据媒体援引消息人士报道，芯片巨头英伟达(NVDA.US)计划发布一款全新处理器，专门为人工智能 （AI）研究公司OpenAI及其他客户打造，以帮助他们构建更快速、更高效的工具。 知情人士透露，英伟达正在设计一套新的推理计算系统。这一新平台预计将在下个月于圣何塞举行的英 伟达GTC开发者大会上发布，并将整合由初创公司Groq设计的芯片。 推理计算是一种让AI模型能够对用户提问作出响应的处理方式，该领域已成为行业激烈竞争的焦点。 谷歌和亚马逊等公司已经设计出与英伟达旗舰系统相竞争的芯片。 而科技行业中自动化编程的迅猛发展，也催生了对新型芯片的需求，这些芯片需要更高效地处理复杂的 AI相关任务。 知情人士表示，OpenAI已同意成为这款新处理器的最大客户之一，这对英伟达而言是一项重大胜利。 英伟达Hopper、Blackwell和Rubin系列GPU被认为是训练超大型AI模型的行业标杆产品，价格也处于高 位。 然而，自AI热潮兴起以来，英伟达首次面临其旗舰产品的局限性。随着市场重心从训练转向推理，一 些客户开始向英伟达施压，要求其推出更高效驱动AI应用的芯片。 过去一年，随着企业部署AI代理及其他工具，先进算力的需 ...

Nvidia - Nvidia has posted a remarkable return of 1,170% since January 2023, yet analysts believe the stock remains undervalued with a median target price implying a 43% upside from the current price of $182 [1][9] - The company reported a 73% increase in revenue to $68 billion in the fourth quarter, driven by significant growth in data center sales, particularly in compute (58%) and networking (263%) segments [5] - Nvidia's gross margin expanded by 1.7 percentage points, indicating strong pricing power despite competitive pressures [5] - The market is underestimating Nvidia's potential in the AI infrastructure market, with adjusted earnings expected to grow at 49% annually over the next two years, making its current valuation of 39 times earnings appear cheap [7] Robinhood Markets - Robinhood has seen an 875% return since January 2023, with a median target price suggesting a 62% upside from the current price of $76 [1][9] - The company is well-positioned for growth, having nearly double the number of millennial and Gen Z accounts compared to its closest competitor, as this demographic is expected to inherit over $100 trillion in assets [10] - Robinhood is gaining market share across various brokerage services, including cryptocurrency and equities, and has rapidly expanded its prediction markets segment [11] - The introduction of AI features through its Cortex suite, including personalized portfolio insights, is expected to enhance user engagement and drive growth [12] - Despite a 48% decline from its peak, Robinhood's total platform assets increased by 67%, indicating strong user engagement with equities and options [13]

Core Viewpoint - NVIDIA's stock is experiencing a decline as investors shift focus from strong earnings to uncertainties surrounding growth in 2027 [1][2]. Financial Performance - NVIDIA reported a record full-year revenue of $215.9 billion, reflecting a 65% increase from the previous year [6]. - The company achieved revenue growth of approximately 73% in its January quarter and guided for roughly 79% growth at the high end for the April quarter, exceeding prior expectations [5][6]. Market Reaction - Following the earnings report, NVIDIA's shares fell about 9% in the two sessions, contrasting with a 2% decline in the Nasdaq Composite [2]. - Despite strong fundamentals, investor anxiety may persist, potentially keeping the stock rangebound [5]. Growth Projections - Street estimates for NVIDIA's growth in 2027 suggest about 31% growth, an increase from earlier projections [2][3]. - The debate among investors centers on whether these growth estimates are too low or too uncertain to support current valuations [3]. Industry Insights - Data center sales now account for nearly 90% of NVIDIA's revenue, with hyperscalers making up about half of the business [4]. - The AI infrastructure cycle is still in its early stages, with rising capital spending plans from major cloud players indicating strong demand for advanced AI chips [4]. Stock Performance - As of the latest trading session, NVIDIA shares closed at $177.19, down 4.16%, with a 7.41% decline over the past five days [6]. - NVDA ranks strongly on quality in Benzinga's Edge Stock Rankings, showing a positive long-term price trend but weaker trends in the short and medium term [7].

英伟达正计划推出一款新型处理器，专为帮助OpenAI等客户开发更快、更高效的AI工具而设计。此举 有望重塑AI竞争格局。 知情人士透露，英伟达正在开发一套全新的"推理"计算系统——即AI模型响应用户查询的处理方式。该 平台将于下月在圣何塞GTC开发者大会上亮相，并将采用初创公司Groq设计的芯片。 股票频道更多独家策划、专家专栏，免费查阅>> 责任编辑：山上 据称，OpenAI已同意成为这款新处理器的最大客户之一，这对英伟达来说是一次重要胜利。 ...

2月28日，据华尔街日报援引消息人士称，英伟达计划推出一款全新的处理器，该处理器专为帮助 OpenAI及其他客户开发更快速、更高效的工具而设计。此举将对其业务产生重大影响，并有望重塑AI 领域的竞争格局。 美股频道更多独家策划、专家专栏，免费查阅>> 责任编辑：山上 知情人士透露，英伟达正在设计一套全新的"推理"计算系统。推理计算是一种能够让AI模型响应查询的 处理方式。知情人士称，该新平台将于下月在圣何塞举行的英伟达GTC开发者大会上亮相，并将采用由 初创公司Groq设计的芯片。一些知情人士表示，OpenAI已同意成为这款新处理器的最大客户之一，这 对英伟达而言无疑是一项重大胜利。 ...

IT之家 2 月 28 日消息，科技媒体 Wccftech 昨日（2 月 27 日）发布博文，报道称联发科进军硅光子技术领域，通过子公司斥资约 9000 万美元（IT之家注： 现汇率约合 6.18 亿元人民币）入股硅光子初创公司 Ayar Labs。 联发科通过其子公司 Digimoc Holdings Limited，向硅光子初创公司 Ayar Labs 投资了约 9000 万美元，借此交易获得了 172 万股特别股，占据 Ayar Labs 约 2.4% 的股份。 该媒体解读认为，英伟达、英特尔和 AMD 此前均已参股该公司，联发科此次入局显然意在硅光子技术逐渐成熟之际，为自己争取到下一代芯片制造领域的 关键话语权。 通过部署光学链路，I/O 功耗最高可以降低 80%，同时为边缘 AI 和即将到来的 6G 蜂窝网络解锁更高的数据吞吐量。此外，硅光子技术还可应用于共封装 光学（CPO）领域，将光学引擎直接集成在 SoC 封装内，从而缩短电气走线并减少延迟与发热。 由于光信号的传输速度远超电信号，且几乎不产生热量，这项技术在数据包传输时能实现更高的带宽、更低的延迟以及显著降低的能耗。 Ayar Labs 专门 ...

Group 1 - Nvidia plans to launch a new processor aimed at enhancing the speed and efficiency of AI systems for OpenAI and other customers [1] - The report about Nvidia's new chip was cited by the Wall Street Journal, indicating a focus on AI processing capabilities [1]

Core Insights - The chip giant is responding to competitive pressure by launching a new product aimed at enhancing the rapid processing of AI queries specifically for 'inference' demand [1] Group 1 - The new product is designed to address the growing need for efficient AI processing capabilities in the market [1] - The focus on 'inference' demand indicates a strategic shift towards applications that require quick decision-making and data analysis [1] - This move is likely a direct response to increasing competition within the semiconductor industry, particularly in the AI sector [1]

【长江电子】Feynman问世在即，LPU芯片开启PCB又一增长极 英伟达计划在GTC 2026大会上推出Feynman架构芯片，该产品的发布节奏较市场预期有所提前。 这一架构的核心技术突破在于采用3D堆叠方式，将专为推理任务优化的LPU芯片直接集成在GPU计算核心之上， 从而实现通用计算与专用计算在物理层面的深度融合。 新架构LPU芯片主要用于推理，以高多层方案为主，单芯片PCB价值量有望达到300-500美金，核心供应商建议关 注#胜宏科技、沪电股份、深南电路、景旺电子。 #坚定看好高确定性CoWoP技术+正交背板方案 CoWoP方案有望提前至27年底小批量、28年大批量，PCB单平米价值量或提升数倍至十倍，该方向下推荐关注 正交背板当前仍在正常稳步推进中，3月初计划进行新一轮样品测试，该方案预计在27年H2步入批量生产阶段，该 方向下看好低估值龙头公司，性价比凸显，推荐关注 英伟达计划在GTC 2026大会上推出Feynman架构芯片，该产品的发布节奏较市场预期有所提前。 这一架构的核心技术突破在于采用3D堆叠方式，将专为推理任务优化的LPU芯片直接集成在GPU计算核心之上， 从而实现通用计算与专用计算 ...

从训练走向极致推理—LPU架构重塑算力底座【东北计算机】 技术核心： 不同于GPU依赖HBM，LPU倾向于采用大规模片上SRAM直接存储模型参数，消除了内存访问延 迟；同时利用静态时序调度，将计算路径精确锁定在时钟周期内。 这种ASIC化设计旨在追求推理端的绝对高吞吐与低延迟。 范式转移：推理端的"低延迟革命"催生LPU架构 随着大模型进入大规模应用期，算力需求正从"暴力计算"向"极致交互"演进。 传统的GPU架构在处理LLM推理的Decode阶段时，往往面临高延迟瓶颈。 LPU（Language ProcessingUnit）架构应运而生。 技术核心： 不同于 从训练走向极致推理—LPU架构重塑算力底座【东北计算机】 范式转移：推理端的"低延迟革命"催生LPU架构 随着大模型进入大规模应用期，算力需求正从"暴力计算"向"极致交互"演进。 传统的GPU架构在处理LLM推理的Decode阶段时，往往面临高延迟瓶颈。 LPU（Language ProcessingUnit）架构应运而生。 产业链正加速转向M9级以上基材，其核心标准在于： 树脂端： 必须使用极低损耗的特种树脂体系。 电子布： 传统的玻璃布在介电一致 ...