8月29日消息，英伟达（NASDAQ: NVDA）今晨公布截至2025年7月27日的2026财年第二季度（2025年 自然年二季度）业绩报告。 财报显示，第二季度，英伟达营收467.43亿美元，环比增长6%，同比增长56%，略超此前分析师预期 462.3亿美元；净利润（GAAP）264.22亿美元，同比增长57%；毛利率达72.4%，相比去年同期的75.1% 下降了2.7个百分点；稀释每股收益1.08美元，比去年同期增长61%。 其中，二季度英伟达数据中心营收为411亿美元，较上一季度增长5%，较去年同期增长56%。而 NVIDIA Blackwell数据中心的营收环比增长17%。 英伟达创始人兼CEO黄仁勋表示："Blackwell是全世界翘首以盼的AI平台，它实现了非凡的跨越式发展 ——Blackwell Ultra的产量正在全速提升，市场需求也异常旺盛。NVIDIA NVLink机架级计算技术具有 AI芯片一哥英伟达（NVIDIA）上一季度业绩持续增长，但增速低于预期。 革命性，它的到来恰逢推理AI模型推动训练和推理性能数量级提升的时代。AI竞赛已拉开帷幕，而 Blackwell正是这场竞赛的核心平台。 ...

Core Insights - NVIDIA is collaborating with European nations and industry leaders to develop the Blackwell AI infrastructure, aiming to enhance digital sovereignty and economic growth in Europe [1][14] - The initiative will provide over 3,000 exaflops of computing resources for sovereign AI, enabling secure development and deployment of AI applications across various sectors [3][15] Group 1: National Collaborations - France, Italy, Spain, and the U.K. are key nations involved in building domestic AI infrastructure, partnering with technology and telecommunications providers [2][11] - In France, Mistral AI is developing a cloud platform powered by 18,000 NVIDIA Grace Blackwell systems, with expansion plans for 2026 [7] - The U.K. plans to deploy 14,000 NVIDIA Blackwell GPUs to enhance AI capabilities for businesses [8] - Germany is establishing the world's first industrial AI cloud for manufacturers, utilizing 10,000 NVIDIA Blackwell GPUs [9] - Italy is advancing its AI capabilities through collaboration with Domyn and NVIDIA, focusing on regulated industries [10] Group 2: AI Technology Centers - NVIDIA is expanding AI technology centers in Germany, Sweden, Italy, Spain, the U.K., and Finland to foster research and workforce development [4][13] - These centers will support various research fields, including digital medicine and embodied AI, and provide training through the NVIDIA Deep Learning Institute [21] Group 3: Telecommunications Partnerships - NVIDIA is partnering with leading European telecommunications companies to create secure and scalable AI infrastructure [11][12] - Companies like Orange, Fastweb, and Telefónica are developing enterprise-grade AI solutions using NVIDIA's infrastructure [16]

Core Viewpoint - Cerebras has developed the world's largest computer chip, the Cerebras WSE, which integrates an impressive 4 billion transistors and achieves AI inference speeds that are approximately 2.5 times faster than comparable NVIDIA clusters [1][4]. Group 1: Chip Specifications and Performance - The Cerebras WSE measures 8.5 inches (22 cm) on each side and has set a world record for AI inference speed, processing 2,500 tokens per second, surpassing NVIDIA's Llama 4, which reached 1,000 tokens per second [1][4]. - The WSE's performance is attributed to its 4 billion transistors, which is significantly higher than Intel's Core i9 with 3.35 billion transistors and Apple's M2 Max with 6.7 billion transistors [4]. - The chip features 44GB of the fastest RAM, allowing for integrated computing without the need for external processing, which is a limitation in NVIDIA's architecture [4][5]. Group 2: Evolution of Chip Technology - The WSE represents a significant evolution in chip design, moving beyond traditional CPU dominance and GPU reliance, introducing a new GPU-accelerated architecture that is not based on x86 or ARM [5]. - This development is characterized as a leap rather than an incremental improvement in technology, indicating a transformative shift in the semiconductor industry [5]. Group 3: Market Implications - The speed of AI engines is becoming increasingly critical as businesses seek to implement AI solutions that can handle complex, multi-step tasks efficiently [3][4]. - Independent verification from Artificial Analysis confirmed the WSE's speed claims, stating it outperformed NVIDIA's Blackwell in inference solutions for Meta's flagship models [4][5].

Core Insights - NVIDIA announced that leading CAE software vendors, including Ansys, Altair, Cadence, Siemens, and Synopsys, are enhancing their simulation tools by up to 50 times using the NVIDIA Blackwell platform [1][2] - The integration of NVIDIA Blackwell with CUDA-X libraries allows industries such as automotive, aerospace, energy, manufacturing, and life sciences to significantly reduce product development time, cut costs, and improve design accuracy while maintaining energy efficiency [2][3] Ecosystem Support - A growing ecosystem of software providers is integrating Blackwell into their offerings, including companies like Altair, Ansys, Cadence, Siemens, and Synopsys, enabling customers to develop real-time digital twins with enhanced interactivity [4][3] - Rescale has launched a CAE Hub that streamlines access to NVIDIA technologies and CUDA-accelerated software, providing high-performance computing and AI technologies in the cloud powered by NVIDIA GPUs [8] Industry Applications - Cadence is utilizing NVIDIA Grace Blackwell-accelerated systems to tackle challenges in computational fluid dynamics, achieving multibillion cell simulations in under 24 hours, which previously required extensive CPU resources [5][6] - Boom Supersonic plans to use NVIDIA Omniverse Blueprint and Blackwell-accelerated CFD solvers on Rescale CAE Hub to design and optimize its new supersonic passenger jet, enabling 4 times more design explorations [9][10] Performance Enhancements - The collaboration between NVIDIA and various software providers is leading to significant performance improvements, with GPU-based simulations being up to 1.6 times faster compared to previous generations [7] - The combination of NVIDIA Blackwell architecture with Siemens' digital twins is expected to drastically reduce development times and costs, enhancing efficiency in design and manufacturing processes [7]