 1,030%, respectively, since early 2023. With gains of that magnitude, it isn't surprising that investors are turning over every stone looking for the next big AI winner.One area th ...

Nebius Group NBIS stock surged on Monday after Goldman Sachs analyst Alexander Duval initiated coverage with a Buy rating and announced a price target of $68.Duval noted that Nebius is emerging as a key player in the fast-growing AI Neoclouds space — a niche segment within the broader GPU-as-a-Service (GPUaaS) market. The analyst pointed out that this market allows AI startups, enterprises and hyperscalers to rent GPU infrastructure remotely via the cloud to run large-scale AI training and inference workloa ...