Nvidia(NVDA)
Search documents
明年超越英伟达Rubin?400亿国产GPU大消息
Zhong Guo Ji Jin Bao· 2026-01-26 15:20
2025年,天数天枢架构超越Hopper; 2026年,天数天璇架构对标Blackwell; 2026年,天数天玑架构超越Blackwell; 2027年,天数天权架构超越Rubin; 2027年之后,将转向突破性计算芯片架构设计。 未来3年,天数智芯将基于此次发布的四代架构,陆续发布多款产品,持续提升计算性能。 单天逸还介绍了公司四代架构的关键细节: 天数天枢架构,支持从高精度科学计算到AI精度计算,AI芯片在执行注意力机制相关计算时,算力的实际有效利用效率达到90%以上; 天数天璇架构,新增ixFP4精度支持; 天数天玑架构,实现全场景AI与加速计算覆盖; 天数天权架构,融入更多精度支持与创新设计。 据介绍,基于相关核心技术创新,目前天数智芯已实现在DeepSeek V3场景平均比Hopper架构高约20%性能。 天数智芯董事会主席盖鲁江同时兼任公司CEO,其于2020年7月加入公司,并于2025年5月调任执行董事,主要负责公司的日常管理及整体策略规划。在此 之前,盖鲁江拥有约17年财务及投资经验。 2026年1月8日,天数智芯在港股上市。1月26日,天数智芯股价以188.2港元/股收盘,大跌7.65%, ...
明年超越英伟达Rubin?400亿国产GPU大消息
中国基金报· 2026-01-26 15:19
【导读】天数智芯发布四代架构路线图 中国基金报记者 卢鸰 在港股上市不久的国产GPU公司天数智芯,发布了一份标注日期的、超越英伟达相关产品的 路线图。 2026年,天数天璇架构对标Blackwell; 2026年,天数天玑架构超越Blackwell; 2027年,天数天权架构超越Rubin; 2027年之后,将转向突破性计算芯片架构设计。 未来3年,天数智芯将基于此次发布的四代架构,陆续发布多款产品,持续提升计算性能。 在1月26日召开的天数智芯"智启芯程"合作伙伴大会上,天数智芯AI与加速计算技术负责人单 天逸公布了超越英伟达相关产品的四代架构路线图: 2025年,天数天枢架构超越Hopper; 天数智芯董事会主席盖鲁江同时兼任公司CEO,其于2020年7月加入公司,并于2025年5月 调任执行董事,主要负责公司的日常管理及整体策略规划。在此之前,盖鲁江拥有约17年财 务及投资经验。 2026年1月8日,天数智芯在港股上市。1月26日,天数智芯股价以188.2港元/股收盘,大跌 7.65%,最新市值为478亿港元。 编辑:江右 校对:王玥 制作: 鹿米 审核:陈墨 注:本文封面图由AI生成 单天逸还介绍了公 ...
The Magnificent 7 in 2026: What Should Investors Expect?
Yahoo Finance· 2026-01-26 15:18
If you invested in all of the Magnificent 7 tech stocks in 2025, congratulations — your return for the year was well above the S&P 500. But if you invested in only five of them, your return might have lagged the broader stock markets. Only two Magnificent 7 stocks — Alphabet (GOOGL) and Nvidia (NVDA) — outperformed the S&P 500 in 2025, Bloomberg reported. The five that fell short of the mark were Amazon (AMZN), Apple (AAPL), Meta Platforms (META), Microsoft (MSFT) and Tesla (TSLA). It was the first time ...
新能源+AI周报(第40期20260118-20260124):储能量价齐升,太空、AI主题延续-20260126
Tai Ping Yang Zheng Quan· 2026-01-26 15:01
Investment Rating - The report does not provide specific investment ratings for the industry sectors mentioned [2]. Core Insights - The overall industry strategy focuses on the simultaneous rise in energy storage volume and price, with ongoing themes in space and AI [3]. - The new energy vehicle supply chain is entering an upward cycle, benefiting companies like CATL and EVE Energy due to the electrification upgrade and optimization of energy storage patterns [3]. - By the end of December 2025, China's electric vehicle charging infrastructure is expected to reach 20.092 million units, a year-on-year increase of 49.7% [3]. - Global energy storage battery shipments are projected to reach 640 GWh in 2025, a year-on-year increase of 82.9%, with CATL maintaining a leading position [3][25]. - The solid-state battery sector is entering a critical phase of engineering and industrialization, with companies like Xiamen Tungsten and Peking University Technology benefiting [4]. Summary by Sections Energy Storage and New Energy - The energy storage trend continues to improve, with companies like Sungrow Power and Huaneng Power benefiting from a significant increase in domestic procurement, which exceeded 100 GWh for the first time [5]. - The average price of lithium battery storage systems has rebounded by 6.39% to 0.5226 yuan/Wh [5]. - AI expansion and global grid upgrades are driving demand for power equipment, benefiting companies like TBEA and Sanyuan Electric [5]. Lithium Carbonate and Battery Materials - The supply and demand for lithium carbonate are exceeding expectations, with companies like Salt Lake Potash and Hunan Youneng benefiting from favorable market conditions [4]. - The cost of phosphoric iron lithium cathode materials has increased, with processing fees rising by 318.7 yuan/ton compared to November averages [4][27]. AI and Robotics in New Energy - The integration of AI and humanoid robots in the new energy sector is gaining traction, with companies like Zhejiang Rongtai and Keda Li benefiting from advancements in robotics [7]. - Tesla's shift towards becoming a robotics company is expected to create new growth cycles, with significant implications for the automotive industry [7][26]. Market Trends and Projections - The report highlights that the global energy storage battery shipment is expected to reach 1,090 GWh in 2026, a year-on-year increase of 70% [25]. - The market for commercial energy storage products is evolving, with larger capacity batteries becoming mainstream and driving innovation in the sector [29].
美股异动|CoreWeave大涨超15% 英伟达向公司追加投资20亿美元
Ge Long Hui A P P· 2026-01-26 14:58
格隆汇1月26日|云计算服务商CoreWeave(CRWV.US)大幅上涨超15%。消息面上,英伟达公司宣布向 CoreWeave Inc.追加投资20亿美元,支持后者在2030年前增加超过5吉瓦的AI算力基础设施。这笔投资是 英伟达利用其庞大资源推动人工智能产业发展的最新举措,也标志着这家全球市值最高的公司继续深化 其在AI基础设施领域的布局。 ...
Nvidia Just Invested Another $2 Billion Into CoreWeave. Time to Buy Again?
247Wallst· 2026-01-26 14:53
Core Investment - Nvidia initially invested approximately $100 million in CoreWeave in 2023 [1] - Nvidia increased its stake around CoreWeave's IPO in March [1]
英伟达注资20亿美元 CoreWeave涨15%
Xin Lang Cai Jing· 2026-01-26 14:45
核心要点 2026 年 1 月 20 日,在 瑞士达沃斯举行的世界经济论坛上,CoreWeave 首席执行官迈克尔・因特拉特做客美国消费者新闻与商 业频道(CNBC)《财经论谈》节目。 英伟达宣布向人工智能基础设施供应商 CoreWeave 投资 20 亿美元后,后者股价在周一暴涨 15%。 CoreWeave 的核心营收来源是建设并出租搭载英伟达图形处理器(GPU)的算力数据中心。这类图形处 理器是训练人工智能模型和运行大规模人工智能计算任务的关键硬件。部分投资者将该公司归为 "新型 云计算服务商",在日益紧密互联的人工智能基础设施合作网络中,CoreWeave 已成为不可或缺的关键 参与者。 事实上,英伟达早已是 CoreWeave 的主要投资方。 今年 9 月,CoreWeave 在美国证券交易委员会提交的一份文件中披露,其从英伟达获得了一笔价值至少 63 亿美元的订单。根据协议条款,英伟达有义务在 2032 年 4 月前,收购该公司 "剩余的未售出算力资 源"。 CoreWeave 于今年 3 月在纳斯达克挂牌上市,公司通过债权和股权融资的方式筹集了数十亿美元资金, 其中就包括来自英伟达的投资。 公告 ...
GPU vs ASIC的推理成本对比
傅里叶的猫· 2026-01-26 14:42
以下文章来源于More Than Semi ,作者mofan More Than Semi . More Than SEMI 半导体行业研究 在 AI 技术加速渗透的当下,芯片作为核心基础设施,其性能与成本的平衡直接影响着整个 AI 生态的 发展节奏。前几天高盛的一份报告中,通过构建 "推理成本曲线",量化分析了 GPU 与 ASIC 两类芯片 的竞争态势。 一、推理成本成 AI 芯片竞争关键 AI 芯片的竞争早已不局限于单纯的性能比拼,"性价比" 成为衡量核心竞争力的关键指标,而推理阶段 的成本控制更是重中之重。对于企业和开发者而言,推理成本直接关系到 AI 应用的商业化落地可行 性;对于芯片厂商来说,能否在推理成本上形成优势,决定了其市场份额的归属。 高盛通过对比英伟达、AMD 等厂商的商用 GPU 解决方案,以及谷歌 TPU、亚马逊 Trainium 等定制化 ASIC 方案,构建了跨周期的推理成本曲线。研究发现,随着技术迭代,各类芯片的推理成本均呈下降 趋势,但不同厂商的下降幅度和竞争态势存在显著差异。这一分析框架,不仅解答了投资者对两类芯片 技术路线优劣的疑问,也为行业参与者提供了清晰的竞争坐标。 ...
Nvidia's Unspoken Problem: 40% of Revenue Comes From Companies Developing Their Own AI Chips
247Wallst· 2026-01-26 14:40
Microsoft (NASDAQ:MSFT), Meta (NASDAQ:META), Amazon (NASDAQ:AMZN), and Alphabet (NASDAQ:GOOGL) represent 40-50% of Nvidia (NASDAQ:NVDA) revenue. Every single one is developing custom AI chips. Threat #1: Your Best Customers Are Building Your Replacement Jensen Huang has built a $4.6 trillion empire selling the picks and shovels of the AI revolution. But while he preaches accelerated computing, three existential threats aren't discussed in earnings calls. They're happening now. Google's TPUs power Bard and S ...
Nvidia invests $2B in CoreWeave to accelerate AI data center buildout
Proactiveinvestors NA· 2026-01-26 14:37
About this content About Emily Jarvie Emily began her career as a political journalist for Australian Community Media in Hobart, Tasmania. After she relocated to Toronto, Canada, she reported on business, legal, and scientific developments in the emerging psychedelics sector before joining Proactive in 2022. She brings a strong journalism background with her work featured in newspapers, magazines, and digital publications across Australia, Europe, and North America, including The Examiner, The Advocate, ...